INFLUENCE OF FEMORAL HEAD AND NECK CONFORMATION ON HIP DYSPLASIA IN THE GERMAN SHEPHERD DOG

Introduction

Canine hip dysplasia (CHD), a common skeletal disorder in dogs, is due to abnormal development of the coxofemoral joint.^1,2 The shape of the femoral head, femoral neck, and acetabulum vary by breed.^3–5 German Shepherd dogs and Rottweilers have a moderately deep acetabulum, whereas the Boxer and Labrador Retriever have a shallow acetabulum.⁶ The acetabulum of the Boxer, standard Poodle, and German Shepherd dog is almost circular, whereas in the Bassett the acetabulum is more elliptical.^3,4,7,8 In the German Shepherd dog, a femoral head with a width similar to the femoral neck has been termed broomstick-like femoral head and neck formation.⁹

Because of the wide range of hip joint conformation in dogs, hip scoring is complex and, although breeding programs have led to a general reduction in the prevalence of CHD, this approach is not always successful.^10–12 Additional radiographic parameters could be helpful in narrowing the number of dogs used for breeding in normal and mildly dysplastic groups. If broomstick-like conformation were a phenotypic sign of CHD, it may prove to be an additional criterion for the selection of breeding dogs.

The aims of our study were to (1) determine the prevalence of the broomstick-like conformation in the German Shepherd dog, (2) identify whether a torsional change at the level of the proximal femur is the reason for the radiographic change in size of the femoral head in dogs with a broomstick-like conformation by calculating the antetorsion angle, (3) assess joint laxity by calculating the distraction ratio, (4) assess the influence of broomstick-like conformation on the development of CHD, and (5) evaluate a genetic base for the variation of this femoral head/neck conformation.

Materials and Methods

Pelvic radiographs of 294 German Shepherd dogs, 187 (63.6%) males and 107 (36.4%) females were studied. The median and mean age of the dogs were 15 and 24.4 months, respectively. The prospective part of the study included 84 dogs screened for CHD between 2002 and 2004. The retrospective part involved studying pelvic radiographs of the remaining 210 German Shepherd dogs taken from the archive.

All dogs were premedicated with an intravenous (IV) combination of diazepam^* (1 mg/kg) and atropine sulfate (0.05 mg/kg).† General anesthesia was induced and maintained with a combination of IV ketamine hydrochloride (3 mg/kg)‡ and xylazine hydrochloride (0.3 mg/kg).§

The radiographic projections evaluated were as follows: standard hip-extended ventrodorsal view (VD), abducted VD (frog-leg view), a distraction view using a wooden lath¹³ (distraction view) (Fig. 1A), and a faux-profile view of the right and left femur.¹⁴ All radiographs were obtained with a vertical beam and the same focus–film and object–film distance.

For the distraction view, the dog was positioned in dorsal recumbency. A wooden, trapezoid-shaped board (length 80 cm, caudal width 8 cm, cranial width 16 cm) was placed between the hind legs.¹³ Contrary to the described method,¹³ we extended the hind legs to the point where further extension would have resulted in lateral deviation of the stifle joints (Fig. 1A and B). The centering point was the cranial aspect of the pubic symphysis. For the faux-profile view (Fig. 2), the dog was in lateral recumbency with the lowermost leg under investigation. A foam wedge was placed under the tarsal joint to level it with the stifle, which resulted in superimposition of the femoral condyles. The stifle was flexed, while holding the tibia at an angle of 90° relative to the femur with a sandbag. The uppermost leg was pulled cranially and held parallel to the body with a sandbag. The centering point was the mid femur.

The hip-extended radiograph was taken in all 294 dogs. This was the only view available for all the 210 dogs in the retrospective group. The complete set of five radiographic projections was obtained in the 84 dogs in the prospective group.

The diagnosis of a broomstick-like conformation (Group 1) was based on the appearance of the femoral head in the hip-extended radiograph. The morphologic features of the broomstick-like conformation were (a) same width of femoral head and neck and (b) absence of a transition zone between the femoral head and neck (Fig. 3A and B). Dogs with a unilateral broomstick-like conformation were not studied. Dogs without a broomstick-like conformation were classified as normal (Group 2). The group classification was carried out by one radiologist (A.W.) and verified by a board-certified radiologist (B.T.).

Calculations were performed in the 84 animals of the prospective part of the study, 51 of which belonged to Group 1 with broomstick-like conformation, and 33 to Group 2 with a normal conformation of the femoral head. Dogs from the retrospective part of the study were only used to determine the overall frequency, to identify the CHD score, and to identify the heritability of the broomstick-like conformation.

The size of femoral head and acetabulum was assessed on the hip-extended and frog-leg views. A circle was drawn around the femoral head and its center was identified. The contour of the acetabulum was drawn and extrapolated to form a full circle. After measuring the radius of the femoral head and acetabulum, the sectional plane of the femoral head and the acetabulum were calculated according to the formula for the area of a circle (A=πr² with A is the area, π is the mathematical constant approximately equal to 3.14159, and r is the radius of the circle). Henceforth, we will refer to the calculated area of the femoral head as femoral head size and to the calculated area of the acetabulum as acetabular size.

To identify if antetorsion of the proximal femur could be the reason for the broomstick-like conformation the antetorsion angle was calculated. Hip-extended and faux-profile views of the femora were used to calculate the antetorsion angle. The measurements were performed on both views in a similar fashion as in Fig. 2. The narrowest diameter of the femoral shaft was identified. At that level a line perpendicular to the cranial cortex was drawn extending from the periosteal aspect of the cranial cortex to the periosteal aspect of the caudal cortex. The same was carried out 2 cm proximal and distal to this line. A vertical line (line A) was drawn from the acetabulum to the distal pole of the patella, bisecting these three lines, thus identifying their center points. A circle was drawn around the femoral head and its center was determined. The center point of this circle and line A were connected by line B in such a way that a 90° angle ensued at the intersection of the two lines. Line B provides the y-value on the radiograph in the hip-extended view and the x-value on the faux-profile view. The x/y ratio represented the tangent for the desired angle α, which was calculated by further mathematic transformation.¹⁴

The distraction ratio was calculated by dividing the distance between the center of femoral head and ipsilateral acetabulum by the radius of femoral head.^13,15

The CHD assessment was carried out on the hip-extended view as required by the official guidelines of the Féderation Cynologique Internationale, where the grades are identified as follows: A=no sign of hip dysplasias, excellent; B=borderline, good; C=mild hip dysplasia; D=moderate hip dysplasia; E=severe hip dysplasia.⁴

For calculation of heritability (h²) of broomstick-like conformation, the pedigree information of the 180/294 dogs with this conformation was obtained from the German Shepherd Dog Breeding Association. Heritability (h²) was estimated according to the following formula:

where is the additive genetic variance and is the residual variance.¹⁶

In 20 randomly selected dogs all measurements were repeated five times to check the intraobserver repeatability of the results.

The measured femoral head and acetabular size were compared between Groups 1 (bilateral broomstick-like conformation) and 2 (no broomstick-like conformation) and within each group between the hip-extended and frog-leg position. The results for the antetorsion angle and the distraction ratio were compared between Groups 1 and 2. Statistical analysis was carried out using analysis of variance (ANOVA).¶ A P-value smaller than 0.05 was considered significant.

Results

Broomstick-like conformation was found in 61.2% of all German Shepherd dogs examined. Of the animals from the prospective part 51 and of the animals from the retrospective part 129 were affected, resulting in a total of 180 affected animals. Unaltered femoral head conformation was present in 33 dogs from the prospective and 81 from the retrospective part of the study, resulting in a total of 114 unaffected dogs. Calculations performed on all 84 German Shepherd dogs from the prospective part revealed that the average size of the femoral head in dogs with a broomstick-like conformation (51/180) on the hip-extended view was 4.5±0.6 cm². In German Shepherd dogs with a normal appearance (33/114), the average size (4.8±0.6 cm²) on the hip-extended view was significantly larger (P=0.036). There was no difference in the mean size of the femoral head on the frog-leg view between the two groups (Group 1, 5.5±0.7 cm²; Group 2, 5.7±0.6 cm²; P=0.194). There was no significant difference in the mean size of the acetabulum between groups on the hip-extended (6.4±0.7 cm² for Group 1 and 6.5±0.8 cm² for Group 2; P=0.461) and frog leg (6.6±0.7 cm² for Group 1 and 6.8±0.8 cm² for Group 2; P=0.113) views. Also, there was no significant difference in the distraction ratio (Group 1, 0.39±0.1; Group 2, 0.36±0.1; P=0.205) or the antetorsion angle (Group 1, 28.1°±6.1°; Group 2, 29.3°±5.9°; P=0.343) between the two groups. There was no difference between Groups 1 and 2 with respect to age or gender.

The official hip score was available for all 294 dogs examined (Table 1). There was no significant difference between the CHD score between Group 1 (bilateral broomstick-like conformation) and Group 2 (normal conformation). In Group 1, 39.5% of the dogs had an excellent and 29.4% had a good hip score. In Group 2, only 36% were scored as excellent and more hip joints were classified as good (36.8%). In each group, <3% of the dogs were characterized as severely dysplastic. No significant difference was found regarding the CHD score in Groups 1 and 2 when forming subgroups on the basis of age (Fig. 4A and B).

The heritability for all 180 dogs of Group 1 was 0.3±0.1 for broomstick-like conformation, indicating the presence of a genetic base. However, there was a considerable variance of the calculated heritability of broomstick-like conformation in randomly chosen samples from dogs of the same year/range of years of birth (Table 2).

The intraobserver repeatability was high for all values examined (0.92–0.75).

Discussion

Sixty-one percent of the German Shepherd dogs examined in our study had the broomstick-like conformation. To establish the size of femoral head and acetabulum, the accepted model of a circle was chosen.^3–5,7,8 The circular shape corresponded well with the outline of these two anatomic structures in the hip-extended and frog-leg views. The size of the femoral head in dogs with a broomstick-like conformation on the hip-extended view was smaller than in the frog-leg view. In normal dogs, there was no difference between the size of the femoral head in these two views. As a standard radiographic technique was used, the differences in size cannot be attributed to technique. A likely explanation is that in dogs with a broomstick-like conformation the femoral head is an ellipse rather than a sphere.

Computed tomography (CT) would have been valuable in understanding the exact anatomic formation of the femoral head in dogs with the broomstick-like conformation, but CT was not available to us during the study period.

Acetabular size between Groups 1 and 2, and within each group, was not significantly different when comparing the hip-extended with the frog-leg view. The size of the acetabulum on the hip-extended and the frog-leg views does not change significantly, which can be explained by the unchanged position of the pelvis, independent of the position of the hind legs. The authors have no explanation as to why the acetabular size is the same in Groups 1 and 2 while there is a significant difference in size of the femoral head. Further anatomic studies are necessary to identify the reason for this discrepancy.

The mean antetorsion angle for dogs in our study ranged from 28.1°±6.1° (Group 1) to 29.3°±5.9° (Group 2). These results agree with previously published data where the reference range for all breeds was between 24° and 36°, with a mean of 30.1°±6° for German Shepherd dogs.¹⁴ Others have measured a lower antetorsion angle.^17,18 This discrepancy is most likely due to differences in technique, because both reports are based on transverse radiographs where the antetorsion angle could be directly measured in contrast to the indirect method used in our study.^17,18

There was no difference in the distraction ratio between groups (Group 1, 0.39±0.1; Group 2, 0.36±0.1). Therefore, the broomstick-like conformation does not contribute to hip joint laxity when using this distraction method. Our values are lower than those of other published data.^19,20 These differences may be due to different inclusion criteria, a different study population, or our technique, which may have resulted in spiral twisting of the fibrous elements of the joint capsule leading to an increase in joint stability and decreased laxity.¹⁵

It has been theorized that the OFA database in the United States of America is biased toward better hip scores presumably due to voluntary submission of radiographs into the database pool.¹² The same may be true for the SV database in Germany. All dogs radiographed in the given time period were included in our study; therefore, a bias toward better hips does not exist in our study. It is, however, possible that preselection was made by owners, only submitting clinically normal dogs.

The excellent hip score in 39.5% of Group 1 dogs and 36.0% of Group 2 dogs are lower than reported by the SV database, where >60% scored excellent.²¹ Our results may represent a more accurate estimation of CHD in this breed because we did not preselect the animals and included all dogs in our calculation.

It seems that older dogs achieved slightly better hip dysplasia scores than younger dogs. One would expect the reverse because degenerative changes due to hip dysplasia become more apparent with age. A possible explanation is that owners only bring in older dogs for official hip scoring when they wish to use them for breeding based on good physical performance or on the basis of a negative radiographic examination of the pelvis obtained in the dog's youth.

The heritability for the broomstick-like conformation has not been previously reported. Heritability is defined as the fraction of the total phenotypic variance that is due to genetic differences.²² An h² of 0.3 or greater means that the genetic base of a sign is high enough to select breeding dogs for the desired characteristic and that the occurrence of the sign cannot be explained by environmental factors alone. A sign with an h² value of 1 would be entirely genetically based.²² Our results suggest that there is a genetic base for the broomstick-like conformation. The variation of the calculated h² in the four chosen sample groups indicates the need of a larger number of dogs for the calculation, because to determine an exact value for h² more than 1000 animals have to be used.^23,24

Limitations of our study are that dogs with a unilateral broomstick-like conformation were excluded, though only 17 of 311 reviewed dogs were thus affected, and that the gender effect could not be addressed due to the small size of the sample groups.

In conclusion, the broomstick-like conformation is common in German Shepherd dogs and suggests a genetic basis. The size of the femoral head in the broomstick-like conformation varied significantly between the hip-extended and frog-leg views, suggesting that the femoral head in dogs with a broomstick-like conformation is not spherical. The antetorsion angle between dogs with and without the broomstick-like conformation was similar, eliminating antetorsion as a possible cause for this anatomic variation. The distraction ratio was not significantly different in dogs with and without the broomstick-like conformation, indicating that the broomstick-like conformation does not contribute to hip joint laxity. The hip score in German Shepherd dogs with the broomstick-like conformation is similar to dogs with a normal hip formation. It is therefore unlikely that the presence of the broomstick-like conformation can be interpreted as another sign of hip dysplasia in the German Shepherd dog.