Pathological and clinical features associated with palmar/plantar osteochondral disease of the metacarpo/metatarsophalangeal joint in Thoroughbred racehorses

Authors


Summary

Reasons for performing study

Palmar/plantar osteochondral disease (POD) of the metacarpo/metatarsophalangeal (Mc/MtPh) joints is common in Thoroughbred racehorses, but there is little information on predilection sites, associated pathologies and clinical signs.

Objectives

To describe the distribution of POD at post mortem examination and to examine the association of POD with other pathologies and clinical histories.

Methods

The Mc/MtPh joints of 102 Thoroughbred racehorses were examined post mortem. Gross pathology of associated osteochondral elements, suspensory ligament and flexor tendons was recorded. Clinical data were extracted from computerised records. Relationships between all Mc/MtPh joint pathologies, clinical data and grade of POD in each condyle were determined using multilevel, ordinal logistic regression models.

Results

Palmar/plantar osteochondral disease was recorded (grade 1–3 in at least one condyle) in 80.4% (95% confidence interval 71–88) of horses. The grade of POD was significantly higher in the forelimb, where the medial condyle had higher grades, than in the hindlimb, where the POD grade was higher in the lateral condyle. There was a significant positive relationship between POD grade and multiple other joint pathologies, whereas grades of linear fissures had a negative relationship with grade of POD. There was no relationship between grade of POD and tendon and suspensory ligament gross pathology, although horses with a history of tendon injury had significantly lower POD grades. Horses that had a clinical history of a Mc/MtPh joint problem had higher POD grades. In addition, horses that had been retired from racing for more than 2 months at the date of death had significantly lower POD grades.

Conclusions and potential relevance

Palmar/plantar osteochondral disease has specific predilection sites. Horses with a clinical history of Mc/MtPh-associated lameness had higher POD grades, whereas those with a history of superficial digital flexor tendon injury and retired horses had lower grades. The progression of POD and its ability to heal requires further evaluation.

Introduction

Injuries involving the metacarpophalangeal and metatarsophalangeal (Mc/MtPh) joints are common causes of lameness and days lost from training in the Thoroughbred racehorse [1-3].

Palmar/plantar osteochondral disease (POD) of the distal condyles of the third metacarpal and metatarsal (Mc/MtIII) bones [4], previously termed traumatic ostechondrosis [5], is a common injury of Mc/MtPh joints. The condition is believed to be associated with material fatigue of subchondral bone following repetitive overload on the distal Mc/MtIII during high-speed locomotion [6]. Pain associated with disease of subchondral bone of the distal condyles of Mc/MtIII is recognised as a common syndrome in most racing breeds [7, 8]; however, definitive diagnosis of all but severe POD lesions is difficult to achieve without specialised imaging modalities, such as magnetic resonance imaging (MRI) [9].

Trope et al. [10] demonstrated that increased radiopharmaceutical uptake (RDU) of the palmar/plantar Mc/MtIII condyles is the most common sign on scintigraphic examination of the Mc/MtPh region in Thoroughbred racehorses. Furthermore, horses with moderate to marked RDU of the palmar or plantar condyles had fewer starts and earned less prize money per start than horses without these findings [10], demonstrating the performance-limiting effects of these lesions. Powell [9] found that 54.9% of 131 horses with Mc/MtPh joint pain had POD demonstrable by low-field standing MRI.

In a study of 64 Thoroughbred racehorses, POD was identified in 67% at post mortem [4], and associations with other pathologies affecting the Mc/MtIII condyles (wear lines, cartilage ulceration, dorsal impact injuries and linear fissures) were demonstrated. However, that study did not show any significant predilection sites between fore and hind or lateral and medial condyles.

The present study builds on our earlier observations [4]. The aims were to describe the prevalence, distribution and epidemiology of POD lesions at post mortem in a further, larger population of flat-racing Thoroughbreds. A principal objective was to evaluate concurrent gross pathology of the articular surfaces of the proximal sesamoid bones and distal condyles of Mc/MtIII and of the flexor tendons and the suspensory ligament (SL) and to determine associations of pathology recorded with POD. In addition, the relationships of POD at post mortem to previous history of lameness, Mc/MtPh joint disease, flexor tendon or SL injury and intra-articular corticosteroid medication were assessed. Our hypotheses were as follows: 1) that particular limbs and condyles would show predilection for POD; 2) that POD lesions would be associated with concurrent SL injury; and 3) that POD grades post mortem would be associated with a history of lameness and frequent medication of the affected joint with corticosteroids.

Materials and methods

All horseracing in Hong Kong is under the direct control of the Hong Kong Jockey Club. Racing is conducted at 2 racecourses, both of which are on the flat. Approximately 88% of races are on turf, while the remainder are conducted on a dirt track (sand, silt and organic fibre). The vast majority of training is on the same dirt surface. Horses are all imported, predominantly as 3-year-old males. Approximately 50% of imports are previously unraced, and those which have raced previously have to meet minimal performance criteria for acceptance. Horses are compulsorily retired when they reach 10 years of age, but premature retirement may occur at any time during their racing career, and the mean retirement age is 5.7 years. Following retirement, suitable horses are transferred for use in riding schools but are still under the treatment/care of veterinary surgeons of the Hong Kong Jockey Club. All clinical care is provided by the Department of Veterinary Clinical Services, Hong Kong Jockey Club.

Material was examined from 102 Thoroughbred racehorses that were either in active race training up to the time of death or had previously retired from active race training at the Hong Kong Jockey Club between 24 November 2005 and 25 March 2009. This was an entirely separate cohort of horses to those studied in our previous report [4]. Of these 102 horses, one had never raced in Hong Kong and 16 had raced but had been retired for more than 2 months at the time of euthanasia.

The reasons for death/euthanasia included the following: 20 due to a fracture associated with the Mc/MtPh joint (including proximal sesamoid bone(s) [n = 10], proximal phalanx [n = 1], distal condyles of Mc/MtIII [n = 6] or multiple sites [n = 3]); 9 due to fractures at other sites (carpal [n = 3], pelvis [n = 2] and one each of humerus, scapula, distal phalanx and proximal McIII); 20 due to advanced Mc/MtPh joint disease (e.g. osteoarthritis); 16 due to severe tendon/ligament injury (superficial/deep digital flexor tendon [n = 7] and suspensory ligament [n = 9]); 20 due to other lameness conditions; and 17 (including the one horse that never raced) were subjected to euthanasia for clinical reasons unrelated to orthopaedic disease.

Metacarpo/metatarsophalangeal joints were examined by gross observation at post mortem examination performed immediately after death, predominantly by one observer (C.M.R.). The observer did not refresh his knowledge of history of training and racing records prior to post mortem examination. The scoring system used by Barr et al. [4] was modified and extended, and each condyle and associated structures were assigned a score (Table 1).

Table 1. Scoring system of pathology of distal condyles of the third metacarpal/metarsal bone (Mc/MtIII) and associated structures
FeatureScoreDescription
Palmar/plantar osteochondral disease0Normal
1Discolouration (bruising) of subchondral bone only; no or minimal disruption of overlying articular cartilage
2Discolouration (bruising), with mild to moderate disruption of articular cartilage
3Established palmar/plantar osteochondral disease lesions. Discolouration and disruption/collapse of articular surface
Wear lines (distal condyles, proximal sesamoid bones and proximal phalanx)0Normal, i.e. wear lines absent
1Partial-thickness wear lines in cartilage
2Full-thickness wear lines in cartilage
Cartilage loss, associated with transverse ridge, condylar groove and central condylar region0Normal
1Partial-thickness cartilage loss (fibrillation)
2Full-thickness cartilage loss (ulceration)
3Extensive full-thickness cartilage loss, with exposure of the subchondral bone
Linear fissures (within condylar grooves)0Normal
1Faint groove, with intact cartilage visible along length
2Well-defined groove, with partial thickness split in cartilage
3Well-defined groove, with full-thickness split in cartilage
Marginal remodelling of proximal sesamoid bones0Normal/marginal remodelling absent
1Marginal remodelling present
Dorsal impact injuries (dorsoproximal aspect of proximal phalanx against dorsodistal aspect of Mc/MtIII)0Normal
1Mild (thickening/inflammation of synovial pad and mild scalloping of underlying Mc/MtIII; mild erosion of dorsoproximal proximal phalanx)
2Severe (thickened, inflamed synovial pad and marked scalloping of Mc/MtIII; full-thickness cartilage loss and/or chip fracture of dorsoproximal proximal phalanx)
Cartilage loss associated with proximal sesamoid bones0Normal
1Partial-thickness cartilage loss (fibrillation)
2Full-thickness cartilage loss (ulceration)
3Extensive full-thickness cartilage loss, with exposure of the subchondral bone
Tendon and ligament in gross sample0Normal
1Inflammation
2Grossly enlarged
3Ruptured

Detailed clinical data for the 102 horses were extracted from the computerised records held by the Hong Kong Jockey Club and imported into an access database. Four separate databases were used. The clinical database holds all details of all examinations, including those resulting from ultrasound, radiographic or scintigraphic studies. The official veterinary examination database contains findings from all inspections conducted by veterinary officials in the Department of Veterinary Regulation and International Liaison, Hong Kong Jockey Club, relating to follow-up of significant findings reported by the Department of Veterinary Clinical Services and incidents during or after racing. The ‘racecheck’ database consists of subjective observations made by veterinarians of the Department of Veterinary Regulation and International Liaison during the prerace inspections that are compulsory for all horses on the day before racing. These observations included notes on lameness, joint effusion, tendon and ligament enlargement and subjective gait abnormalities, such as ‘choppy,’ ‘wide’ and ‘plaiting’ (placing a foot directly in front of the other). Horses were coded as ‘positive’ if these observations were made at any time during the horse's racing career in Hong Kong. Fetlock abnormalities were grouped together and included findings such as ‘effusions,’ ‘fetlock wear and tear,’ swelling, joint enlargement, ‘fetlock osteoarthritis’ and lateral suspensory branch thickening. The drug treatment database records all drug administrations, and data on the administration of intra-articular corticosteroids were extracted from this. These databases began in 2003, and horses that raced prior to 2003 therefore did not have complete clinical data available and were excluded from the present analyses.

Data analysis

Databases were linked using Microsoft Access 2007a and data extracted and exported to a spreadsheet (Microsoft Excel 2007)a. Descriptive statistics were generated for the frequency of pathologies and clinical findings. The association of grades of pathologies at the condylar level with the ordinal outcome of POD grade (0–3; grade 0 as the reference category) and of clinical data and POD was performed using a multilevel, ordered proportional odds model with a logit link function [11], with horse included as a random effect to allow for the clustering of outcomes at the level of the individual condyle within horse. The assumption of proportional odds was assessed for all variables in the final model by formally testing (using the Wald test) whether the coefficients fitted individually for each response category were significantly different from each other. Models were fitted using penalised quasi-likelihood with 2nd order Taylor series expansion [11].

Data were analysed using Minitab® 15 statistical softwareb and multilevel models used the MLwiN statistical software packagec.

Results

A total of 805 condyles were examined, and POD was recorded (grade 1–3 in at least one condyle) in 82 of 102 horses, a prevalence by individual of 80.4% (95% confidence interval [CI] 71–88). Table 2 shows the summary numbers of condyles affected by each grade of pathology. It was not possible to score accurately all features of all condyles on account of extensive pathology associated with acute trauma to the joint, and data are therefore missing for these condyles. In addition, the tendons and suspensory ligament were not scored in a proportion of horses owing to logistical constraints.

Table 2. Summary of numbers (and percentages) of structures affected by each grade of the different pathologies scored in 805 condyles in 102 horses
PathologyNumber in each grade (% of the total row)
Grade 0Grade 1Grade 2Grade 3Total
Palmar osteochondral disease447 (55.5%)246 (30.6%)81 (10.1%)31 (3.9%)805
Wear lines423 (52.8%)322 (40.2%)56 (7.0%)-801
Cartilage loss, condyles431 (54.0%)308 (38.6%)45 (5.6%)14 (1.8%)798
Linear fissures446 (55.7%)317 (39.6%)37 (4.6%)1 (0.1%)801
Dorsal impact injury369 (46.2%)292 (36.6%)137 (17.2%)-798
Marginal remodelling578 (72.0%)225 (28%)--803
Cartilage loss, proximal sesamoid bones744 (93.7%)27 (3.4%)12 (1.5%)11 (1.4%)794
Superficial digital flexor tendon364 (97.1%)2 (0.5%)7 (1.9%)2 (0.5%)375
Deep digital flexor tendon373 (99.5%)0 (0%)1 (0.3%)1 (0.3%)375
Suspensory ligament718 (95.7%)16 (2.1%)10 (1.3%)6 (0.8%)750

One horse had never raced, and 4 horses had raced prior to 2003, and so did not have complete clinical data available. In total, 16 horses had retired 2 months or more prior to the date of death. Of the 97 horses with clinical data, 52 (54%) had a record of one or more Mc/MtPh joint problems in life, including 31 with osteoarthritis, 21 with a fracture, 5 with sesamoiditis and 11 with osteochondral fragmentation. A total of 34 (35%) horses also had a history of lameness due to a Mc/MtPh joint problem identified from official veterinary records, and 34 horses had a Mc/MtPh joint problem/lameness identified from race check records, of which 29 and 21, respectively, also had a record of one or more Mc/MtPh joint problems in the clinical database.

A total of 72 (74%) horses had received at least one treatment of intra-articular corticosteroids in their lifetime (median 7 treatments, maximum 53), and this included a Mc/MtPh joint in 56 horses (median 2 treatments, maximum 34). In the 12 months prior to euthanasia, 64 horses (66%) had received at least one treatment of intra-articular corticosteroids, and 51 of these were in a Mc/MtPh joint.

Intercept-only models showed that there was significant clustering of POD pathology within horse (variance estimate 3.43 [s.e. 0.6]), indicating that horse accounted for approximately 51.1% of clustering of POD outcomes.

Multilevel ordinal regression analysis showed that the grade of POD was significantly higher in the forelimb compared with the hindlimb (Table 3); however, there was a significant interaction between fore and hind and medial and lateral such that in the forelimb there were higher POD grades medially, whereas in the hindlimb the grade of POD was higher in the lateral condyle (see also Table S1). There appeared to be no difference between left and right forelimbs, and this was consistent even after analysing fore- and hindlimbs separately.

Table 3. Relationships of pathologies of the metacarpal/metatarsal joints and palmar osteochondral disease (POD) grade
VariableOdds ratio

Lower

95% confidence interval

Upper

95% confidence interval

P value
  1. Results are from univariable multilevel, ordered proportional odds models, allowing for within-horse clustering (n = 776 because any with missing pathology data were excluded). Outcome is POD grade 0–3. Palmar osteochondral disease grade 0 was treated as the reference category; therefore, odds ratios less than one indicate that the probability of being in the lower categories is increased and odds ratios greater than one indicate that the probability of being in the higher categories is increased. #There was a significant interaction between lateral/medial and fore/hindlimb, reflecting differences between forelimbs and hindlimbs (see Table S1). Odds ratios including interaction term: fore lateral odds ratio = 1.0, fore medial odds ratio = 2.5, hind lateral odds ratio = 0.9, hind medial odds ratio = 0.5. For brevity, the separate intercept for each grade of POD and the within-horse variance estimate for each analysis are not shown.
LimbLeftReference   
 right0.870.641.180.4
LimbForeReference   
 Hind0.450.330.62<0.001
LimbLateralReference   
 Medial1.280.941.740.1
Fore/hind vs. lateral/medial interaction#    
Hind0.900.581.400.60
Medial2.501.623.85<0.001
Fore/hind vs. lateral/medial interaction0.250.130.46<0.001
Marginal remodelling    
NoReference   
Yes2.641.763.96<0.001
Wear lines    
Grade 0Reference   
Grade 12.401.613.57<0.001
Grade 26.833.1614.77<0.001
Cartilage loss, condyles    
Grade 0Reference   
Grade 15.623.638.72<0.001
Grades 2 and 3 combined22.4211.1445.14<0.001
Linear fissure    
Grade 0Reference   
Grade 10.440.300.64<0.001
Grades 2 and 3 combined0.540.231.270.20
Dorsal impact injury    
Grade 0Reference   
Grade 11.731.132.640.01
Grade 23.892.306.59<0.001
Cartilage loss, sesamoids    
Grade 0Reference   
Grade 12.120.825.470.10
Grades 2 and 3 combined71.4524.70206.71<0.001

There was a significant positive relationship between grade of POD and the following pathologies (i.e. as the grade of pathology increased the probability of higher grades of POD increased; Table 3): 1) grade of cartilage loss on the condyles; 2) grade of dorsal impact injuries; 3) grade of cartilage loss on the proximal sesamoid bones; and 4) grade of wear lines. Horses with signs of marginal remodelling were also more likely to have higher grades of POD (Table 3).

Grade of linear fissures had a significant negative relationship with grade of POD (i.e. as the grade of pathology decreased, the probability of higher grades of POD increased). There was no significant relationship between grade of POD and gross tendon or suspensory ligament lesions, although these injuries were not commonly identified.

Univariable analysis of clinical data (Table S2) showed that horses with a clinical record of an Mc/MtPh joint problem, either in the last year (odds ratio [OR] 3.7, 95% CI 1.7–7.9) or in their lifetime (OR 2.9, CI 1.3–6.3), had an increased probability of having higher POD grades. Horses that had a greater number of doses of intra-articular triamcinolone in the last 12 months, either in any joint (OR 1.1 [per additional dose], 95% CI 1.0–1.1) or in the Mc/MtPh joints (OR 1.1 [per additional dose], 95% CI 1.0, 1.2), also had an increased probability of higher POD grades. Horses with a history of superficial digital flexor tendon problems (OR 0.2, 95% CI 0.1–0.6) and horses that were retired more than 2 months before the date of death (and therefore the date of POD scoring; OR 0.2, 95% CI 0.1–0.6) had a decreased probability of higher POD grades. However, there was no significant association between history of suspensory ligament injury and POD grade. Horses with a record of abnormal gaits on prerace inspection in the hindlimbs, such as ‘choppy behind’ (OR 3.0, 95% CI 1.4–6.6) and ‘plaiting’ (OR 2.2, 95% CI 1.0–4.8), had an increased probability of higher POD grades. There was no significant association with the observation of lameness or fetlock abnormalities on prerace examination and POD grade.

All the clinical data satisfied the assumption of proportional odds. However, 3 of the pathology variables did not. Cartilage loss in condyles (grade 2/3), wear lines (grade 2) and cartilage loss in proximal sesamoid bones (grade 2/3) all demonstrated a lesser effect across POD categories POD 0 to POD 1 and greater effects across the remaining (i.e. suggesting even higher probability of POD grades 2 or 3 if higher grades of these pathologies were also present). However, in all cases the numbers with these higher grades of pathologies were low, so that results from proportional odds models are presented.

Discussion

Palmar/plantar osteochondral disease was very common in this sample of animals; 80% of all horses had a POD lesion in at least one condyle, and 14% of identified lesions were severe, graded 2 or 3. However, the sample was potentially biased because it was comprised of horses that were suffering from conditions, many of which involved musculoskeletal disease (either acute or chronic), which warranted euthanasia. Therefore, it is not unreasonable to speculate that POD is over-represented in this group and that the prevalence reported here is likely to be an overestimate for the population as a whole.

Palmar osteochondral disease lesions were significantly more severe in forelimbs compared with hindlimbs. This is at odds with the findings from our previous, smaller study on a different set of horses from the same population conducted at a different time point [4], but in agreement with observations reported by others [3, 5]. Furthermore, the grade of POD was most severe in the medial condyle in the forelimb and lateral condyle in the hindlimb, which concurs with the distribution of increased RDU in the Mc/MtPh region demonstrated by Trope et al. [10]. There was no significant difference in prevalence or severity of POD between left and right limbs, which is also in agreement with previous reports into common conditions of the Mc/MtPh joints [3, 4, 12].

There was a significant positive relationship between POD grade and severity of cartilage loss on the condyles, wear lines and grade of dorsal impact injuries, which substantiates our previous findings [4]. There was also a positive relationship between POD and grade of cartilage loss on the proximal sesamoid bones and signs of marginal remodelling on the base of the proximal sesamoid bones. While grossly detectable marginal remodelling of the condyles is relatively rare, as reported previously [4], it is much more common at the base and apex of the proximal sesamoid bones and is a well-recognised radiographic feature of osteoarthritis of the Mc/MtPh joint. The scoring system used in the present study was therefore modified to record modelling at the base of the proximal sesamoid bones. Correlation between POD and marginal new bone at the base of the proximal sesamoid bones supports the hypothesis that POD is associated with progression of generalised joint disease.

Grade of linear fissures had a significant negative relationship with grade of POD, which is in agreement with other studies [4, 13]. This may reflect different patterns of loading in individual animals, which predispose to one or the other pathology, although further work is required to elucidate this relationship more thoroughly. Evidence of suspensory ligament and superficial digital flexor injury were uncommon among the horses examined and, therefore, lack of significant association of these injuries with POD is considered an unreliable finding.

Although the clinical effect of different grades of POD on performance has yet to be established, our findings demonstrate that horses with higher grades of POD were more likely to have been presented for an Mc/MtPh joint problem whilst in training than those without POD. This was a consistent observation across grades of POD and was frequently associated with a history of diagnostic investigations, including imaging studies. Horses with relatively increased RDU of the condyles of both forelimbs and hindlimbs have been shown to have fewer starts, reduced earnings and reduced earnings per start compared with control horses [10]. In addition, horses with RDU of the palmar/plantar condyles are more likely to return to racing later than control animals following a rest period [10]. Given that POD is likely to be associated with RDU, it is reasonable to combine these observations and to speculate that POD has a negative effect on performance. Others have demonstrated that lameness in racehorses due to injury of the Mc/MtPh joint, other than fracture, contributes significantly to days lost from training [2, 14], and Bailey et al. [14] showed that the recurrence of Mc/MtPh joint problems was high. It is conceivable that POD, at varying stages/grades, is at the root of a high proportion of these degenerative Mc/MtPh joint pathologies in the Thoroughbred racehorse.

Horses with higher POD grades were significantly more likely to have undergone intra-articular medication with triamcinolone acetonide (corticosteroid) while in training than horses without POD. Horses that had received more doses of intra-articular triamcinolone acetonide in the Mc/MtPh joints in the last 12 months were more likely to have higher POD grades. Triamcinolone acetonide is a commonly administered intra-articular therapy for joint disease in racehorses [15]. Overuse of intra-articular corticosteroids has been incriminated as a potential cause of joint disease [16]. It is not possible to determine cause or effect from a cross-sectional study such as this; while it could be postulated that use of corticosteroid may lead to advancement of POD, it is more likely that this relationship was due to higher use of triamcinolone acetonide in the affected population, because these horses had greater lameness or other clinical signs that warranted this therapy.

While it seems reasonable to propose that POD is a progressive condition and that the different grades represent stages in a continuum, this is as yet unproven. Also, it is unknown whether POD lesions can resolve and, if so, if there is a threshold beyond which repair is impossible. Powell [9] reported that, given an extended rest period, a number of horses with MRI-confirmed POD subsequently trained and raced successfully. In the present study, POD grades were lower both in horses that had been retired from racing for at least 2 months and also in horses that had a history of superficial digital flexor tendonitis. Horses with superficial digital flexor tendon lesions may have had extended periods of rest or may have been in training at lower intensities, and this may be a proxy measure for time off training. Alternatively, the biomechanical aetiopathogenesis of superficial digital flexor tendon lesions may differ from those of POD lesions [17]. It is not yet known whether these lesions can heal, and prospective studies, using imaging techniques where POD can be identified and graded early, are needed. A recent study from the USA reported an excellent prognosis for return to successful racing in horses with evidence of pain localised to the condyles and associated with RDU in the absence of radiographic findings (which is likely to be due to early POD lesions) given 60 days of paddock rest [18].

Fitting a model with an ordered categorical response outcome based on cumulative response probabilities, rather than fitting a response for each separate category of POD, takes advantage of the ordering in the categories to simplify the model structure [11, 19]. However, inherent in this model is the proportional odds assumption, which states that the odds ratios of the explanatory variables are constant across response categories. There was some evidence that cartilage loss in condyles (grade 2/3), wear lines (grade 2) and cartilage loss on proximal sesamoid bones (grade 2/3) did not entirely satisfy this assumption. In all cases, however, the probabilities of these grades of pathology were even higher in POD grades 2 or 3. In all cases, numbers with these higher grades of pathologies were low, so that results from proportional odds models are presented.

The multilevel model showed that there was significant clustering of POD pathology within horse, suggesting that some horses are more likely to have multiple condyles affected with higher grades and that multilevel models are needed to obtain estimates of associations accurately. The fact that individual horses are more likely to have multiple condyles affected suggests some predisposition to POD in that individual or that all limbs are subjected to the same environmental influences that contribute to repetitive, overuse injuries. Potential intrinsic risk factors include variation in susceptibility of tissues to repetitive load and variation in loading due to morphology of the distal condyles of Mc/MtIII, conformation of the limb and fine musculoskeletal coordination during fast exercise.

In conclusion, this study identifies that POD is a common finding at post mortem in the deceased racing Thoroughbred and that it occurs more frequently in the forelimb, with a predilection for the medial condyle of the forelimb and the lateral condyle of the hindlimb. There is an association between severity of POD and other pathologies associated with generalised osteoarthritis and with a history of fetlock lameness. Grade of POD may decrease following a period of rest, suggesting that healing is possible. Further study is now required to understand the pathogenesis and progression of this disorder and to identify factors associated with its occurrence.

Authors' declaration of interests

No competing interests have been declared.

Source of funding

This project was funded by the Horserace Betting Levy Board and the Hong Kong Jockey Club.

Acknowledgements

We are grateful to veterinarians within the Department of Veterinary Regulation and International Liaison, Hong Kong Jockey Club, in particular Dr Brian Stewart and Dr Kenneth Lam, for granting us access to their databases for part of this study. We are also grateful for the assistance of colleagues in the Department of Veterinary Clinical Services, Hong Kong Jockey Club, particularly Iris Yu for assistance with data extraction.

Authorship

The study was conceived by C.M.R., P.D.C and A.B. All authors contributed to the design of the study. C.M.R. collected all data for the study, and all statistical analyses were performed by G.L.P. G.L.P. drafted the manuscript, and all authors contributed to the final version.

Manufacturers' addresses

  1. aMicrosoft Corporation, Redmond, Washington, USA.

  2. bMinitab Inc. 2010, State College, Pennsylvania, USA.

  3. cMLwiN version 2.18, Centre for Multilevel Modelling, University of Bristol, Bristol, UK.

Ancillary