Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses

Livestock Science ] (]]]]) ]]]–]]]

Contents lists available at ScienceDirect

Livestock Science journal homepage: www.elsevier.com/locate/livsci

Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses Dorothea Hilla, Ottmar Distl n Institute for Animal Breeding and Genetics, University of Veterinary Medicine Hannover (Foundation), Bünteweg 17p, 30559 Hannover, Germany

a r t i c l e i n f o

abstract

Article history: Received 20 February 2014 Received in revised form 19 August 2014 Accepted 10 September 2014

In the present study, results of radiographic examinations were used for estimation of genetic parameters for osteoarthrosis (OA) in distal interphalangeal joints, proximal interphalangeal joints, fetlock, hock and stifle joints, radiographic changes in the navicular bones of the forelimbs (RCNB) and sidebone in 7396 Hanoverian Warmblood horses. Radiographic changes were recorded as ordinal traits in dependence of the number of joints affected or, in case of RCNB, of their severity score. In addition, their genetic correlations with osteochondrosis (OC) and osteochondrosis dissecans (OCD) in fetlock, hock and stifle joints as well as palmar/plantar osteochondral fragments (POFs) and dorsodistal osteochondral fragments in fetlock joints were determined. The prevalence for OA in hock joints was 21.4%, for OA in fetlock joints 7.4%, for RCNB 33.6% and for sidebone 9.6%. Horses with OA in most cases were affected at only one limb while RCNB and sidebone in most cases were found bilaterally. The heritabilities estimated in a linear animal model and transformed onto the liability scale were at 0.17 for OA in fetlock joints, 0.40 for OA in hock joints, 0.19 for RCNB and 0.59 for sidebone. Additive genetic correlations among OA in fetlock joints, OA in hock joints, RCNB and sidebone were moderate (rg ¼  0.19 to rg ¼ 0.18). The highest additive genetic correlations were found between OA in fetlock joints and OC in fetlock joints (rg ¼ 0.71) as well as OCD in fetlock joints (rg ¼0.62). RCNB showed positive genetic correlations with OC and OCD in the different joints as well as with OFs in hock and stifle joints. POFs and RCNB were genetically negatively correlated. The size of the heritability estimates for OA in fetlock joints, OA in hock joints, RCNB and sidebone seems to be sufficient high that breeding measures can be recommended in order to reduce their prevalence. & 2014 Published by Elsevier B.V.

Keywords: Osteoarthrosis Navicular bone disease Sidebone Osteochondrosis Heritability Genetic correlation

1. Introduction The presence of radiographic changes in the context of a pre-purchase examination has a significant effect on the market value of horses (van Hoogmoed et al., 2003).

n Corresponding author at: Bünteweg 17p, 30559 Hannover, Germany. Tel.: þ 49 511 953 8875; fax: þ 49 511 953 8582. E-mail address: [email protected] (O. Distl).

Osteoarthrosis (OA) (Oliver et al., 2008), radiographic changes in the navicular bones (RCNB) (Stock and Distl, 2006a) and the ossification of the cartilages of the distal phalanx of the forelimbs, also called sidebone (Ruohoniemi et al., 1993; Willms et al., 1999) are common findings in the limb joints of horses. Osteochondrosis (OC) and osteochondrosis dissecans (OCD) are most prevalent in growing horses and these conditions prone horses to increased risk of developing orthopaedic problems later in life (Distl, 2013). OA may lead to lameness (Gough and Munroe,

http://dx.doi.org/10.1016/j.livsci.2014.09.015 1871-1413/& 2014 Published by Elsevier B.V.

Please cite this article as: Hilla, D., Distl, O., Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses. Livestock Science (2014), http://dx.doi.org/10.1016/j.livsci.2014.09.015i

D. Hilla, O. Distl / Livestock Science ] (]]]]) ]]]–]]]

2

1998) and reduced performance (Stock and Distl, 2006b). RCNB are not necessarily associated with navicular disease, but in affected horses are often more severe (Dik et al., 2001). The clinical relevance of sidebone remains to be clarified (Ruohoniemi et al., 2003; Verschooten et al., 1996). Inheritance seems to play an important role for the occurrence of OC, OCD, OA, RCNB and sidebone (Axelsson et al., 2001; Bos et al., 1986; Distl, 2013; Ruohoniemi et al., 2003; Stock and Distl, 2006a). Heritability estimates for OA in Warmblood horses ranged between 0.16 and 0.31 (KWPN, 1994; Stock and Distl, 2006a), for RCNB in Hanoverian Warmblood horses between 0.09 and 0.40 (Stock and Distl, 2006c) and for sidebone in Finnhorses between 0.31 and 0.50 (Ruohoniemi et al., 2003). Heritability estimates for OC and OCD in warmblood horses were at 0.06–0.32 in fetlock joints, at 0.19–0.46 in hock joints and at 0.02–0.23 in stifle joints and for trotters between 0.21 and 0.45 (Distl, 2013; Hilla and Distl, 2014). To achieve genetic progress for orthopedic health traits, the size of the heritability of the target traits in the selection index and their genetic correlation structure have to be taken into account. The aim of this study was to estimate genetic parameters for OA, RCNB and sidebone in a large sample of Hanoverian Warmblood horses and to analyse their correlations with osteochondrosis (OC) and osteochondrosis dissecans (OCD) in fetlock, hock and stifle joints as well as with palmar/plantar osteochondral fragments (POFs) and dorsodistal osteochondral fragments (DOFs) in fetlock joints. This is the first report on additive genetic correlations of OA, RCNB and sidebone with OC, OCD, POFs and DOFs. The results of the present analysis should help to define a set of traits to be included in a merit index for orthopedic health in warmblood horses. 2. Material and methods Data of 7396 Hanoverian Warmblood horses that were presented for radiographic examinations in 2005–2012 was included in the present study. These horses took part in pre-selection for sale at auction as riding horses by the Association of Hanoverian Warmblood breeders in Verden (Aller), Germany, (n¼4259), were pre-selected for stallion licensing during the same time period (n ¼865) or were presented for other reasons, particularly for pre-purchase veterinary examination including a radiographic examination (n¼2272). Of these horses, 4480 were male and 2916 were female. The mean age of the horses was 3.9 years, with 776 horses aged less than 3 years, 4923 aged 3–4 years and 1698 aged more than 4 years. More details on the horses and the radiographic examination can be found elsewhere (Hilla and Distl, 2013, 2014). Pedigree information was provided from the unified animal ownership database (Vereinigte Informationssysteme Tierhaltung w.V., vit) in Verden/Aller, Germany. The horses were offspring of 675 sires, with a mean of 11 progeny per sire, ranging from 1 to 317, and of 5712 dams with a mean of 1.3 progeny per dam, ranging from 1 to 6. All horses underwent a standardized radiographic examination, with a minimum of twelve x-rays of the limbs, including laterolateral projections of the front and hind digits, dorsoproximal–palmarodistal projections of

the region of the navicular bone of the forelimbs according to Oxspring (1935), 451 and 1151 projections of the hock joints and 901 projections of the stifle joints. According to McIlwraith (1982), changes of the contour of the joint by periarticular osteophytes or exostoses and narrowing or loss of the joint space were classified as signs of OA. Their presence was recorded in distal interphalangeal joints (DIJ) and proximal interphalangeal joints (PIJ), fetlock, hock and stifle joints. Diagnostic criteria for RCNB were shape, symmetry, contour and structure of the navicular bone as well as form, size, number and location of the canales sesamoidales (Brunken, 1986; Diesterbeck et al., 2007; MacGregor, 1986; Stock and Distl, 2006a; Ueltschi, 2001). In addition, the presence of sidebone was regarded. Criteria for diagnosis of OC and OCD included all radiographic findings consistent with OC and OCD at the specific predilection sites (Distl, 2013; van Weeren, 2006). These sites were in fetlock joints the dorsoproximal part of the sagittal ridge of the third metacarpal or metatarsal bone, in hock joints the sagittal ridge of the distal tibia, the lateral and medial malleolus of the tibia, the lateral and medial trochlea and the basis of the talus, and in stifle joints the lateral and medial trochlea of the femur, the sulcus intertrochlearis and the patella (van Weeren, 2006; van Grevenhof et al., 2009). Only those osteochondral fragments were classified as OCD that were within one of these predilection sites. A distinction has to be made to other osteochondral fragments, such as those palmar or plantar in fetlock joints and at the attachment sites of the short sesamoidean ligaments on the proximal phalanx (POFs) (Carlsten et al., 1993; Sandgren et al., 1993) or those dorsodistal in fetlock joints (DOFs). In fetlock joints, we distinguished between the traits OC, OCD, DOFs and POFs (Hilla and Distl, 2013, 2014).

2.1. Statistical analysis OA in DIJ, PIJ, fetlock, hock and stifle joints, just as sidebone, were analysed as ordinal traits in dependence of the number of joints affected. The score ranged from 0 to 4 (0¼ free from radiographic changes for the specific condition, 1 to 4¼ radiographic changes at a particular localisation in one, two, three or four limbs). RCNB were classified into a severity score from 0¼free from radiographic changes to 3¼severe radiographic changes. Few conical canales sesamoidales or slight alterations of the contour of the navicular bone were classified as slight, deformed canales sesamoidales or changes of the structure of the navicular bone as moderate and branched or bulb-ended canales sesamoidales as severe radiographic changes of the navicular bone (Diesterbeck et al., 2007; Stock and Distl, 2006a). Prevalences of OA in DIJ, PIJ, fetlock, hock and stifle joints, sidebone and RCNB were analysed using the procedure FREQ of SAS, version 9.4 (SAS, 2013). Analysis of variance was employed to evaluate the influences of gender, month of birth, age group, year of examination and reason for radiographic examination on the prevalence of OA, RCNB and sidebone. The analyses were performed using the procedure MIXED of SAS.

Please cite this article as: Hilla, D., Distl, O., Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses. Livestock Science (2014), http://dx.doi.org/10.1016/j.livsci.2014.09.015i

D. Hilla, O. Distl / Livestock Science ] (]]]]) ]]]–]]]

Heritabilities, genetic, residual and phenotypic correlations for OA in DIJ, PIJ, fetlock, hock and stifle joints, RCNB and sidebone were estimated using Residual Maximum Likelihood (REML) with VCE-5, Version 5.1.2 (Variance Component Estimation, Kovač and Groeneveld, 2003) performing multivariate analyses with the following linear animal model: yijklmn ¼ μ þ Genderi þMonthj þ Agek þYearl þ Categorym þ animaln þ eijklmn with yijklmn ¼respective radiographic findings of the ijklmn-th horse encoded as categorical traits, μ ¼model constant, Genderi ¼fixed effect of the gender (i¼1–2), Monthj ¼fixed effect of the month of birth (j¼1–5; 1 ¼November–February, 2 ¼March, 3 ¼April, 4 ¼May– October and 5¼not available), Agek ¼ fixed effect of the age group (k¼ 1–3; 1 ¼horses younger than 3 years, 2 ¼horse aged 3 to 4 years and 3 ¼horses older than 4 years), Yearl ¼fixed effect of the year of examination (l ¼1– 7), Categorym ¼fixed effect of the reason for the radiographic examination (m ¼1–3, 1 ¼pre-selection for sale at auction, 2¼pre-selection for stallion licensing, 3 ¼other reasons), animaln ¼ random effect of the nth animal (n ¼1– 39,265, with 6165 base animals and 33,100 non-base animals from a six generation pedigree) and eijklmn ¼ residual error. When only few observations are present at the boundaries of a scale, then these observations may have much influence on the outcome of an analysis. Thus, we employed classes for age and month. Warmblood horses have seasonal foaling from December to June with much fewer births outside these months. The age of the horses at examination was centered between 3 to 4 years and a low number was older than 5 years. Multivariate analyses were performed using following combinations of traits: OA in fetlock joints, OA in hock joints, RCNB and sidebone or OA in DIJ, PIJ, fetlock, hock and stifle joints or OA in fetlock joints, OA in hock joints, RCNB and sidebone and each one trait representing OC or OCD in fetlock, hock or stifle joints, OFs in fetlock, hock or stifle joints, POFs or DOFs in fetlock joints of the hindlimb. The genetic correlations were estimated among the examined traits and OC, OCD, POFs and DOFs in fetlock joints as well as OC and OCD in hock and stifle joints. To countervail against underestimation, heritability estimates, residual and phenotypic correlations obtained by the linear animal model were transformed onto the liability scale according to Vinson et al. (1976). Traits are designed with subscripts m and n and thresholds for the traits with subscripts i and j. h  σ 2obs ¼ obs
2 1 ∑ti  ¼ 1 Z im 2

2 hliab

r e; liab ¼

r p; liab ¼ 2

r e;obs σ m σ n
 t  1  ∑j ¼ 1 Z in

1 ∑it  ¼ 1 Z im

r p;obs σ m σ n
 t  1  ∑j ¼ 1 Z in

1 ∑it  ¼ 1 Z in

with hliab ¼estimated heritability of the trait m on the 2 underlying continuous scale, hobs ¼estimated heritability

3

of the trait m on the observed scale, σ 2m ¼phenotypic variance of the trait m, Z im ¼ordinate of a standard normal distribution at the threshold point i for the trait m, r e; liab ¼ residual correlation among the traits m and n on the underlying continuous scale, r e; obs ¼residual correlation among the traits m and n on the observed scale, σ m ¼ phenotypic standard deviation of the trait m, σ n ¼ phenotypic standard deviation of the trait n, Z im ¼ordinate of a standard normal distribution at the threshold point i for the trait m, Z jn ¼ordinate of a standard normal distribution at the threshold point j for the trait n, r p; liab ¼phenotypic correlation among the traits m and n on the underlying continuous scale and r p; obs ¼phenotypic correlation among the traits m and n on the observed scale. In contrast to residual and phenotypic correlations, additive genetic correlations estimated in a linear model are identical to those estimated in a threshold model (Stock et al., 2005). In addition, data was analysed using a linear mixed sire model and the procedure MIXED of SAS as well as a threshold sire model and the procedure GLIMMIX with a cumulative logit link function for multinomial traits:

λijklmn ¼ μ þ Genderi þ Monthj þ Agek þ Yearl þ Categorym þ siren with λijklmn ¼logit for the radiographic finding of the ijklmn-th horse encoded as a categorical trait and siren ¼random effect of the nth sire (n¼ 1–675). Heritability estimates from the linear mixed sire model were transformed onto the liability scale according to Vinson et al. (1976) and compared with those estimated by the threshold sire models in order to validate the estimates obtained by the linear models. 3. Results 3.1. Prevalence The prevalences of the different radiographic findings examined in this study are given in Table 1. A total of 28.3% of the examined horses were affected by OA in at least one limb joint. The prevalence of OA in fetlock joints was 7.4% and in hock joints 21.4%, whereby most horses were affected at only one limb. RCNB were present in 33.6% of the horses. The majority of horses thereby showed slight RCNB. Sidebone was diagnosed in 9.6% of the horses, in most cases bilaterally. 3.2. Analysis of variance The p-values for the fixed effects obtained by a linear sire model are presented in Table 2. All fixed effects were highly significant for the occurrence of sidebone and all but the month of birth on the occurrence of RCNB. The effect of the year of examination was significant for OA in DIJ, PIJ, fetlock joints and hock joint, while the effect of the reason for the radiographic examination was significant for OA in DIJ and in PIJ. The 95% confidence intervals of the least square means (CI) of the effect of the different reasons for radiographic examination were significantly different for OA in DIJ, PIJ and fetlock joints, RCNB and sidebone (Table 3). With a 95% CI of 0.19–0.25 females

Please cite this article as: Hilla, D., Distl, O., Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses. Livestock Science (2014), http://dx.doi.org/10.1016/j.livsci.2014.09.015i

D. Hilla, O. Distl / Livestock Science ] (]]]]) ]]]–]]]

4

Table 1 Prevalence (%) of osteoarthrosis (OA), radiographic changes in the navicular bone (RCNB) and sidebone in the Hanoverian warmblood horses examined (n¼7396). Radiographic finding

In total

Number of limbs affected 1

OA DIJ OA PIJ OA fetlock OA hock OA stifle OA total RCNB1 RCNB2 RCNB3 RCNB Sidebone

1.6 0.8 7.4 21.4 0.4 28.3 30.5 2.7 0.3 33.6 9.6

2

1.2 0.6 4.6 13.5 0.3

0.4 0.2 2.2 7.9 0.1

3.0 0.4 0.1 3.5 1.2

27.5 2.3 0.2 30.0 8.4

3

4

0.0 0.0 0.4

0.0 0.0 0.2

RCNB1 ¼ slight radiographic changes, RCNB2 ¼moderate radiographic changes, RCBN3¼ severe radiographic changes, DIJ ¼ distal interphalangeal joints, PIJ ¼proximal interphalangeal joints.

Table 3 95% confidence interval of the least square means for the effect of the reason for radiographic examination on the occurrence of osteoarthrosis (OA) in distal (DIJ) and proximal interphalangeal joints (PIJ), fetlock, hock and stifle joints, radiographic changes of the navicular bones of the forelimbs (RCNB) and sidebone. Radiographic finding and limb joint

Horses intended for sale at auction

Horses intended Horses presented for for stallion other reasons licensing

OA DIJ OA PIJ OA Fetlock OA Hock OA Stifle RCNB Sidebone

 0.001–0.015  0.003–0.008 0.104–0.146 0.262–0.328 0.001–0.007 0.346–0.403 0.211–0.271

0.002–0.030 0.003–0.023 0.036–0.111 0.219–0.328 0.000–0.013 0.337–0.434 0.075–0.174

0.024–0.041 0.006–0.018 0.083–0.127 0.238–0.306  0.001–0.006 0.283–0.344 0.121–0.183

and sidebone these estimates were lower by 5.1 and 1.9%, respectively. 3.4. Genetic correlations

Table 2 P-values of the analysis of variance for the effect of gender, month of birth, year of examination, age group and reason for radiographic examination on the occurrence of osteoarthrosis (OA) in distal interphalangeal joints (DIJ), proximal interphalangeal joints (PIJ), fetlock, hock and stifle joints, radiographic changes of the navicular bones of the forelimbs (RCNB) and sidebone in 7396 Hanoverian Warmblood horses. Radiographic finding and limb joint

Gender

Month of birth

Year

Age group

Reason

OA DIJ OA PIJ OA Fetlock OA Hock OA Stifle RCNB Sidebone

0.055 0.122 0.402 0.052 0.940 o 0.001n o 0.001n

0.092 0.145 0.649 0.051 0.199 0.061 o 0.001n

o 0.001n o 0.001n o 0.001n 0.019n 0.080 o 0.001n o 0.001n

0.139 0.122 0.697 0.930 0.759 o0.001n o0.001n

o0.001n 0.006n 0.056 0.364 0.504 o0.001n o0.001n

n

The additive genetic correlation between OA in fetlock and in hock joints was positive and among OA in hock joints and sidebone negative (Table 4). A highly positive additive genetic correlation was found among OA in fetlock joints and OC in fetlock joints and a moderately to highly positive correlation between OA in fetlock joints and OCD in fetlock joints (Table 6). RCNB was positively genetically correlated with OC and OCD in stifle joints. Between RCNB and POFs in total as well as POFs of the hindlimbs, the additive genetic correlations were negative. The residual correlations were generally low and ranged at 0.02–0.07 after transformation on the liability scale (Table 4). An exception was the residual correlation between OA in hock joints and sidebone with re,liab ¼0.20. Phenotypic correlations were close to zero, varying at 0.01–0.03.

Significant p-value.

4. Discussion were considerably more frequently affected by sidebone than males with a 95% CI of 0.10–0.15.

3.3. Heritability estimates The highest heritability estimate after transformation onto the liability scale was found for sidebone with 0.59, followed by OA in hock joints with 0.40, RCNB with 0.19 and OA in fetlock joints with 0.17 (Table 4). Heritability estimates for OA in DIJ with 0.19 70.07, for OA in PIJ with 0.20 70.10 and for OA in stifle joints with 0.37 70.17 were not significantly different from zero due to high standard errors. In the threshold sire model, heritability estimates for OA were at 0.06 and 0.27 (Table 5). The highest heritability was estimated for sidebone with 0.51. The heritability estimate obtained from the linear sire model after transformation on the liability scale exceeded the heritability from the threshold sire model by 4.1% for OA in fetlock joints and for OA in hock joints by 3.4%, while for RCNB

In the present study, genetic parameters for OA in DIJ, PIJ, fetlock, hock and stifle joints as well as for RCNB and sidebone were estimated in a large sample of Hanoverian Warmblood horses. The current study revealed that, as in Icelandic horses (Björnsdóttir et al., 2000) and Thoroughbreds (Oliver et al., 2008), OA is also a common finding in Hanoverian Warmblood horses. The prevalences for OA in the examined limb joints in this study were considerably higher (28.3% versus 15.6%) than those determined in a previous study by Stock et al. (2004a). In this latter study, Hanoverian Warmblood horses that were selected for sale at auction as riding horses were analysed. The difference was largest for OA in fetlock joints, where Stock et al. (2004a) only found a prevalence of 1.2%. A possible reason for these large differences may be positive time trends and improvements in diagnosis through digital radiography. In the present study, we could use digital radiographs for all horses examined, whereas in the previous studies by Stock et al. (2004a,b) and Stock and Distl (2006a–c) digital radiographs were not yet available. With about one third

Please cite this article as: Hilla, D., Distl, O., Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses. Livestock Science (2014), http://dx.doi.org/10.1016/j.livsci.2014.09.015i

D. Hilla, O. Distl / Livestock Science ] (]]]]) ]]]–]]]

5

Table 4 Heritability estimates (on the diagonal), additive genetic correlations (above the diagonal), residual correlations (below the diagonal) with their standard errors and phenotypic correlations (below the diagonal, last line) for osteoarthrosis (OA) in interphalangeal (DIJ), proximal interphalangeal (PIJ), fetlock, hock and stifle joints, radiographic changes of the navicular bone (RCNB) and sidebone. Radiographic finding

OA DIJ

OA PIJ

OA fetlock

OA hock

OA stifle

RCNB

Sidebone

OA DIJ

0.02 7 0.01 0.1970.07 0.40 7 0.11 0.03  0.02 7 0.06 0.01 0.047 0.05 0.01  0.117 0.11 0.00 0.08 7 0.11 0.01 0.017 0.01 0.00

0.517 0.28

0.707 0.19

0.28 70.17

0.1770.19

0.03 7 0.19

0.03 70.19

0.017 0.01 0.207 0.10 0.017 0.08 0.008  0.157 0.07 0.00  0.277 0.11 0.00 0.0470.11 0.00 0.02 70.01 0.00

0.83 70.13

0.65 70.16

0.96 7 0.19

0.08 7 0.19

0.02 70.17

0.05 70.01nn 0.17 70.04nn 0.077 0.03 0.03  0.28 7 0.10 0.00 0.05 70.02 0.01 0.047 0.03 0.02

0.20 70.12

0.99 7 0.03

 0.05 7 0.06

0.107 0.04n

0.217 0.02nn 0.407 0.04nn  0.14 70.08 0.00 0.03 70.02 0.02 0.20 70.05 0.02

0.29 7 0.16

0.08 7 0.03

 0.19 70.07n

0.01 70.01 0.377 0.17 0.01 70.05 0.00  0.0770.05  0.01

 0.047 0.07

0.78 70.40

0.157 0.02nn 0.1970.02nn 0.02 7 0.03 0.02

0.147 0.07n

OA PIJ OA fetlock OA hock OA stifle RCNB Sidebone

0.20 70.02nn 0.597 0.05nn

Heritability estimates are given in bold after transformation onto the liability scale according to Vinson et al. (1976). Residual and phenotypic correlations are also transformed onto the liability scale according to Vinson et al. (1976). n Additive genetic correlations significantly different from zero. nn Heritability significantly different from zero.

Table 5 Heritabilities estimated using at threshold sire model (TSM) and a linear mixed sire model after transformation onto the liability scale according to Vinson et al. (1976) (LSMVin). Radiographic finding

TSM h2 7SE

LSM(Vin) h2 7SE

OA fetlock OA hock RCNB Sidebone

0.06 70.01nn 0.27 70.01nn 0.22 70.01nn 0.51 70.02nn

0.10 70.01nn 0.30 70.01nn 0.17 70.01nn 0.49 70.01nn

OA ¼ osteoarthrosis; RCNB ¼radiographic changes of the navicular bone. nn Heritability estimate significantly different from zero.

affected horses, RCNB was the most frequent finding in agreement with previous studies in warmblood horses (KWPN, 1994; Stock and Distl, 2006a; Willms et al., 1999). Like in previous studies, most of the horses showed only slight alterations. Prevalences of RCNB decreased with year of examination in the present study. The prevalence determined for sidebone (9.6%) was in the range of that one reported by Verschooten et al. (1996), who found sidebone in 10% of 2100 Warmblood horses. Studies in Finnhorses (Ruohoniemi et al., 1997, 1993) and Coldblood horses (Holm et al., 2000; Verschooten et al., 1996) revealed much higher prevalences of 52.6 to 80.0%. In two other studies, also very high prevalences were determined in Warmblood horses (Melo E Silva and Vulcano, 2002; Willms et al., 1999), but these studies were performed in very small horse samples (n o500) and thus, possibly not representative for the respective populations. The prevalence of sidebone was significantly higher in females than in males in accordance to studies in Finnhorses (Ruohoniemi et al., 1997), Brazilian jumper horses (Melo E Silva and Vulcano, 2002) and Norwegian coldblood horses (Holm et al., 2000). The higher prevalence of sidebone in female horses seems to be caused by the significantly faster progression of ossification in females

compared to males (Ruohoniemi et al., 1997). The gender effect was also significant for RCNB with significantly higher prevalences in males. Other studies did mostly not show significant effects of gender on RCNB (Stock et al., 2004b). The horses in the current study were assorted in three different categories according to the different reasons for radiographic examination, namely pre-selection for sale at auction, pre-selection for stallion licensing and for other reasons (Hilla and Distl, 2014). The effect of the category in the current study was not significant for OA in fetlock, hock and stifle joint. In the case of OA in PIJ and DIJ, the observed significance may be due to the low prevalence of both findings in the examined population. Although horses envisaged for sale at auction or stallion licensing were pre-selected by their breeders or owners and by the Hanoverian Horse Breeding Society, the prevalence for RCNB was lower in horses examined due to other reasons compared to the other two categories. A pre-selection based on conformation therefore does not seem to be sufficient to reduce the prevalence of RCNB. Sidebone was most prevalent in horses pre-selected for auctions. The lower prevalence of sidebone in horses intended for stallion licensing is mainly caused by the gender effect and possibly stronger pre-selection for conformation. Sires with 45 progeny were represented in all three categories. We assume that sampling of horses for preselection for sale at auction or pre-selection for stallion licensing or other reasons did not influence the estimates of heritabilities. We could show that variance components for the interaction of sire by category were close to zero and not significant for all traits studied here. Accordingly, the analyses for heritabilities in the examined population across categories do not differ from those within category. Large sample sizes allow estimation of genetic parameters with low standard errors and thus, return estimates with higher reliabilities. In small samples, the distribution of radiographic findings may even not be

Please cite this article as: Hilla, D., Distl, O., Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses. Livestock Science (2014), http://dx.doi.org/10.1016/j.livsci.2014.09.015i

D. Hilla, O. Distl / Livestock Science ] (]]]]) ]]]–]]]

6

Table 6 Additive genetic correlations (rg) and their standard errors among osteoarthrosis (OA) in fetlock and hock joints, radiographic changes of the navicular bone (RCNB), sidebone and osteochondrosis (OC), osteochondrosis dissecans (OCD) as well as different types of osteochondral fragments in fetlock, hock and stifle joints. Radiographic finding

OA fetlock

OA hock

RCNB

Sidebone

OC fetlock OC hock OC stifle OCD fetlock OCD hock OCD stifle OFs fetlock OFs hock OFs stifle POFs total POFs hindlimb DOFs total DOFs hindlimb

0.7170.13n 0.1070.10 0.0470.18 0.62 70.14n 0.08 70.05  0.34 70.20 0.4770.17n  0.02 70.10  0.4170.21  0.03 70.19  0.02 70.17 0.03 70.06  0.1170.18

0.10 70.10 0.117 0.09  0.047 0.10 0.017 0.11 0.087 0.08  0.047 0.10 0.087 0.05 0.117 0.09  0.027 0.11 0.077 0.10 0.077 0.11 0.087 0.03 0.047 0.05

0.1170.06 0.1170.09 0.25 70.11n 0.02 70.04 0.1670.08n 0.43 70.14n  0.1770.08n 0.1970.08n 0.40 70.13n  0.3570.10n  0.3370.09n  0.0470.03 0.1370.10

0.067 0.10 0.067 0.03  0.09 7 0.08 0.107 0.12 0.047 0.03  0.167 0.06n 0.067 0.05 0.05 7 0.02n  0.217 0.07n  0.047 0.04  0.02 7 0.03  0.09 7 0.04  0.107 0.07

POFs ¼palmar or plantar osteochondral fragments in fetlock joints, DOFs ¼dorsodistal osteochondral fragments in fetlock joints, OFs ¼osteochondral fragments in general (including OCD, POFs and DOFs). n Additive genetic correlations significantly different from zero.

representative for most of the sires employed in the population. When the number of progeny per sire is low ( o5), estimates of genetic parameters may be considerably influenced by extreme values of progeny means (Stock et al., 2007). Particularly, in small samples, heritabilities may be over- or underestimated in sire models. To counteract this problem in this study, genetic parameters were estimated using a linear animal model and subsequently transformed onto the liability scale. Simulation studies revealed that threshold models perform more accurately in analyses with binary data than linear models (Stock et al., 2007). Heritabilities estimated under a linear animal model and afterwards transformed onto the liability scale were similar to the true estimates determined by a Monte Carlo simulation in a previous study of Stock et al. (2005). Validation of the heritability estimates using a threshold sire model in the present study indicated either under- or overestimation at a maximum of 4–5% in regard to the estimates in a linear sire model and transformation onto a liability scale. In contrast to Stock et al. (2005), where underestimation could be up to 9% and overestimation up to 18%, linear estimates transformed onto the liability scale were more close to the estimates from a threshold model in the present study. The heritabilities estimated in the current study had low standard errors for the most part and thus, were significantly different from zero. Exceptions were heritability estimates for OA in DIJ, PIJ and stifle joints, where heritability estimates were not at least three times higher than their standard errors. Due to very low prevalences of these conditions in the examined population (0.4–1.6%), reliable statements on the size of their heritabilities cannot be made. Standard errors of heritability estimates for OA in DIJ and PIJ obtained in a study of Stock et al. (2004a) were similarly high. To the authors’ best knowledge there are no other studies that investigated the heritability of OA in stifle joint in horses. In humans, OA in hands, knees and hips was shown to be heritable with estimates of h2 ¼0.28–0.68 (MacGregor et al., 2009).

The heritability estimate for OA in hock joints after transformation onto the liability scale was higher than estimates from other studies in Warmbloods (KWPN, 1994; Stock and Distl, 2006b). In contrast to the study by Stock and Distl (2006b), OA in hock joints in the current study was analysed as multinomial trait and not as a binomial trait. In Icelandic horses, for OA in hock joints a familial predisposition was found (Axelsson et al., 2001). Heritabilities were at 0.10–0.33 (Árnason and Björnsdóttir, 2003). In contrast to OA in hock joints, the heritability estimate for OA in fetlock joint was lower in Hanoverian Warmblood horses. Studies on the heritability of OA in fetlock joints are rare. A study in Dutch Warmblood horses revealed a heritability of 0.26 in a linear animal model, but the estimate had a high standard error of 0.15 (KWPN, 1994) and thus, was not significantly different from zero. The situation was similar in a study of Stock et al. (2004a). The moderate heritability indicates that this radiographic finding is more likely influenced by factors others than inheritance. The heritability estimate for RCNB was in the range of other studies in Warmblood horses (KWPN, 1994; Stock et al., 2004b; Willms et al., 1999), but with a much lower standard error. For sidebone, heritability estimates in Finnhorses ranged between 0.31 and 0.50 and thus, were lower than in the current study (Ruohoniemi et al., 2003). Comparing these studies, it has to be taken into account that different evaluation schemes were used. While in the current study the ossification of the cartilages of the distal phalanx was regarded as all-or-none trait and only uni- or bilaterality of ossification signs was regarded, the evaluation scheme of Ruohoniemi et al. (2003) included a severity score from 0 to 5 and thus, results might be different. Furthermore, the population of Ruohoniemi et al. (2003) with 964 Finnhorses was considerably smaller than in the current study. This is the first time that additive genetic correlations with OC and OCD in fetlock, hock and stifle joints as well

Please cite this article as: Hilla, D., Distl, O., Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses. Livestock Science (2014), http://dx.doi.org/10.1016/j.livsci.2014.09.015i

D. Hilla, O. Distl / Livestock Science ] (]]]]) ]]]–]]]

as POFs and DOFs in fetlock joints were estimated for OA in fetlock and hock joints, RCNB and sidebone. Highly positive genetic correlations were found between OA in fetlock joints and OC as well as OCD in fetlock joints. Neither POFs nor DOFs were significantly genetically correlated to OA in fetlock joint. This supports the results by Hilla and Distl (2014), who found that OCD, POFs and DOFs in fetlock joints should be regarded as different traits to avoid bias of results. Despite the moderate heritability of OA in fetlock joints (h2 ¼0.17), highly positive genetic correlations with OC (rg ¼0.71) and OCD in fetlock joints (rg ¼0.62) were evident. The prevalence of OA in fetlock joints could probably be reduced through correlated responses caused by the traits OC and OCD in fetlock joints in selection for orthopedic health traits. A moderate positive additive genetic correlation (rg ¼0.43) was also determined among RCNB and OCD in stifle joints what should help to reduce these both conditions in horse breeding programs. However, the negative additive genetic correlation among RCNB and POFs may reduce selection response for these traits. The residual correlations after transformation onto the liability scale among most of the examined traits were low. Merely among OA in hock joints and sidebone, a higher residual correlation was estimated. In a study of Stock and Distl (2006c) the residual correlation between OA in hock joints and RCNB with 0.10 was also higher. Phenotypic correlations, which were also transformed onto the liability scale, were close to zero. In a study of Willms et al. (1999), the phenotypic correlation among OA in hock joints and sidebone was in the same range. OA should be included in the breeding objectives because this condition may lead to lameness (Gough and Munroe, 1998) and reduced performance (Stock and Distl, 2006b). Although clinical RCNB is not necessarily associated with navicular disease (Dik et al., 2001) and the clinical relevance of sidebone is questionable (Ruohoniemi et al., 2003; Verschooten et al., 1996), the presence of these radiographic findings may influence the sale of horses (van Hoogmoed et al., 2003). Nevertheless, decreasing their incidences would be desirable. 5. Conclusions Common findings in the limb joints of Hanoverian Warmblood horses were OA in hock joints, RCNB and sidebone. Heritability estimates are sufficient for a genetic response to be expected on these three conditions. Due to their clinical relevance and prevalence, besides osteochondrosis in fetlock, hock and stifle joints, POFs in the hindlimbs and DOFs, OA in hock joints and RCNB in frontlimbs should be included in an orthopedic health index. Despite its large heritability, sidebone may not be recommended as a selection objective due to its unclear clinical relevance but should be included as an optional trait for information of the breeders. Conflict of interest statement The authors declare that there are no conflicts of interest related to this study.

7

Acknowledgements We would like to thank the Association of Hanoverian Warmblood breeders (Grant no. DI-HAN-1/1) for support in data collection and the involved veterinarians for making the radiographs available. References Árnason, T., Björnsdóttir, S., 2003. Heritability of age-at-onset of bone spavin in Icelandic horses estimated by survival analysis. Livest. Prod. Sci. 79, 285–293. Axelsson, M., Björnsdóttir, S., Eksell, P., Haggstrom, J., Sigurdsson, H., Carlsten, J., 2001. Risk factors associated with hindlimb lameness and degenerative joint disease in the distal tarsus of Icelandic horses. Equine Vet. J. 33, 84–90. Björnsdóttir, S., Árnason, T., Axelsson, M., Eksell, P., Sigurðsson, H., Carlsten, J., 2000. The heritability of degenerative joint disease in the distal tarsal joints in Icelandic horses. Livest. Prod. Sci. 63, 77–83. Bos, H., van der Meij, G.J.W., Dik, K.J., 1986. Heredity of navicular disease. Vet Q. 8, 68–72. Brunken, E., 1986. Röntgenologische Verlaufsuntersuchung am Strahlbein des Pferdes Tierärztliche Hochschule Hannover. Hannover. Carlsten, J., Sandgren, B., Dalin, G., 1993. Development of osteochondrosis in the tarsocrural joint and osteochondral fragments in the fetlock joints of Standardbred trotters I. A radiological survey. Equine Vet. J. 16, 42–47. Diesterbeck, U.S., Hertsch, B., Distl, O., 2007. Genome-wide search for microsatellite markers associated with radiologic alterations in the navicular bone of Hanoverian warmblood horses. Mamm. Genome 18, 373–381. Dik, K.J., van den Belt, A.J., van den Broek, J., 2001. Relationships of age and shape of the navicular bone to the development of navicular disease: a radiological study. Equine Vet. J. 33, 172–175. Distl, O., 2013. The genetics of osteochondrosis. Vet. J. 197, 13–18. Gough, M., Munroe, G., 1998. Decision making in the diagnosis and management of bone spavin in horses. Practice 20, 252–259. Hilla, D., Distl, O., 2013. Prevalence of osteochondral fragments, osteochondrosis dissecans and palmar/plantar osteochondral fragments in Hanoverian Warmblood horses. Berl. Münch. Tierärztl. Wochenschr 126, 236–244. Hilla, D., Distl, O., 2014. Heritabilities and genetic correlations between fetlock, hock and stifle osteochondrosis and fetlock osteochondral fragments in Hanoverian Warmblood horses. J. Anim. Breed. Genet. 131, 71–81. Holm, A.W., Bjornstad, G., Ruohoniemi, M., 2000. Ossification of the cartilages in the front feet of young Norwegian coldblooded horses. Equine. Vet. J. 32, 156–160. Kovač, M., Groeneveld, E., 2003. VCE-5 Users Guide and Reference Manual Version 5.1.2. Instititute for Animal Science and Animal Behaviour. Federal Agricultural Research Centre (Bundesforschungsanstalt für Landwirtschaft, FAL), Mariensee/Neustadt, Germany. KWPN, 1994. The Frequency and Heredity of Navicular Disease, Sesamioditis Fetlock Joint Arthrosis, Bone Spavin and Osteochondrosis of the Hock. A Radiographic Progeny Study. KWPN (Koninklijke Vereiniging Warmbloed Paardenstamboek Nederland), ZeistKWPN (Koninklijke Vereiniging Warmbloed Paardenstamboek Nederland), Zeist. MacGregor, C., 1986. Radiographic assessment of navicular bones, based on changes in the distal nutrient foramina. Equine Vet. J. 18, 203–206. MacGregor, A.J., Li, Q., Spector, T.D., Williams, F.M.K., 2009. The genetic influence on radiographic osteoarthritis is site specific at the hand, hip and knee. Rheumatology 48, 277–280. McIlwraith, C.W., 1982. Current concepts in equine degenerative joint disease. J. Am. Vet. Med. Assoc. 180, 239–250. Melo E Silva, S.R.A., Vulcano, L.C., 2002. Collateral cartilage ossification of the distal phalanx in the Brasilian jumper horse. Vet. Radiol. Ultrasound 43, 461–463. Oliver, L.J., Baird, D.K., Baird, A.N., Moore, G.E., 2008. Prevalence and distribution of radiographically evident lesions on repository films in the hock and stifle joints of yearling Thoroughbred horses in New Zealand. N.Z. Vet. J. 56, 202–209. Oxspring, G.E., 1935. The radiology of navicular disease, with observation on its pathology. Vet. Rec. 15, 1433–1447. Ruohoniemi, M., Tulamo, R.M., Hackzell, M., 1993. Radiographic evaluation of ossification of the collateral cartilages of the third phalanx in Finnhorses. Equine Vet. J. 25, 453–455. Ruohoniemi, M., Laukkanen, H., Ojala, M., Kangasniemi, A., Tulamo, R.M., 1997. Effects of sex and age on the ossification of the collateral

Please cite this article as: Hilla, D., Distl, O., Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses. Livestock Science (2014), http://dx.doi.org/10.1016/j.livsci.2014.09.015i

8

D. Hilla, O. Distl / Livestock Science ] (]]]]) ]]]–]]]

cartilages of the distal phalanx of the Finnhorse and the relationships between ossification and body size and type of horse. Res. Vet. Sci. 62, 34–38. SAS, 2013. Statistical Analysis System. Cary, NC, USA. Ruohoniemi, M., Ahtiainen, H., Ojala, M., 2003. Estimates of heritability for ossification of the cartilages of the front feet in the Finnhorse. Equine Vet. J. 35, 55–59. Sandgren, B., Dalin, G., Carlsten, J., 1993. Osteochondrosis in the tarsocrural joint and osteochondral fragments im the fetlock joints in Standardbred trotters I. Epidemiology. Equine Vet. J. 16, 31–37. Stock, K.F., Distl, O., 2006a. Genetic correlations between osseous fragments in fetlock and hock joints, deforming arthropathy in hock joints and pathologic changes in the navicular bones of Warmblood riding horses. Livest. Sci. 105, 35–43. Stock, K.F., Distl, O., 2006b. Correlations between sport performance and different radiographic findings in the limbs of Hanoverian Warmblood horses. Anim. Sci. 82, 83–93. Stock, K.F., Distl, O., 2006c. Genetic analyses of the radiographic appearance of the distal sesamoid bones in Hanoverian Warmblood horses. Am. J. Vet. Res. 67, 1013–1019. Stock, K.F., Hamann, H., Distl, O., 2004a. Variance component estimation on the frequency of deforming arthropathies in limb joints of Hanoverian Warmblood horses. J. Anim. Breed. Genet. 121, 269–288. Stock, K.F., Hamann, H., Distl, O., 2004b. Variance component estimation on the frequency of pathologic changes in the navicular bones of Hanoverian Warmblood horses. J. Anim. Breed. Genet. 121, 289–301. Stock, K.F., Hamann, H., Distl, O., 2005. Estimation of genetic parameters for the prevalence of osseous fragments in limb joints of Hanoverian Warmblood horses. J. Anim. Breed. Genet. 122, 271–280.

Stock, K.F., Hoeschele, I., Distl, O., 2007. Estimation of genetic parameters and prediction of breeding values for multivariate threshold and continuous data in a simulated horse population using Gibbs sampling and residual maximum likelihood. J. Anim. Breed. Genet. 124, 308–319. Ueltschi, G., 2001. Zur Röntgendiagnostik des Strahlbeines. Pferdeheilkunde 18, 217–224. van Hoogmoed, L.M., Snyder, J.R., Thomas, H.L., Harmon, F.A., 2003. Retrospective evaluation of equine prepurchase examinations performed 1991-2000. Equine Vet. J. 35, 375–381. van Grevenhof, E.M., Ducro, B.J., Van Weeren, P.R., Van Tartwijk, J.M., Van den Belt, A.J., Bijma, P., 2009. Prevalence of various radiographic manifestations of osteochondrosis and their correlations between and within joints in Dutch warmblood horses. Equine Vet. J. 41, 11–16. van Weeren, P.R., 2006. Etiology, diagnosis, and treatment of OC(D). Clin. Tech. Equine Pract. 5, 248–258. Verschooten, F., VanWaerebeek, B., Verbeeck, J., 1996. The ossification of cartilages of the distal phalanx in the horse: an anatomical, experimental, radiographic and clinical study. J. Equine Vet. Sci. 16, 291–305. Vinson, W.E., White, J.M., Kliewer, R.H., 1976. Overall classification as a selection criterion for improving categorically scored components of type in holsteins. J. Dairy Sci. 59, 2104–2114. Willms, F., Röhe, R., Kalm, E., 1999. Genetic analysis of different traits in horse breeding by considering radiographic findings—1st communication: Importance of radiographic findings in breeding sport horses. Züchtungskunde 71, 330–345.

Please cite this article as: Hilla, D., Distl, O., Genetic parameters for osteoarthrosis, radiographic changes of the navicular bone and sidebone, and their correlation with osteochondrosis and osteochondral fragments in Hanoverian warmblood horses. Livestock Science (2014), http://dx.doi.org/10.1016/j.livsci.2014.09.015i