Risk Factors for Injuries in Thoroughbred Racehorses. RIRDCInnovation for rural Australia

Transcription

1 Risk Factors for Injuries in Thoroughbred Racehorses RIRDCInnovation for rural Australia

2 Risk Factors for Injuries in Thoroughbred Racehorses A report for the Rural Industries Research and Development Corporation by Naomi Cogger, David Evans, Nigel Perkins, David Hodgson and Stuart Reid June 2006 RIRDC Publication No 06/050 RIRDC Project No US-129A

3 2006 Rural Industries Research and Development Corporation. All rights reserved. ISBN ISSN Risk Factors for Injuries in Thoroughbred Racehorses Publication No. 06/050 Project No. US-129A The information contained in this publication is intended for general use to assist public knowledge and discussion and to help improve the development of sustainable industries. The information should not be relied upon for the purpose of a particular matter. Specialist and/or appropriate legal advice should be obtained before any action or decision is taken on the basis of any material in this document. The Commonwealth of Australia, Rural Industries Research and Development Corporation, the authors or contributors do not assume liability of any kind whatsoever resulting from any person's use or reliance upon the content of this document. This publication is copyright. However, RIRDC encourages wide dissemination of its research, providing the Corporation is clearly acknowledged. For any other enquiries concerning reproduction, contact the Publications Manager on phone Researcher Contact Details Naomi Cogger EpiCentre, Massey University, Palmerston North, New Zealand Phone: Fax: N.Cogger@massey.ac.nz In submitting this report, the researcher has agreed to RIRDC publishing this material in its edited form. RIRDC Contact Details Rural Industries Research and Development Corporation Level 2, 15 National Circuit BARTON ACT 2600 PO Box 4776 KINGSTON ACT 2604 Phone: Fax: rirdc@rirdc.gov.au. Web: Published in June 2006 Printed by Union Offset ii

4 Foreword Lameness, or Musculoskeletal (MS) injuries, are the most common health problem in Thoroughbred racehorses and represent an economic cost to industry in terms of treatment cost and lost opportunity to race. In addition the high number of MS injuries in racehorses, particularly in two-year-olds, has also raised welfare concerns. In May 2000, a 27-month RIRDC funded study investigating risk factors for MS injuries commenced. The study convenience sampled 14 trainers with facilities at metropolitan and provincial racetracks in New South Wales, Australia. In the 2000/01 and 2001/02 racing season 321 and 128 two-year-olds, respectively, were enrolled in the study. During the study period 428 MS injuries were recorded in 248 horses. The most common site of MS injury was the third metacarpal bone, and the most common injury type at this site was shin soreness. Overall, the incidence rate for all MS injuries was higher in two-year-olds than in three-year-olds. In addition the first start was significantly increased in horses that sustained an MS injury. The median time to recovery was 5.5 months. Multivariate statistical models were used to explore risk factors for MS injuries. The results suggest that MS injuries involving structures in the lower forelimb (carpus to fetlock inclusive) could be reduced by limiting exposure to high-speed exercise, in particular gallops at fast speed ( 890 m/minute). This supports the proposition that training injuries are caused by the accumulation of micro damage. The results also suggest there are a number of other factors that vary at the trainer level that may be risk factors for injuries, in particular joint injuries. These include unmeasured variables such as the rate of increase in distance galloped at fast-speed, conformation of the horse, skill of the riders, and farrier and veterinary involvement. This project was funded from industry revenue which is matched by funds provided by the Australian Government. This report is an addition to RIRDC s diverse range of over 1500 research publications. It forms part of our Horse R&D sub-program which aims to research common problems of horses in training. Most of our publications are available for viewing, downloading or purchasing online through our website: downloads at purchases at Peter O Brien Managing Director Rural Industries Research and Development Corporation iii

5 Acknowledgments Funding for this project was provided by the Rural Industries Research and Development Corporation and NSW Racing Research Fund. AAPT provided free access to their commercial database of racing and barrier trial results for the duration of the study. The authors would also like to say a very sincere thank you to the participating trainers and their staff. There were moments when the data collection was an intrusion in their busy work schedules, but they always found time to complete the questionnaires. The dedication and commitment of these people made the data collection process an enjoyable experience and were a major contributor to the success of the study. Thank you also to the veterinarians and racing officials who provided assistance through the study and Sebastian Werner for design of the study database. Abbreviations > Greater than Greater than or equal to < Less than Less than or equal to 95% CI 95% Confidence interval HS High-speed IR Incidence rate IRR Incidence rate ratio MS Musculoskeletal iv

6 Contents Foreword... iii Acknowledgments...iv Abbreviations...iv Executive Summary...vi What is known about musculoskeletal injuries?...1 Introduction...1 Frequency and impact of MS injuries...1 Risk factors for MS injuries...3 Study design and management...11 Introduction...11 Recruitment and enrolment of study population...11 Data collection...13 Conclusion...15 Incidence and impact of musculoskeletal injuries...16 Introduction...16 Classification of MS injuries...16 Frequency of MS injuries...16 Impact of MS injury...20 Conclusion...22 Profiles of training preparations and spell periods...23 Introduction...23 Overview of analytical methods...23 Factors influencing profiles...23 Conclusion...28 Risk factors for musculoskeletal injuries when horses are first exposed to high-speed exercise..29 Introduction...29 Risk factors for MS injuries...29 Conclusion...30 Risk factors for incident cases of shin soreness and lower forelimb joint injuries...31 Introduction...31 Overview of analytical methods...31 Risk factors...32 Conclusion...33 General discussion...34 Frequency of MS injuries...34 Impact of MS injuries...34 Risk factors for MS injuries...35 Recommendations for future research...37 Conclusion...37 References...38 v

7 Executive Summary Introduction Musculoskeletal (MS) injuries have been identified as the most common health problem in Thoroughbred racehorses. The injuries represent an economic cost to industry in terms of treatment cost and lost opportunity to race. The high number of MS injuries in racehorses, particularly in twoyear-olds, has also raised welfare concerns. After reviewing the literature it is clear that there is only a limited understanding of the risk factors for MS injuries. In particular, there is almost a complete absence of any investigations examining training related factors for MS injuries. Studies that utilize daily training records are fundamental to the understanding of training related risk factors for MS injuries. An understanding of the risk factors for MS injuries would facilitate the design of strategies that may reduce the occurrence and impact of MS injuries. The aim of this study was to investigate the epidemiology of MS injuries in Australian two- and threeyear-old Thoroughbred racehorses. More specifically the study aimed to: describe the incidence of MS injuries; determine the impact of MS injuries using a variety of measures of performance; and determine risk factors for MS injuries, in particular shin soreness. Study design Fourteen trainers were convenience sampled from those at metropolitan and provincial racetracks in New South Wales, Australia. The study commenced in May 2000 and concluded at the end of the 2001/02 racing season. In the 2000/01 and 2001/02 racing season respectively 321 and 128 two-yearolds were enrolled in the study and followed either until the end of the study or until they were lost to follow-up. A horse was classified as lost to follow-up if it was not in the stable in the last month of the study (July 2002). The investigator visited trainers at approximately fortnightly intervals to enrol eligible horses and collect training and injury data for those horses present at the stable. Information on training methods, starts in races or barrier trials, and training surface were collected using a questionnaire. All injury information was provided by the trainer, and a veterinarian may or may not have been involved in the diagnosis. A problem involving the musculoskeletal system was classified as an injury if it resulted in the horse leaving the stable for more than seven days and was the direct result of training (i.e. injuries such as stone bruises were excluded). Frequency of MS injuries During the study period 428 MS injuries were recorded in 248 horses; 132, 74, 26, 10 and six horses sustained one, two, three, four and five injuries, respectively, while enrolled in the study. The most common site of injury was the third metacarpal bone (n = 184) and the most common injury at the site was shin soreness. This is consistent with the results of previous research by Mason and Bourke (1973), Bailey et al. (1999) and Perkins et al. (2004a). The incidence rates (IR) of the first MS injury in two-year-olds and three-year-olds was 5.57 per 1000 training days and 1.83 per 1000 training days, respectively. When specific types of MS injuries were considered the IR for all categories of injuries, except for tendon injuries, were higher in two-yearolds than three-year-olds. However, only the differences for injuries involving the third metacarpal bone were statistically significant. vi

8 Comparisons of the incidence rates indicated that two-year-olds were 5.1 times more likely than three-year-olds to sustain an injury involving the third metacarpal bone. This is consistent with a study by Perkins et al. (2004a) who reported that the rate of shin soreness was highest in two-year-olds and decreased with increasing age and that the IR for soft tissue injuries were highest in horses more than five years of age. The results of the current study and of Perkins et al. (2004a) would seem to suggest that young horses are at increased risk of shin soreness and older horses are at increased risk of soft tissue injury. The increased risk of soft tissue injuries in older horses may be the result of long term exposure to training rather than an effect of age. As the majority of horses commence training as two-year-olds, the increased risk of shin soreness in this age group could be due to the onset of training. This is supported by anecdotal evidence that older horses commencing training will also develop shin soreness (Buckingham and Jeffcott 1990). Sex was only associated with increased risk of injuries involving the third metacarpal bone. The comparison of IRR indicated that males were 1.6 times more likely to sustain an incident injury involving the third metacarpal bone. These differences between male and female horses may represent a true increased risk in male horses. However, the differences may also be due to differences in exposure to training variables. For example colts may complete more slow speed exercise than fillies. Impact of MS injuries The impact of MS injuries was assessed at both the horse and preparation/spell level. The two horse level outcomes were: (i) time to first start and (ii) time to recovery. These outcomes were chosen as they represent the time taken before owner(s) have an opportunity to receive a return on their investment. The outcomes considered at the preparation level included the duration of the spell period and the preparation. Time to first start The presence of an MS injury increased the time from the commencement of training until the first start. Horses that did not sustain an MS injury had a median of approximately 19 weeks until their first start while the horses that sustained an MS injury did not start for approximately 33 weeks. Therefore, horses that sustained an MS injury were unproductive for 14 more weeks than those that did not sustain an MS injury. After adjusting for injury status the time to first start varied significantly between trainers suggesting that factors other than the presence of MS injury play a role in the horses not starting in races. These may include management decisions to commence training but to delay the commencement of racing for a variety of reasons. It is also possible that time to first start was delayed due to the presence of health problems other than MS injury, such as respiratory disease. Recovery from injury Of the 248 horses that sustained an MS injury, 70% returned to training after their first MS injury. The median time to recovery was 5.5 months. Time to recovery was significantly associated with the intensity of exercise prior to the onset of MS injury, with the time to recovery decreasing as the intensity of exercise prior to injury increased. This may be a true association or it may be biased by a number of factors. For example it may reflect a decision by owners and/or trainers to spell a horse for a longer period if it sustained an MS injury at low exercise intensity as they require time to grow, thereby increasing the duration of recovery. vii

9 Duration of spell period The median duration of spell periods that were associated with MS injury was 11 weeks, while those not associated with an MS injury had a median duration of 10 weeks. Incomplete spell periods were excluded from the analysis. This may have resulted in an underestimation of the duration of spell periods because it was considered more likely that longer spell periods would be incomplete. When interpreting the results it is important to note that spells not associated with an MS injury may have been associated with other health problems such as respiratory disease. Preparation variables For the 70% (n = 173) horses that returned to training after an MS injury the duration of preparations, time to first start, duration of the interval from first start until the end of the preparation and the number of starts per 100 training days were not adversely affected by previous MS injury. This suggests that MS injuries in young horses may not have a long term impact on racing performance. Risk factors for MS injuries Previous exposure to high-speed exercise In the first preparation to involve exposure to high-speed exercise the risk of MS injury involving structures in the lower forelimb increased as the total distance trained at high speed increased. The risk of MS injury in this preparation varied depending on the maximum gallop speed in the preparation. This suggested that the risk associated with cumulative exercise at speeds 15 seconds per furlong, > 15 seconds per furlong and starts in races and barrier trials may differ. The risk of forelimb joint injuries was significantly associated with cumulative distance trained at high-speed. The results showed that compared to horses that were not exposure to high-speed exercise, the risk of joint injury was 4.13 times higher in horses that had accumulated 2000 meters of high-speed exercise in a seven day period. Therefore, limiting the exposure to high-speed exercise should reduce the incidence of joint injuries in two- and three-year-old Thoroughbred racehorses. The hazard of shin soreness was also associated with the cumulative distance trained at high-speed. Compared to horses that had not been exposed to high-speed the hazard of shin soreness was 4.06 (95% CI = ), 7.11 (95% CI = ) and 6.92 (95% CI = ) times higher, respectively, in horses that had accumulated between 1 and 999 metres, between 1000 and 1999 metres and 2000 metres. The increased risk of shin soreness associated with exposure to high-speed exercise and shin soreness is supported by Verheyen et al. (2004). The results of the current study and Verhyen et al. (2005) support a hypothesis that shin soreness and joint injuries are due to cumulative exposure to high-speed exercise that causes accelerated bone remodelling. Once accelerate remodeling has commenced continual loading of the bone will increase the size of the resorptive cavities, resulting in the appearance of micro fractures that extend into the cortex and cause marked reduction in bone strength (Riggs and Evans 1990, Nunamaker 1996). At a microscopic level, the first signs of accelerated remodelling are vascular congestion, thrombosis and resportion of the bone tissue (Brunker et al. 1999, pp 8-9). viii

10 Previous MS injury The results showed that in the first preparation that horses were exposed to high-speed exercise those horses that had previously sustained an MS injury were at increased risk of injury. This is supported by research that found pre-existing subclinical or mild suspensory apparatus injury was a risk factor for all types of MS injuries and suspensory apparatus injury (Hill et al. 2001). Age There were conflicting results regarding the association between age and risk of injury. In the multivariable logistic analysis of risk factors for MS injury involving the forelimb in the first training preparation to include exposure to high-speed exercise, the effect of age was investigated using age in months at the commencement of training, at either the start of the two-year-old season or at the start of the preparation. In the multivariable logistic model there was no significant association between age and the outcome variable. This suggests that delaying the age, in months, at which a horse was first exposed to exercise at speeds greater than or equal to 15 seconds per furlong will not alter the risk of an MS injury. The results were consistent with those reported by Wilson et al. (1996) who found that two-year-olds with their first start early in the two-year-olds season were at no greater risk of MS injury than those horses with their first start at the end of the season. Age at the time of enrolment in the study was identified as a risk factor for joint injuries but not for shin soreness. The results indicated that horses enrolled in the study at 24 months were 46% less likely than horses that were enrolled at < 24 months to sustain a joint injury. Owing to the design of the study enrolment corresponded to the commencement of training. Hence these results suggest that the incidence of joint injuries may be reduced if horses do not commence training until they are 24 months of age. Sex The risk of shin soreness was 1.77 times more in males than females. This may represent a true biological difference. Alternatively, the results could be biased by confounders such as body size and shape or differences in training not included in this model. For example one trainer in the study reported cantering male horses twice around the sand to keep the weight off them. The risk of fetlock joint and carpal joint problems did not differ between male and female horses. Future research The results of the current study and those of Bailey (1999) and Mason and Bourke (1973) have contributed to our understanding of the frequency and impact of MS injuries in Australian two- and three-year-old Thoroughbred racehorses. However, there is a distinct lack of information regarding the epidemiology of MS injuries in older horses. Results from Perkins et al. (2004 a,b) indicated that the type of MS injuries and impact of these injuries vary between age groups. Therefore future research in Australia should describe the incidence and impact of MS injuries in horses more than three years old. The results of this study show that exposure to high-speed exercise is associated with an increased risk of injury. Therefore, future research should focus on identification of ways in which high-speed exercise can safely be introduced into a training period. The results of Verhyen et al. (2005) also indicate that distances trained at canter may be risk factors for MS injuries. Therefore studies should collect information relating to distances trained at slower speeds. Ideally, these future studies should collect this information using less subjective measures than in the current study as the results of studies in New Zealand have suggested that use of subjective gait as a true measure of speed may be problematic (Rogers and Firth 2004). Recent advances in ix

11 microelectronics and GPS technology mean that similar systems could be implemented in other countries with relative ease. Future studies should also consider the role of track-related factors needs to be considered. The first step in such a study would be determine appropriate techniques that can be used to monitor the physical characteristics of racetracks, because such techniques do not presently exist (Stubbs 2004). A longitudinal study could be undertaken to determine variables relating to track performance that, after adjusting for training, are associated with MS injury. When designing studies to investigate safer ways to expose horses to high-speed exercise and trackrelated factors it is important to note that the results of the current study have shown that risk factors for different types of MS injuries do vary. Therefore, analyses need to consider the risk factors for different injuries separately. However, this does not mean that the studies need to examine only one outcome. On the contrary, to rationalise the resources it would be advisable to undertake prospective studies that investigate a number of injuries (Samet and Munoz 1998b, Szklo 1998). It is the authors belief that training information is the most difficult to collect. Thus collecting information on a number of different outcomes creates a more cost effective study. Furthermore, it did not appear that collection of a number of outcomes impacted negatively on the quality of injury information. Conclusion The research presented in this report confirmed previous research that MS injuries are a common problem in two- and three-year-old Thoroughbred racehorses. The results showed that injury prior to the first start had considerable impact on time to first start. However, preparations after the first MS injury do not appear to be adversely affected. The investigation of risk factors showed that increasing distance galloped at high speed increased the risk of MS injury. Future studies need to use more accurate measures of training to examine the relationship between distance galloped and other training related factors, and risk of MS injury. Finally, when considering future research the importance of country, and even region, specific studies can not be emphasised enough as differences in typical training and management strategies make it difficult to generalise the results. x

12 What is known about musculoskeletal injuries? Introduction Musculoskeletal (MS) injuries, or lameness, are the most common health problems in Thoroughbred racehorses (Rossdale 1989, Kobluk et al. 1990a, Lindner and Dingerkus 1993, Bailey et al. 1999, Perkins et al. 2004a, Perkins et al. 2004b). Studies of horses in training have reported that the percentage of horses that sustained an MS injury ranged from 36% (Rossdale 1989) to over 50% (Kobluk et al. 1990a, Lindner and Dingerkus 1993, Bailey 1998). These MS injuries incur both direct and indirect costs. Direct costs included lost opportunity to race and treatment cost. Bailey (1999) reported that 45% of horses did not race during their two-year-old season. Bailey concluded that the principal reason for this was the high number of cases of low grade injuries and other health problems. However, no statistical tests were performed to determine if those horses that sustained an injury were less likely to race than those that did not sustain an injury. Indirect costs associated with MS injuries are difficult to quantify. They include risk of death or injury to the jockey and welfare implications of MS injuries. While all MS injuries are a welfare concern, those that occur during a race are in the full view of race goers and are frequently replayed in television news reports. The injuries are often dramatic and impact negatively on the public perception of racing. The direct and indirect cost to the Thoroughbred racing industry associated with MS injuries makes it desirable to understand factors that affect the risk of MS injuries. Increased understanding of these factors may facilitate the development of strategies to reduce the incidence and/or impact of MS injuries. Frequency and impact of MS injuries There have been a number of studies in different countries that have described the frequency and impact of racing- and training-related injuries involving the musculoskeletal system. The rate of injuries per 1000 racing starts from these studies are summarised in Table 1. There are some differences in rates for fatal and non-fatal racing-related injuries. These differences may be because of regional differences, but may also be due to differences in study design, population and definitions of injuries. The regional differences may be true differences but may also reflect differences in training and management strategies rather than true regional differences. Table 1: Details of studies and incidence rates for all musculoskeletal injuries (MS injury) and fatal MS injury, per 1000 starts in flat racing. Racing MS injuries (per 1000 starts) Source Location Duration of Study Number of tracks Bailey, 1997a Australia Bailey, 1998 Australia Hill et al., 1986 New York, USA Pelso et al., 1994 Kentucky, USA Estberg et al.,, 1996 California, USA NR 1.7 McKee, 1995 UK NR 0.8 Williams et al., 2001 UK NR NR-Not reported Fatal MS injuries (per 1000 starts) 1

13 Racing related MS injuries A two-year longitudinal study examining causes of death in Thoroughbreds and Quarter horses at California racetracks reported 496 fatalities during racing (208), training (195) and non-exercise related activities (93) (Johnson et al. 1994). The majority of fatalities were in Thoroughbreds (432) and MS injuries accounted for approximately 80% of deaths. Similarly, a three-year retrospective study of fatal and non-fatal health problems at British flat and National Hunt racetracks found that 82% involved the musculoskeletal system (Williams et al. 2001). Furthermore, 27% of all problems resulted in death or euthanasia. Several studies have reported that the majority of racing-related injuries involve the forelimb (JRA 1991, Johnson et al. 1994, McKee 1995, Williams et al. 2001). Johnson et al. (1994) reported that the most common types of fatal MS injuries were fractures (83%) and ruptured ligaments (10%). The most common fracture sites were the proximal sesamoid bones, third metacarpal bone and humerus. Similarly, fatal injuries in flat and National Hunt flat races, in the UK, commonly involved the metacarpal and carpal bones (McKee 1995). In contrast, the shoulder was the most common site for fracture in hurdles and steeplechases. A 17-month retrospective study investigating fatal and non-fatal racing injuries in Kentucky, USA, found that 86% of injuries were located between the carpus and metacarpophalangeal joint (Peloso et al. 1994). The suspensory apparatus was involved in 45% of injuries. Fatal injuries were more likely to involve the left forelimb, sesamoid and third metacarpal, whereas non-fatal injuries were more likely to involve the superficial digital flexor tendon. Similarly, a three-year investigation of fatal and non-fatal injuries at flat and National Hunt races in the UK found that the flexor tendons and suspensory ligaments were involved in 46% of injuries (Williams et al. 2001). Training related MS injuries In studies of ranging in the duration from 81 days to two-years the percentage of MS injuries in horses during training has been reported to range from 36% to 57% (Mason and Bourke 1973, Rossdale 1989, Kobluk et al. 1990a, Lindner and Dingerkus 1993, Bailey et al. 1999, Aida et al. 2001). Furthermore, the MS injuries were found to recur in affected animals (Lindner and Dingerkus 1993, Bailey 1998, Perkins et al. 2004a). This is supported by a nine-month prospective investigation of 308 Thoroughbreds trained at a racetrack in Germany reported that 70% horses suffered from at least one episode in which training was reduced or prevented due to health problems (Lindner and Dingerkus 1993). Of the horses that had an episode in which training was reduced or prevented, 115 had one episode, 75 had two, 22 had three and four horses had more than three. Fifty-seven percent of these episodes were due to lameness. However, this study did not describe the types of lameness. A prospective study at Canterbury Downs, USA, reported that 61% of horses sustained an injury severe enough to result in the reduction or prevention of training (Kobluk et al. 1990a). Thirty-six percent of injuries involved bony/ hard tissues, 57% were soft tissues and 7% were not musculoskeletal in origin (Robinson et al. 1988). While the results are consistent with other studies, it is important to note that this was a pilot study that went for 81 days and involved only 95 horses (Kobluk et al. 1990a). Furthermore the estimate of prevalence of injuries in this study may have been imprecise because of numerous losses to follow up and incomplete records. In the Newmarket area, UK, a two-year prospective study involving 314 racehorses reported that 36% of horses suffered from a lameness that resulted in training being reduced or prevented (Rossdale et al. 1985). Lameness was reported in a further 16% of horses, but the trainer did not consider the problem severe enough to modify training. It was also found that lameness accounted for 68% of all modified training days. The most common causes of lameness were foot problems (19%), muscular problems (18%), carpal joint problems (14%), fetlock joint problems (14%), tendon problems (14%) and shin soreness (9%). Overall 68% of all lost training days were associated with lameness. 2

14 Mason and Bourke (1973) reported that at the end of the racing season 40% of horses in Melbourne, Australia were reported as unsound at the end of the racing season. However, the definition for unsound was not provided. The most common problems were shin soreness (46%), carpal problems (6.8%), splints (5.4%), fetlock problems (3.8%) and sesamoiditis (2.2%). In Sydney, Australia, a two-year prospective study of 169 young Thoroughbred racehorses reported that 85% of horses suffered some health problem that resulted in a modified training day and/or time resting at pasture (Bailey 1998). The majority of problems encountered by this cohort were low-grade injuries and disease. The most common health complaints were shin soreness and fetlock problems, which affected 42% and 25% of horses respectively. Of the horses that suffered from shin soreness, 40% developed the problem for a second or third time by the end of the 3-year-old racing season. Similarly, the recurrence of fetlock problems was 48%. Bailey (1998b) also found that shin soreness accounted for the greatest percentage of lost training days during the two- and three-year-old racing season. Shin soreness was also associated with the greatest percentage of weeks spent resting at pasture in the two-year-old racing season. In the threeyear-old racing season, fetlock problems accounted for the greatest percentage of weeks spent resting at pasture. The horses in study conducted by Bailey (1998b) were selected from the premier yearling sales. These horses represent a subset of the horse population that are thought to have superior conformation and perceived breeding value. Furthermore, the cost of these horses is more than others and as such the owners may differ in their expectations, either protective or pushy. In order to follow these horses a subpopulation of the best trainers in country must be enrolled. Risk factors for MS injuries Sex Studies investigating the association between sex and risk of catastrophic injury have produced conflicting results. In the USA a study of racing injuries found that in two-year-old Thoroughbreds, colts were over-represented (Wilson et al. 1996). This is supported by Rooney (1983b) who found significantly more fractures in stallions and colts than in geldings. However, neither of these studies used multivariable techniques to control for potential confounders. Mohammed et al. (1991), Bailey et l. (1997a) and Bailey et al. (1998) found no association between sex and risk of catastrophic injury. In contrast, other studies have found males to be at increased risk of catastrophic injury (Estberg et al. 1998a, Estberg et al. 1998b, Hernandez et al. 2001). Estberg et al. (1998) postulated that a possible reason for the differences between the two sexes may be because an owner may be more prepared to pay treatment costs for a female as it can be used for breeding. Studies of non-catastrophic injuries in populations of horses in training have also produced conflicting results. A study of suspensory apparatus injuries in Thoroughbred racehorses during training has shown that gender was not a risk factor (Hill et al. 2001). Univariable analysis by Perkins et al. (2004a) reported that males were 2.5 times more likely than females to sustain a tendon or ligament injury and 0.7 times less likely than females to sustain a non-fracture case of lameness. This was supported by the multivariable analyses that found males were 2.57 times more risk of suspensory apparatus injury and 1.74 time higher risk of superficial digital flexor tendon injury than females (Perkins et al. 2005). As with investigations of fatal MS injuries when investigating the relationship between sex and injury it is important to consider differences in economic value of male and female horses. When a gelding stops racing there is little or no opportunity for an owner to receive any economic return on their investment. Similarly, only a few entire males are of economic value in the breeding industry. In contrast, females can be used for breeding and therefore retain some economic value when they cease racing. 3

15 Age Descriptive studies of training-related MS injuries have reported that the rate of injuries varies between age groups, with the highest levels reported in two-years olds (Lindner and Dingerkus 1993, Perkins et al. 2004a, Perkins et al. 2004b). When different types of injuries are considered the rate of shin soreness is highest in two-year-olds and decreases with increasing age (Perkins et al. 2004a). A study found that the risk of fatal MS injury was less in two-year-old Thoroughbred racehorses than older horses (Wilson et al. 1996). The study also reported that two-year-olds sustained significantly fewer soft tissue injuries and carpal fractures and more upper limb long bone fractures than older horses. A descriptive analysis of racing injuries in Quarterhorses found that two-year-olds were not any more likely to sustain an injury (Cohen et al. 1999a). However, none of the studies used analytical techniques to control for confounding. When controlling for confounding several studies have reported that the risk of a horse sustaining a severe injury is greatest in older horses (Mohammed et al. 1992, Bailey et al. 1997, Bailey et al. 1998, Carrier et al. 1998, Estberg et al. 1998b, Williams et al. 2001). This is supported by a study that reported the risk of non-fatal suspensory apparatus injury was 2.3 times more in horses five years or older than in horses less than five (Hill et al. 2001). Caution is advised when interpreting the results of these studies as the results may be biased by confounders such as exposure to training and racing may bias the relationship between age and the risk of MS injury. Generally speaking, older horses have been in training for longer periods of time and therefore are likely to have been exposed to more high-speed exercise. In contrast, younger horses tend to be commencing training for the first time. It is possible to determine if the higher rate of injuries in two-year-olds is due to age or the commencement of training. Anecdotal evidence suggests that older horses commencing training for the first time will also develop shin soreness (Buckingham et al. 1990). To date there have been no well design prospective studies to determined if older horses are at less risk at the commencement of training than younger horses. In addition there have been no studies to determine if older horses that have had less training are at less risk of injury than horses of the same age. In the absences of such information it is difficult to determine if age, the degree of exposure to training or a combination of these factors are risk factors for MS injuries. Skeletal immaturity A survey of a small number of Victorian trainers and veterinarians found that the majority believed immaturity and too much too soon were the major risk factors for shin soreness (Buckingham and Jeffcott 1990). This is supported by Mason and Bourke (1973) who reported that horses commencing racing or hard training when skeletally immature, as measured by closure of the distal radial epiphyseal, had a higher incidence of carpal problems and shin soreness. However, the authors did not define hard training and their analysis did not control for potential confounders. Moreover, a study of 113 Standardbred racehorses failed to find any correlation between distal radius physeal closure and incidence of injury (Gabel et al. 1977). Based on this studies it is apparent that further research is required to determine if there is a causal link between skeletal immaturity and risk of MS injury. Bone strength Bone strength will influence its ability to withstand repeated loading. The total strength of the bone is determined by its stiffness or elasticity, mineral density and shape (Firth 2004). Studies in Quarter horses have shown that those with a greater cortical mass in the dorsal and medial metacarpal at the commencement of a training program had a lower injury rate (Nielsen et al. 1997). The significances of these findings is difficult to assess because the authors did not detail the types of injuries that were included. Davies et al. (1999) have measured the cortical thickness and develop a radiographic index to describe the shape of the bone. The index relates the width of the dorsal cortex to that of the palmar cortex, and using this index the group reported that it was possible to predict when a Thoroughbred racehorse would become shin sore. Use of this technique in a clinical setting would require careful 4

16 standardisation and this may limit its widespread application. Furthermore, methods such as those used by Davies et al. (1999) do not take into consideration intra-cortical porosity and can not detect changes in trabecular bone (Firth 2004). Studies investigating the relationship between bone strength and injury are complicated by the dynamic nature of bone. Studies have shown that bone mass and geometry change in response to changes in the normal loading pattern. If the load decreases then loss of bone mass occurs. In contrast, if the load increases then the bone mass and strength will increase. Investigations of the relationship between changes in bone mass and geometry are limited by lack of cost effective measures to measure bone response to exercise (Firth 2004). Body size and composition Body size affects the magnitude of forces applied to skeletal tissue and could theoretically be a risk factor for an MS injury. In human studies there are conflicting results regarding the relationship between body size and composition (Knutzen and Hart 1996). Some studies have reported that taller, heavier individuals are at greatest risk of injury, and other studies have reported no effect. To the author s knowledge there have been no studies to investigate these factors in Thoroughbred racehorses, although, a cross-sectional study of Icelandic horses found sound horses were 6 mm taller, at the croup, than unsound horses (Axelsson et al. 2001). However, the definition of sound and unsound was not provided and the clinical usefulness of a 6 mm difference is unclear. Clearly, further research is required to assess the impact of body size and composition on risk of injury. Conformation A few studies have examined the relationship between MS injury and conformation. One prospective study measured several variables relating to body shape (Kobluk et al. 1990a). These variables are described in Table 2. There was no significant association between the ratios of measurements in the forelimb and hindlimb and MS injury. A cross-sectional study subjectively described horse shape and conformation and measured a number of angles in the forelimb and hindlimb (Axelsson et al. 2001). Multivariate analysis found that an eight-degree reduction in the angle between the axis of the tarsal and the third metatarsal bone was the only measure of conformation associated with an increased risk of degenerative joint disease in the distal tarsus of Icelandic horse. The size of the difference limits its clinical usefulness. Table 2: Conformation measurements obtained from 95 horses in an 81-day prospective study. Limb Index Measurement Fore-limb Antebrachium length Elbow joint to mid carpus 3 rd Metacarpal length Mid carpus to fetlock joint Pastern length Fetlock joint to coronary band Hind-limb Tibia length Stifle joint to tibio-tarsal joint 3 rd Metatarsal length Tibio-tarsal joint to fetlock joint Pastern length Fetlock joint to coronary band Adapted from Kobluk et al. (1990) Shoeing and hoof angles A study in Japan found that horses fitted with plates for racing and training had lower injury rates than those that were not fitted with plates (JRA 1991). The reduction in risk may be because shoeing and hoof trimming can alter hoof angles and correct conformation faults such as bucked carpals and standing under (Bushe et al. 1990). A prospective study involving 95 horses measured the hoof angles and subjectively evaluated the foot for balance (Kobluk et al. 1990b). The results showed an association between low fore hoof angles and MS injury, but no association between steep hoof angles and injury. However, the analysis did not use multivariable techniques to control for any potential confounders. Furthermore the outcome in this study was all MS injury and the heterogeneous nature of the outcome may have biased results towards the null. 5

17 Using multi-variable techniques, at least two studies have demonstrated an association between shoeing and risk of MS injury. A case-control study in California, USA, involving horses that died or were euthanased with an MS injury reported that changes to shoeing could be used to modify the risk of fatal MS injury (Kane et al. 1998). In this study, controls were selected from horses that died or were euthanased for reasons other than an MS injury. Thus the exposure of controls to other risk factors may have been different to the cases and these differences may have biased the results. In a prospective study, the risk of suspensory apparatus was found to increase when a toe grab was used, and when the hoof was trimmed to perfect mediolateral symmetry (Hill et al. 2001). The authors reported that reducing the difference between toe and heel angles decreased the risk of suspensory apparatus injury. It is worth noting that shoeing practices differ between countries, due to differences in the rules of racing. The regional differences may limit the usefulness of these studies to the wider population of horses and the relative importance of shoeing as a risk factor for MS injuries. This highlights the importance of conducting country, or region, specific research. Previous or pre-existing injury Case studies have shown that injured horses often have either a previous injury or a pre-existing low grade MS injury (JRA 1991, Stover et al. 1992, Stover et al. 1993). This suggests that previous or preexisting injury may be a risk factor for MS injury. However, case studies do not examine the underlying distribution of the risk factor in the population. Therefore, it is necessary to review the results from analytical studies to determine if previous or pre-existing injuries are risk factors for MS injury. Several analytical studies have reported a relationship between a previous or pre-existing complaint and risk of MS injury. A case-control study found that abnormality in the suspensory ligament detected during the pre-race inspection was associated with a threefold increased in the risk of any kind of MS injury during the race and five fold increase in risk of an injury involving the suspensory apparatus (Cohen et al. 1999b). This is supported by a 90-day longitudinal study, that found a preexisting sub clinical or mild suspensory apparatus injury was a risk factor for all types of MS injuries and suspensory apparatus injuries (Hill et al. 2001). In three- and four-year-old Standardbred racehorses, the presence of radiographic changes before the age of two resulted in a two-fold increase in the risk of lameness (Gaustad et al. 1995). The results of these studies highlight the need for a prospective study to examine the relationship between factors that occur in the event and post event stage of an injury which may increase or decrease the risk of another MS injury. Fatigue Rooney (1983a) reported that fatigue is a significant factor in the cause of lameness, particularly in horses working at speed. In order to quantify fatigue, 234 races were examined. The velocity drop between the last and next-to-last segments was calculated and used to give the races a fatigue rating. Analysis of the data showed that a high fatigue rating was associated with certain track conditions and an increased number of breakdowns. This may also be because when the muscle is fatigued it is unable to reduce the stresses and strain on the bone (Brunker et al. 1999). Laboratory studies horses have shown that an inability to attain a previous maximum exercise speed in a subsequent run, within the same exercise period, is associated with an increase in bone strain (Davies 1996). Muscle fatigue has also been associated with distinctive changes in stride pattern and length in Thoroughbreds (Leach and Sprigings 1979). It has been suggested that this uncoordinated movement may result in excessive forces on some tendons and may increase the risk of tendon injury (Goodship and Birch 2001). Genetics The relationship between genetics and health problems is a relatively new area in epidemiology (Samet and Munoz 1998a). A study of Standardbreds in Norway revealed that an index relating to sire was associated with MS injury (Gaustad et al. 1995). The authors concluded that the sire index was significant because of genetic factors. Another study involving Icelandic horses found that, after 6

18 controlling for age and a number of other risk factors the prevalence of hind limb lameness varied between offspring groups (Axelsson et al. 2001). More research is required to determine the relationship between genetics, environment and the onset of MS injuries. Training-related factors Several studies have reported differences in injury rates between equine operations. Wilson et al. (1997) reported that there was a significant variation in race day injury rates between trainers. Two prospective studies have also reported that the prevalence of training related MS injuries differed between operations (Rossdale 1989, Ross and Kaneene 1996b). A study of training and health problems Thoroughbred in racehorses showed that there were differences in the training patterns between trainers, and that stables with higher average exercise scores had lower injury rates (Kobluk et al. 1990a). The results of these studies support a hypothesis that factors relating to training may be risk factors for injuries. A study of horses involved a variety of activities reported that horses involved in racing were 1.75 times more likely to suffer an episode of lameness than those not involved in racing (Ross and Kaneene 1996a). This is supported by Hill et al. (2001) who found that Thoroughbred racehorses were 3.5 times more likely to sustain a suspensory apparatus injury in the seven days following a race (Hill et al. 2001). Estberg et al. (1998a) found that the risk of MS injury increased if a horse was exposed to an excessive amount of high speed exercise, either in official workouts or races, in the previous 60 days. High-speed exposure for an individual horse was considered excessive if it exceeded a threshold level that varied by class and calendar year (Table 3). Table 3: Year- and age-specific cut off values for the average rate of distance accumulation of racingspeed exercise in 60 days, above which the time period was classified as a hazard period. Cut off values for accumulation of racing speed (meters) Calendar year 2-year olds 3-year-olds 4-year-olds 5-year-olds < Adapted from Estberg et al. (1998a) Verheyen and Woods (2003) have reported that an association between increasing distances cantered, in a 30-day period, and risk of tibial and pelvic stress fracture. The group found no association between distances trained at high-speed and risk of stress fracture. It is difficult to comment further on the results of these studies as the speeds typically associated canter and high-speed work were not described. Verheyen et al. (2005) reported an increased risk of dorsometcarpal disease with increased exposure to canter ( 15 seconds/furlong) high-speed (>15 seconds/furlong) exercise in one, two and four weeks periods. The results of the Verhyen et al. (2005) contradict Boston and Nunamaker (2000) who reported a protective effect of exposure to speeds 900 m/minute and an increase in risk associated with increased exposure to speeds around 600 m/minute. Verhyen et al. (2005) suggest that the most likely reason for this is that Boston and Nunamaker determined average distances trained in each week by dividing the cumulative sum of distance at each speed during the whole training period by the time at risk. Other possible reason for the difference between the current study and Verhyen et al. (2005) included methodological differences between the studies such as differences in data collection methods. In Verhyen et al. all the data was collected prospectively whilst Boston and Nunamaker (2000) used a combination of prospective and retrospective data collection, in some cases the retrospective records were 20-years old. Over time the accuracy of the record keeping and the extrinsic risk factors could have changed. In addition the commencement of the study may have caused reporting to be altered. 7

19 With the exception of Nunamaker and Boston (2000) the results of the previous studies and support a hypothesis that MS injuries are the result of accelerated micro-damage caused by repeated loading of the bone (Riggs and Evans 1990, Nunamaker 1996). At a microscopic level, the first signs of microaccelerated remodelling are vascular congestion, thrombosis and resportion of the bone tissue (Brunker et al. 1999). Once accelerated remodelling has commenced, continual loading of the bone will increase the size of the resorptive cavities (Riggs and Evans 1990, Nunamaker 1996). This will result in the appearance of micro fractures that extend into the cortex causing a marked reduction in bone strength. Post mortem examination of horses that had undergone gallop exercise and develop bucked shins reported that there was localised congestion and oedema of the periosteum and subcutaneous tissue (Katayama et al. 2001). Furthermore, there was lamellar bone formation and intracortical lytic change evident on micrographs. None of these changes were evident in two unexercised controls. Studies in humans indicate that in the initial stages these changes do not produce any symptoms (Brunker et al. 1999). However, as remodelling progresses mild pain occurs during exercise and even if the loading is stopped then pain will persist even after the completion of exercise. Complete fracture may even occur as there is insufficient bone to withstand the load. Extended breaks for training Laboratory studies have shown that the bone of horses that are box rested undergoes significant changes (Firth 2004). These studies support a hypothesis that extended breaks from training will place the horse at risk of MS injury when they resume training. This is supported by population based studies. In Kentucky research has shown that horses that had not completed an official high-speed work or raced in the previous month were at increased of suffering from a non-fatal racing related MS injury (Cohen et al. 2000). Carrier et al. (1998) also reported that horses returning after a lay-up of more than 60 days were at increased risk of humeral or pelvic fracture. In both these studies the absence of exposure to high speed exercise do not necessarily equate to a break from training. Furthermore, no information was collected as to the reason for the lay-up. Consequently it is not possible to determine if the risk factors were a break from high-speed exercise, previous or preexisting MS injury or a combination of both factors. Track related factors The role that factors associated with racetrack design and management play in the onset of MS injury is unclear. Hill et al. (1986) concluded that there was not a significant difference in injury rate between racetracks used for Thoroughbred racing. Bailey (1998b) commented that while there was no significant difference overall there was a significant difference in fracture rates between Saratoga and Aqueduct racetracks. Peloso et al. (1994) also reported that there was no difference in injury rates at four Kentucky racetracks. However, the results reported by Peloso et al. (1994) do not truly reflect the opportunity for injury at each track as the injury rates were reported per race day rather than per race start (Bailey 1998). Other studies in the USA have reported that the fatality rate for two-year-olds racing at different dirt tracks ranged from 0 to 4.14 per 1,000 (Wilson et al. 1996). Unfortunately, none of these studies used multivariable techniques and the variation in fatality rates may be due to a number of factors including differences in the gender distributions of race entrants (Estberg et al. 1996b). When using a multivariate analysis to control for factors such as age and gender, Mohammed et al. (1991) found that horses racing at one track were at less risk of sustaining an MS injury. The results of this study are supported by Bailey et al. (1998) who reported that one of four metropolitan racetracks in Melbourne, Australia, was associated with a significant increase in risk. In contrast, when using a multivariate approach to analysis of risk factors for injury at two racetracks in Sydney, Australia, there was no significant difference between tracks (Bailey et al. 1997). Comparison of injury rates pre- and post track reconstruction in Thoroughbreds in Japan (Oikawa et al. 1994) and Standardbreds in Australia found a significant reduction in injuries post reconstruction (Evans and Walsh 1997). Whilst the results of these studies are conflicting, there appears to be sufficient evidence to support a hypothesis that factors associated with the design and 8

20 management of racetracks are risk factors for MS injury. These factors may include racetrack geometry, track surface, track condition, camber, starting chutes and thatch accumulation and grass roots. Further details on track related risk factors can be found in the RIRDC report Racetrack design and performance. Seasonal effects Presently the role of season in the onset of an MS injury has not been adequately addressed. A retrospective descriptive study of injuries in Japan noted that fatal and non-fatal fractures whilst racing were more common in early spring (JRA 1991). In contrast training injuries were more common during winter. Similarly, a nine-month prospective investigation study in Germany noted that the incidence of lameness was highest in February (Lindner and Dingerkus 1993). However, neither study controlled for confounders. Therefore it is difficult to determine if the seasonal effects observed in these studies are a true result or due to the effect of exposure to other variables, such as previous racing and track condition. A case-control study using analytical methods to control for a number of potential confounders reported that horses racing in summer were three times more likely to sustain an MS injury than those racing in winter (Mohammed et al. 1991). However, this study relied on racing histories and therefore may have been biased by other unmeasured factors including changes in the level of high-speed exercise coincident with the season. A study of training injuries in Japan reported that more MS injuries occurred in February and April (Aida et al. 2001). The authors noted that the seasonal effect did not appear to correlate to the training menu. However, the authors did not describe the variables that were included in the training menu or the statistical techniques used to investigate the association. Stable Management There are a number of issues relating to the management of a stable that could be risk factors for MS injuries. Analysis of data collected in the Michigan Equine Monitoring System (MEMS) found that horses on larger operations were at less risk of lameness while those on operations with a higher level of veterinary and farrier involvement were at greater risk (Ross and Kaneene 1996b). It is not clear if this relationship represents a true alteration in risk or if the differences reflect the level of detection of MS injuries or higher level of involvement because their injury rates are higher. Rider The risk of fatal injury was higher in horses ridden by amateur than those ridden by professional jockeys in National Hunt Race (McKee 1995). The relationship between rider and risk of injury could have been confounded by a number of factors including an increased likelihood of amateur riders being placed on older horses and horses with pre-existing injuries. Evidence from laboratory studies have shown that riders can redistribute their weight over the forelimbs (Schamhardt et al. 1991). The overall differences were small and there is no further evidence of an association between rider skill and injury. Weight of the rider may also play a role in the onset of MS injury. Studies have demonstrated that increasing the load a horse carries increases the ground reaction forces (Schamhardt et al. 1991). This increase in ground reaction forces could increase the load placed on the skeletal tissue. Therefore it is possible that the use of heavier riders could increase the incidence of injury. It is possible that training without a rider, for example on a treadmill, may help to reduce lameness. Conclusion Musculoskeletal (MS) injuries have been identified as a common health problem in Thoroughbred racehorses (Bailey et al. 1999, Perkins et al. 2004a, Perkins et al. 2004b). The most common MS injuries in two- and three-year olds have been identified as shin soreness (Bailey et al. 1999, Perkins et al. 2004a) and problems involving the carpus and fetlock joints (Bailey et al. 1999). These MS 9

21 injuries incur both direct and indirect costs. These direct costs include lost opportunity to race and treatment costs. The indirect costs relate to the negative impact of MS injuries on the public perception of racing and are difficult if not impossible to quantify. After reviewing the literature it is clear that there is very limited understanding of the risk factors for MS injuries. In particular, there is almost a complete absence of any investigations that have collected data on daily training in an effort to understand the training related risk factors for MS injuries. Studies that utilize daily training records are fundamental to the understanding of training related risk factors for MS injuries. An understanding of the risk factors for MS injuries would facilitate the design of strategies that may reduce the occurrence and impact of MS injuries. 10

22 Study design and management Introduction Presently research has focused on describing the frequencies with which injuries occur (Jeffcott et al. 1982, Rossdale 1989, Lindner and Dingerkus 1993, Bailey et al. 1999, Perkins et al. 2004a, Perkins et al. 2004b, Verheyen and Wood 2004), in identifying risk factors for fatalities that occur either during racing or training (Estberg et al. 1995, Estberg et al. 1996a, Estberg et al. 1996b, Carrier et al. 1998, Estberg et al. 1998a), and non-fatal racing-related injuries (Hill et al. 1986, Mohammed et al. 1991, Mohammed et al. 1992, Peloso et al. 1994, Wilson et al. 1996, Cohen et al. 1997, Wilson et al. 1997, Bailey et al. 1998, Estberg et al. 1998b, Hill et al. 2001, Parkin and Clegg 2004). There has only been a small number of studies investigating risk factors for non-fatal training-related MS injuries (Kobluk et al. 1990b, Kobluk et al. 1990a, Moyer et al. 1991, Moyer and Fisher 1992, Boston and Nunamaker 2000, Hill et al. 2001, Verheyen et al. 2003). The high number of MS injuries in two-year-olds in training and the lack of research in the literature was a major motivation behind the funding of this study. The aim of this study was to use prospective methods to describe the epidemiology of MS injuries in Australian two- and three-year-old Thoroughbred racehorses. Specifically, the study aimed to: describe the incidence of MS injuries; determine the impact of MS injuries using a variety of measure of performance; and determine risk factors for an MS injuries, in particular shin soreness. This chapter describes the design, implementation and management of the longitudinal study. Descriptive and multivariable analysis of this data will be presented in subsequent chapters. Recruitment and enrolment of study population Trainers Data were successfully collected from 14 trainers over a 27 month period. Trainers were sampled based on three criteria: location, expected number of two-year-olds entering training and the perceived willingness of trainers to participate. Location of the stable was important because if it was too far from the study centre the interval between visits would have had to be increased and/or additional people would be required to collect data. This was considered unsuitable as the relatively short interval between visits and the use of only one person for data collection were considered important factors for maintaining a high data quality during the study period. Expected number of horses was an important factor because it was perceived that enrolment of trainers with 10 or more horses would facilitate efficient data collection. The willingness of the trainers to participate in the study was important because data collection required a high level of compliance. Reluctance to provide the information could have compromised the quality of the data. Sampling trainers using non-probability methods was associated with a risk of bias (Thrusfield 1995, p 180). Selection bias may also have occurred because a number of trainers approached declined the offer to participate in the study (n = 11) or were removed for failing to comply with the protocol (n = 4). In addition to creating bias removal of four trainers for failing to comply with the study protocol represented a waste of resources because at the time of removal considerable time and money had been expended. In future, it would be appropriate to attempt to measure trainer compliance prior to enrolment. This could be achieved by requesting that a trainer complete a questionnaire concerning details of the training and management of their horses. Those trainers that did not complete the questionnaire or refused to answer questions would then not be enrolled in the study. It would also be advisable to enrol trainers for a probationary period that was sufficiently long enough to include two to three visits. If data collection during this period was satisfactory the trainer would then be asked if they would like to continue to participate in the study. 11

23 Horses In the 2000/01 and 2001/02 racing seasons all two-year-olds trained by a participating trainer were enrolled in the study when they entered the stable. Some of the eligible horses entered the training stable before 1st August, when they were still officially one-year-old. They were enrolled in the study at that time. Data were collected from the time of enrolment until either the end of the study or until the horse was lost to follow-up. A horse was classified as lost to follow-up if it was not in training in the last month of the study (July 2002). The last training day was recorded as the date of loss to follow-up. Lost to follow-up was defined in this manner because data were not collected when horses were away from the stable, making it difficult to determine when a horse was lost to follow-up. Using this definition, nearly 80% of horses enrolled in the 2000/01 racing season and over 50% of horses enrolled in the 2001/02 racing season were classified as lost to follow-up. Due to differences in the definition of lost to follow-up it is not possible to make direct comparisons to other studies, such as Bailey et al. (1999). The duration of follow-up for the horses enrolled in the 2000/01 season is described in more detail in Figure 1. Similarly, Figure 2 describes the duration of follow-up for horses enrolled in the study during the 2001/02 racing seasons. Figure 1: Graph of the duration of follow-up for two-year-old Thoroughbred horses enrolled in 2000/01 racing season. Dashed lines represent 95% Confidence Interval. Figure 2: Graph of the duration of follow-up for two-year-old Thoroughbred horses enrolled in 2001/02 racing season. Dashed lines represent 95% Confidence Interval. 12

24 Data collection One investigator (NC) visited trainers at approximately fortnightly intervals to enrol eligible horses and to collect training and injury data for those horses present at the stable. The injury status of all horses that left the stable between visits was confirmed during the visit. No other data were recorded when the horse was not in the stable. Enrolment data At the time of enrolment the horses' training and injury history, sex and either the racing name or breeding was noted. If sire and dam or racing name was not known, the horse was identified by its sire and sex, or its stable name. The stable name refers to the name used by stable staff to identify the horse. Sex was categorised as either male or female. The exact date of birth was determined for each horse using the Australian or New Zealand studbook as appropriate. Training data Training information had to be collected directly from the trainer because only barrier trials and races are routinely recorded in official industry databases. A standardised questionnaire was used to collect training information in an attempt to minimise bias. The questionnaire was designed in consultation with the trainers involved in the study, and it used common industry terms that described the gait and speed of exercise training. The categories used on the questionnaire are described in Table 4. The number of furlongs trained at `1/2 pace', `evens', `3/4 pace' and `home on the bit' were also recorded. During the study period there were times when the questionnaire was not completed and on these occasions the training data was entered as missing. The likely cause of missing data was that the person responsible for completing the questionnaire was absent for a period of time. Table 4 Categories for the training activities, gaits and expected speeds used in the questionnaire in a longitudinal study investigating musculoskeletal (MS) injuries in two- and three-year-old Thoroughbred racehorses. Activity Gait Expect speed Box rest N/A N/A Walk only N/A N/A Walk and swim only N/A N/A Trot and/or canter Trot or canter Less than 600 m/minute Evens or 3/4 pace Gallop 800 to 890 m/minute Home on the bit Gallop 890 m/minute Trial Gallop At or near race pace Race Gallop Race pace N/A = Not applicable Validity of the training data was investigated by comparison of the race and barrier trial start information with the information recorded in a commercial database. Results showed that the rate of incorrect entries was higher in the first part of the study and decreased in the later months (Figure 3). This suggests that there may have been a period early in the study when trainers familiarised themselves with the data collection process. Information was not available for other training activities and so other training data could not be validated. It would seem reasonable to assume that the trends observed in the incorrect recording of starts were similar in the whole data set, although the magnitude may differ. Future studies could consider a comparison of the trainer s records of daily training activities with an independent daily record of training activities. Such a study would be able to quantify the measurement error in records of training obtained from the trainer or their staff. 13

25 Rate per 100 starts May/00 Jul/00 Sep/00 Nov/00 Jan/01 Mar/01 May/01 Jul/01 Sep/01 Nov/01 Jan/02 Mar/02 May/02 Jul/02 Study month Figure 3: Number of starts in official barrier trials and races that were incorrectly recorded, per 100 starts by study month. Error bars represent 95% confidence intervals (CI). Injury data The injury status of all horses that left the stable between visits was confirmed during the fortnightly visit. If a horse had a problem involving the musculoskeletal system, additional information relating to the nature of the injury was recorded, including the anatomical location and the type of problem. For example, an injury may have been swelling and heat in the fetlock joint with no known cause, or it may have been caused by degenerative joint disease or sprained suspensory ligament. The information concerning the injury was provided by the trainer, and a veterinarian may or may not have been involved in the diagnosis. A problem involving the musculoskeletal system was classified as an injury if it resulted in the horse leaving the stable for more than seven days and was the direct result of training (i.e. injuries such as stone bruises were excluded). This differs from previous studies that included MS injuries that resulted in modified training (Rossdale 1989, Bailey et al. 1999). The definition used in previous studies was not applied in this study because it was felt that modified training was subjective. In contrast, a horse being removed from the stable was a clearly defined event. However, the severity of injury that would result in a horse ceasing training and thus being removed from the stable was likely to vary between trainers and should be considered as a reason for trainer effects in multi-variable models. Despite this limitation it was felt that defining an MS injury by removal from stable was more consistent that using modified training days. Classification of study time A preparation began on the day that the horse was enrolled in the study, or when a horse returned to training after an absence of more than seven days from the stable. The preparation continued until the horse was lost to follow-up or left the stable for more than seven consecutive days. The period between consecutive preparations was termed a spell-period or spell. A spell period that ended with the commencement of a preparation was classified as complete. A spell-period in which the horse did not return to training before the end of the study was classified as incomplete. 14

26 The movement of the horses in and out of stables during the study period and the varying duration of time in training made it sensible to aggregate data to preparations and spell periods, rather than racing seasons or horse. Training data was also grouped into preparations because in Australia horses often change trainers and owners during a racing season. Therefore trainers and owners can only be expected to influence the risk of injury at the preparation level. Thus in Australia, it is appropriate to report on risk factors for MS injuries at the level of the preparation, rather than the horse. Table 5 describes the number of horses contributing data and the number of preparations, and training days by age class and sex. Table 5: Number of horses contributing data and numbers of preparations and training days contributed by age class, and sex. Horses enrolled in 2000/01 contributed data to both the two- and three-year-old age classes. Age class Sex Horses Preparations Training days 2 Female Male Unknown Total Female Male Unknown Total and 3 Female Male Unknown Total Conclusion This is the largest study undertaken in Australia to examine risk factors for MS injuries in horses in training. During the study period 451 horses were enrolled during their two-year-old racing season. Training and injury information for these horses were collected directly from the trainer. This required a high level of co-operation and therefore trainers were sampled using a non-random method. Validity of the training data was investigated by comparison of the race and barrier trial start information with the information recorded in a commercial database. Information was not available for other training activities and so other training data could not be validated. Future studies could consider a comparison of the trainer s records of daily training activities with an independent daily record of training activities. Such a study would be able to quantify the measurement error in records of training obtained from the trainer or their staff. 15

27 Incidence and impact of musculoskeletal injuries Introduction Musculoskeletal (MS) injuries have been identified as a common problem health problem in Thoroughbred racehorses (Rossdale 1989, Kobluk et al. 1990a, Lindner and Dingerkus 1993, Bailey et al. 1999, Perkins et al. 2004a, Perkins et al. 2004b). However, there have been very few studies that have described the frequency of training injuries using measures that take into consideration time at risk (Bailey et al. 1999, Perkins et al. 2004a, Verheyen and Wood 2004). MS injuries represent a major economic cost to the Thoroughbred racing industry in terms of treatment cost and lost opportunity to race. Bailey (1998) reported that 45% of horses enrolled in a longitudinal study did not race during their two-year-old season. Bailey concluded that the principal reason for this was the high number of cases of low grade injuries and other health problems. However, no statistical tests were performed to determine if those horses that sustained an injury were less likely to race than those that did not sustain an injury. It is possible that the horses that sustained MS injuries sustained them after their first start. Therefore, care should be taken in concluding that MS injuries, or other health problems, are the reason horses do not start at all as two-year-olds. The aims of this chapter were to describe the frequency of MS injury and use multivariable techniques to describe the impact of MS injuries on time to first start and factors that influence recovery from MS injury. Classification of MS injuries A problem involving the musculoskeletal system was classified as an injury if it resulted in the horse leaving the stable for more than seven days and was the direct result of training (i.e. injuries such as stone bruises were excluded). Information relating to the nature of the MS injury was provided by the trainer and did not necessarily involve a veterinarian. As the necessary diagnostic tests were not always conducted, it is possible that there was misclassification of the type of MS injury. For example an injury to the suspensory ligament may have been reported as swelling and/or heat in the fetlock joint with no known cause. Despite the potential for misclassification, it was considered preferential not to rely on confirmation by a veterinarian, as this may have resulted in an underestimation of MS injuries because common conditions such as shin soreness are often not seen by a veterinarian (Perkins et al. 2004a). Frequency of MS injuries Number of injuries During the study period 428 MS injuries were recorded in 248 two- and three-year-old Thoroughbred racehorses; 132, 74, 26, 10 and six horses sustained one, two, three, four and five injuries, respectively, while enrolled in the study. Table 6 describes the anatomical location and nature of the MS injuries recorded during the study period. The most common site of MS injury was the forelimb, in particular the lower forelimb. This is consistent with previous studies of horses in training conducted in Australia (Mason and Bourke 1973, Bailey et al. 1999), New Zealand (Perkins et al. 2004a) and UK (Rossdale 1989). Incidence rate Age-specific incidence rates (IR) were calculated separately for the incident MS injury and second MS injury. When calculating the IR for the incident MS, horses that sustained an injury contributed training days to the time at risk until their first MS injury. Horses that did not sustain an MS injury contributed training days at risk until the end of the study or until they were lost to follow-up. The IR for a second MS injury in the population of horses that had previously sustained any MS injury was 16

28 calculated by dividing the number of affected horses by the time at risk. Horses contributed training days to the time at risk from their return to training after the incident MS injury until they were lost to follow-up or sustained a second MS injury. For each category of MS injury incidence rate ratios (IRR) were calculated to enable comparisons of two- and three-year-olds. Figure 4 describes the IR for incident MS injury by age class. The rate of MS injuries was higher in two-year-old than in three-year-old horses, with two-year-olds 2.99 (95% CI = ) times more likely than three-year-olds to sustain an incident MS injury. When examining specific types of MS injuries the IR of all injuries, except for tendon and ligament, was higher in two-year-olds than threeyear-olds. The differences were only significant for injuries involving the third metacarpal bone. Examination of the IRR indicated that two-year-olds were 5.1 (95% CI = ) times more likely than three-year-olds to suffer from an MS injury involving the third metacarpal bone. IRs for second MS injuries are described in Figure 5. Comparison of the IRR indicated that two-yearolds were 2.89 (95% CI = ) times more likely than three-year-olds to suffer from a second MS injury. When individual types of MS injuries were examined the differences between the age classes were only significant for injuries involving the forelimb fetlock joint (IRR = 4.22; 95% CI = ) and the third metacarpal bone (IRR = 23.74; 95% CI = ). These results are consistent with a study of New Zealand horses in training that reported the rate of shin soreness was highest in two-year-olds and decreased with increasing age and that the IR for soft tissues injuries was highest in horses more than five years of age (Perkins et al. 2004a). This would seem to suggest that young horses are at increased risk of shin soreness and older horses are at increased risk of soft tissue injury. However, as the majority of horse s commenced training as twoyear-olds, the increased risk of shin soreness in this age group could be due to the onset of training. This is supported by anecdotal evidence that older horses commencing training will also develop shin soreness (Buckingham and Jeffcott 1990). Therefore, it is not possible to conclude that delaying the commencement of racing will results in a reduction in MS injures. 17

29 Table 6: Count and percentage of 428 musculoskeletal injuries by anatomical location and nature of injury during a longitudinal study of two- and three-year-old Thoroughbred racehorses. Location Nature of injury Number of injuries % of injuries (n= 428) Forelimb fetlock joint Degenerative changes Fracture/Chip Sesamoiditis Swelling and/or heat with no identified cause Total fetlock injuries Carpal joint Degenerative changes Cyst Fracture/Chip Problems with splints Swelling and/or heat with no identified cause Total carpal joint problems Third metacarpal bone Fracture/Chip Shin soreness Total third metacarpal bone problems Soft tissue Ligament Tendon Total soft tissue Other Back Fracture/Chip Hind limb lameness Muscle soreness Stress fracture Unidentified or unknown cause Total injuries classified as other

30 Rate (per 1000 horse training-days) yr-olds 3-yr-olds Forelimb fetlock joint Carpal joint Third metacarpal bone Soft tissue injury Other Any MS injury Injury Category Figure 4: Incidence rate estimates, per training days, for the first musculoskeletal (MS) injuries in horses. Horses contributed training days at risk either to the day of they experience the first MS injury or until they were lost to follow-up. Error bars indicate the upper 95% confidence interval Rate (per 1000 horse training days) yr-olds 3-yr-olds Forelimb fetlock joint Carpal joint Third metacarpal Soft tissue injury Other Any MS injury bone Injury Category 19

31 Figure 5: Incidence rates estimates, per training days, for the second occurrence of musculoskeletal (MS) injuries in horses. Horses contributed training days at risk either to the day of they experience the second MS injury or until they were lost to follow-up. Error bars indicate the upper 95% confidence interval. Impact of MS injury Survival methods were used to estimates the impact of MS injuries on the number of days from enrolment until the first start in a race or barrier trial. Similarly, time to recovery was taken as the number of days from time of MS injury until the first start in a race or barrier trial. These outcomes were chosen as they represent the time taken before owner(s) have an opportunity to receive a return on their investment. For both the time to first start and the time to recovery both spell and training days were included as many owner(s) incur cost regardless of whether the horse is spelling or training. Although, costs associated with spelling are generally lower than those incurred when training. To the author s knowledge, such values have not been reported for training related injuries and may prove useful in estimating the economic cost of MS injuries. Time to first start in a race or barrier trial The median time to first start was 199 days, or 6.5 months. The estimates in the current study may have been biased because some of the horses included in the analysis may have trained prior to enrolment, although these numbers were expected to be small as the data were collected from May 2000 well before the commencement of the two-year-old season Previous research has shown that 50% of horses catalogued had raced during the two-year-old racing season and 79.8% had started by the end of their three-year-old racing season (Bailey 1998). The results of the current study are not directly comparable to Bailey s report as a different time interval was considered. In Bailey report time to first start was taken as the time from the start of the twoyear-old season. In contrast, in the current study time to first start was taken from the time of enrolment. Therefore horses that had trained prior to enrolment were excluded, as the interval would not reflect the time from the commencement of training until the first start. The presence of an MS injury greatly increased the time to first start (Figure 6). Horses that did not sustain an MS injury had a median of approximately 132 weeks until their first start. The time to first start in horses that sustained an MS injury was approximately 232 weeks. This means that horses that sustained an MS injury were unproductive for 14 more weeks than those that did not sustain an MS injury. This represents a substantial increase in costs to the owner in form of lost opportunity, additional training and adjustment fees and any medical expenses associated with the injury. Even after adjusting for MS injury the time to first start varied significantly between trainers suggesting that factors other than the presence of MS injury play a role in the horses not starting in races. These may include management decisions to commence training but to delay the commencement of racing for a variety of reasons. It is possible that time to first start is delayed due to the presence of health problems other than MS injury, such as respiratory disease. Time to recovery Of the 248 horses that sustained an MS injury 70% (n = 173) returned to training after their first MS injury. The median time to recovery was 5.5 months. Of these horses that returned 133, or 77%, had at least one start in a race or official barrier trial while enrolled in the study. The median time to recovery was 5.5 months, or 166 days (95% CI = ). Time to recovery was significantly associated with the intensity of exercise prior to the onset of MS injury, with the time to recovery decreasing as the intensity prior to injury increased (Figure 7). This may be a true association or it may be biased by a number of factors. For example it may reflect a decision by owners and/or trainers to spell a horse for a longer period if it sustained an MS injury at low exercise intensity, thereby increasing the duration of recovery. The results may also be biased because MS injuries that occur at lower speeds may be due to the presence of other unmeasured variables, such as poor conformation. 20

32 The presence of these variables may make the horse more likely to sustain another MS injury and thereby delay recovery. Figure 6: Graphs of the days till first start in a race or barrier trial in Thoroughbred horses that sustained an MS injury before their first start and those that did not. Figure 7: Graphs of the duration of recovery in Thoroughbred horses that had galloped at speeds 15 seconds per furlong, > 15 seconds per furlong or had at least one start in a race or barrier trial. 21

33 Conclusion This chapter describes training and racing patterns in young Australian racehorses and supports findings both in Australia and overseas that the most common MS injury in two- and three-year-old Thoroughbred racehorses was shin soreness. Overall, the rate of injury was higher in two-year-olds than three-year-olds due mostly to a higher incidence of shin soreness in two-year-olds. The current study also showed that MS injury impacts negatively on the time to first start in a race or barrier trial. Other factors that associated with the time to first start were age and trainer. Time to recovery, as measured by a start in a race or barrier trial, was shortest in horses that had galloped at speed greater 15 seconds or started in a race or barrier trial. 22

34 Profiles of training preparations and spell periods Introduction An understanding of normal training patterns provides information that trainers and owners can use to benchmark the performance of their horses and is of great value in understanding the possible associations between training management and injury-risk in racehorses. A study of horses racing in South-Eastern Queensland used multivariable techniques to describe factors associated with performance and length of racing career (More 1999). The study found that in comparison to horses that were performing poorly, well performing horses were more likely to be male, have started as 2- year-olds and had more starts in the preceding 12 months. Length of racing career was found to be associated with performance in the first year of racing, sex, date of birth and age at first start. However, as noted by the author, interpretation of the results was limited due to confounding trainingrelated factors, clustering at the level of the trainer and the health status of the horses. Analyses presented in the previous chapter showed that the presence of MS injury increased the number of weeks from the commencement of training to the first start. Whilst, these associations are interesting horses frequently change trainers and owners and as such it would be useful to describe the impact of MS injuries at the level of the preparation and spell periods. This chapter presents the results of multivariable analyses to determine factors associated with training preparations and spell periods. Specifically, the analyses examined the impact of age, sex, cohort, trainer and MS injury. Overview of analytical methods Several outcomes were considered to describe training preparations and spell periods. Preparation level outcomes were the duration of preparation, the interval from the start of the preparation to the first start in an official barrier trial or race, interval from the first start until the end of the preparation and number of starts. The duration of preparation, time to first start in the preparation, interval from the first start until the end of the preparation and the spell period interval were investigated using standard methods of survival analysis (Hosmer and Lemenshow 1999, pp ). For those preparations with at least one start in a trial or race the survival analysis was also used to investigate the duration of the interval from the first start until the end of the preparation. Incomplete preparations were treated as censored when estimating the duration of preparations and the duration of the interval from the first start until the end of the preparation. When calculating the time to first start in a preparation, preparations in which the horse started in a race on the first day were excluded from further analysis and preparations with no starts were right censored. When estimating the duration of the spell period, only complete spell periods were included in the analysis. Multivariable Cox regression models were used to investigate the combined effects of a number of variables. The number of starts in official barrier trials and races per 100 training days were stratified by age class, sex, month, number of previous preparations and MS injury status of the preparation. A Poisson regression model was used to determine the combined effects of each of the variables. Factors influencing profiles The data set comprised 1,272 preparations recorded in 451 horses. The median duration of preparations was 54 days. During the study period 589 preparations, or 46% of all preparations, had at least one start in an official barrier trial or race and the median number of days to the first start was 59 days. In the 589 preparations with at least one start the median duration of the interval from the first start until the end of the preparation 34 days and the number starts per 100 horse-training days were Analysis of the complete spell periods found that the median duration was approximately 72 23

35 days. The impact of factors including age, sex, MS injury status and trainer on these profiles was investigated further and the results presented below. MS injury status The median duration of spell periods that were associated with MS injury was 11 weeks, while those not associated with an MS injury was 10 weeks (Figure 8). When conducting this analysis incomplete spell periods were excluded from the analysis. This may have resulted in an underestimation of the duration of spell periods because it was considered more likely that longer spell periods would be incomplete. In interpreting these results it is important to note that spell periods associated with MS injury were compared to those without MS injury. It is possible that these spells not associated with an MS injury may have been associated with other health problems such as respiratory disease. In the univariable screening, preparations before the first MS injury were shorter in duration and had a shorter interval from the first start until the end of the preparation and had less starts per 100 days than those after the first MS injury. However, after accounting for confounders MS injury status of the preparation was not significantly associated with any of the preparation level outcomes. The results of the current study suggest that observed differences between preparations before and after MS injury were biased by confounding. Figure 8: Graphs of the duration of spell periods that did and did not involve a musculoskeletal injury (MSI). 24

36 Age and number of previous preparations Age of the horse at the start of the preparation or age class described some of the variability in all the outcomes considered in this chapter. The results showed that younger horses tended to have shorter preparations (Figure 9), took longer to start in a race or barrier trial (Figure 10), had a shorter interval from the first start to the end of the preparation and fewer starts per 100 training days. Spells in twoyear-olds were also longer than those recorded in three-year-olds. The number of previous preparations was also significantly associated with all the outcome variables except for the time to first start in a preparation. This suggests that some of the age effects may be due to previous exposure to training. In order to examine the relationship between various outcomes, age and exposure to training it would be necessary to compare training of three-year-olds that trained as two-year-olds to those that did not train as two-year-olds. Unfortunately, this could not be done in the current study as all three-year-olds had trained as two-year-olds. Furthermore, it is the author s opinion that in Australia it would be difficult to enrol a group of three-year-olds that had not had some previous exposure to training. Figure 9: Graphs of the duration of preparations in preparations recorded in the two- and three-year-old racing season. 25

37 Figure 10: Graphs of the number of days from the beginning of a the preparation to a start in a race or barrier trial in preparations recorded in the two- and three-year-old racing season Sex Sex was unconditionally associated with the number of starts per 100 training days and the time to first start. However, in the multivariable models that adjusted for confounder, such as trainer and age, sex was not significantly associated with the outcome variable. This suggested that observed differences between male and female horses were the result of differences of age, trainer and other unmeasured covariates. Trainer All of the outcomes considered in this chapter differed significantly between the trainers. The observed trainer differences may have been due to unmeasured variables such as the presence of health problems other than an MS injury or conformation. Another possible reason for the differences between the trainers may be associated with differences in management factors. The presence of significant variability at the level of trainer for various outcomes of interest highlights the need to further investigate factors at the trainer level since this may represent an opportunity to make recommendations that could reduce the occurrence or impact of conditions interfering with training and racing. Calendar month The rate of starts varied between calendar months in the number of starts per 100 training days. The rate of starts was highest in February at 3.01 per 100 training days and lowest in October with 1.67 starts per 100 days. After adjusting for confounders the rate of starts was highest in August and September and lowest in October. This pattern is likely to reflect changes in the racing program. September has the first race for two-year-olds and therefore a large number of horses competed in timed barrier trials to start in those races. September is also the month of the spring racing carnival in Sydney. This carnival ends early in October when the major races move to Melbourne. It is at this stage that trainers will spell their older horses and start preparing their team of horses for the autumn carnival. 26

38 Cohort The duration of preparations, interval from the first start until the end of the preparation, spell periods, and the rate of start differed between cohorts even after adjusting for confounders such as age and trainer. These differences could have occurred for a number of reasons including the timing of races and perceptions of the trainers and owners regarding the appropriate races for the horses entering training that year. In Australia the timing of the autumn carnival is at Easter and therefore the date of the elite Group level races varies from year to year. This may impact on the time available to prepare horses for those races. The perception of the owners and trainers regarding the races that the horses are most appropriate for may also vary from year to year. For example, the horse may be perceived as being suited to early two-year-old races. Alternatively, the owners and/or trainers may perceive that the horse should not start until late in the two-year-old season or early in the three-year-old season Starts per 100 training days two-year-olds three-year-olds Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 11: Graphs of number of starts, per 100 training days, by calendar month in two- and threeyear-old Thoroughbred racehorse. 27

39 Conclusion The results present in this chapter showed that the presence of an MS at the start of the spell period significantly increased the duration of a spell period even after adjusting for trainer and age class. In contrast, the results suggest that MS injuries have limited long term impact on performance. Age of the horse at the start of the preparation or age class accounted for some of the variability in all the outcomes considered in this chapter. In multivariable models the number of previous preparations was significantly associated with all the outcome variables except for the time to first start in a preparation. This suggests that some of the age effects that have been reported in previous studies may have been confounded by previous exposure to training. All the outcomes also varied significantly between trainers. These differences need to be investigated as possible risk factors for MS injuries. 28

40 Risk factors for musculoskeletal injuries when horses are first exposed to high-speed exercise Introduction Both Bailey et al. (1999) and Perkins et al. (2004a) reported that incidence of MS injury is higher in two-year-olds. The current study also found that the incidence rate of MS injuries, in particular shin soreness, was higher in two-year-olds than three-year-olds. A survey of racetrack veterinarians and trainers found that the majority of trainers and veterinarians believed the commencement of fast work was associated with increased risk of shin soreness (Buckingham and Jeffcott 1990). Another prospective study investigating risk factors for MS injuries study concluded that horses commencing racing or hard training when skeletally immature, as measured by closure of the distal radial epiphysis, had a higher incidence of carpal problems and shin soreness (Mason and Bourke 1973). However, the authors did not define hard training and their analysis did not control for potential confounders. The results of these studies suggest that the commencement of training at high-speed (HS) marks the onset of a period in which a horse is at increased risk of MS injury. The aim of this chapter was to identify those factors that influence the risk of MS injury when horses are first exposed to HS exercise, that is gallop speeds greater than or equal to 800 m/minute. Risk factors for MS injuries A multivariable logistic model to determine risk factors for MS injury when Thoroughbred horses are first exposed to HS exercise. Preparations were excluded when more than 10% of the training days had missing data. This was because the exposure to HS exercise would have been underestimated and the estimates of the duration of the slow and HS gallop intervals were more likely to be biased. Similarly, variables for the last and second to last week were also coded as missing if there was more than one training day in the week with missing data as the estimates of exposure to HS exercise were likely to have been underestimated if there was more than one day with missing training data. The remainder of this section discusses the results of the analysis. Highest-gallop speed The results indicate that the risk of injury increased as the distance trained at HS increased. The risk of injury also differed depending on the highest gallop speed recorded in the preparation. The results suggested that different gallop speeds may be associated with a different risk of MS injury. With horses that galloped at speeds greater than 15 seconds per furlong at increased risk compared to those fast gallop speeds was less than or equal to 15 seconds per furlong. The results may have been biased as research has indicated that subjective gait does not correlate well to actual recorded speed (Rogers and Firth 2004). Therefore, future studies may wish to collect training data that included actual speeds. Collection and analysis of such data may facilitate the design of training programs that allowed horses to be exposure to HS exercise in a safe way. Previous MS injury The results of the analysis showed that horses that had previously sustained an MS injury were at increased risk of injury. This is supported by research that found pre-existing sub clinical or mild suspensory apparatus injury was a risk factor for all types of MS injuries and suspensory apparatus injury (Hill et al. 2001). Age In this analysis the effect of age was investigated using age in months at the commencement of training, at the start of the two-year-old season or at the start of the preparation. None of these variables were significantly associated with the outcome variables suggesting that delaying the age, in months, at which a horse is first exposed to HS exercise does not alter the risk of an MS injury. The 29

41 results are supported by Wilson et al. (1996) who found that two-year-olds with their first start early in the two-year-olds season were at no greater risk of MS injury than those horses with their first start at the end of the season (Wilson et al. 1996). Sex Sex was also not significantly associated with MS injury in the univariable screen and when included in the preliminary main effect model (results not present). This is supported by other studies investigating risk factors for suspensory apparatus injuries (Hill et al. 2001), racing injuries (Mohammed et al. 1991, Bailey et al. 1997, Bailey et al. 1998) and pelvic and tibial stress fractures (Verheyen et al. 2003). However, studies of fatal MS injuries have shown that, compared to females, males are at increased risk of injury (Carrier et al. 1998, Estberg et al. 1998b, Hernandez et al. 2001). The difference between the current study and those that found sex was a risk factor may reflect differences in outcomes. In the current study the outcome was mild to moderate MS injuries whilst the other studies investigated risk factors for fatal fractures. Conclusion This analysis focused on the first preparation to include exposure to HS exercise because commencement of training at high speed has been suggested as a risk factor for MS injuries, in particular shin soreness (Buckingham and Jeffcott 1990). The results suggested that changing the age at which horses were first exposed to HS exercise did not alter the risk of injury. The multivariable model demonstrated that exposure to HS exercise, in particular gallop speeds greater than 890 m/minute, was a risk factor for MS injury. The significance of the fixed effect for trainer in the model indicated that there are a number of factors relating to training and management of racehorses that may be risk factors for MS injuries. These may include aspects of the training regime, such as the rate of increase in distance trained at HS or management issues. 30

42 Risk factors for incident cases of shin soreness and lower forelimb joint injuries Introduction Previous studies have identified exposure to high-speed exercise as a risk factor for fatal MS injury (Estberg et al. 1995, Estberg et al. 1996a, Estberg et al. 1997, Estberg et al. 1998a). The outcome of these studies was catastrophic injuries. Therefore, the results are of limited usefulness to understanding of risk factors for non-fatal and less severe injuries that may occur during training and racing. There have only been a small number of studies investigating risk factors for non-fatal training-related musculoskeletal injuries (MSIs) (Kobluk et al. 1990b, Kobluk et al. 1990a, Moyer et al. 1991, Moyer and Fisher 1992, Boston and Nunamaker 2000, Hill et al. 2001, Verheyen et al. 2003). These studies have been conducted in the UK and USA and differences in training and management mean that the results are not directly applicable to Australian conditions. The results of these previous studies and those presented in the previous chapter may also be limited because they aggregated training data over a period of time to one observation per horse. For example in the previous chapter exposure to high-speed exercise was quantified by the cumulative sum of high-speed exercise in during the preparation. Summarizing data in this manner makes it difficult to describe the dynamic nature of training. The aim of this chapter was to investigate risk factors for injury-specific hazards for shin soreness, joint injuries and other forms of lameness. Overview of analytical methods This chapter used survival analysis methods to identify risk factors for shin soreness and joint injuries. Survival methods were considered an appropriate method for a number of reasons. Firstly, the incident MS injury was a well defined event as the horses were typically enrolled in the study when they first started training and injury records were available for those that had previously been in training. Therefore, horses that had previously sustained an injury could be excluded from the analysis. Secondly, the commencement of training was a clearly identifiable beginning time that was consistent across horses. Thirdly, the number of days since the commencement of training was a meaningful metric for measuring time. Exposure to training was quantified by determining the cumulative sum of various activities in a seven day window. The size of this window differs from previous studies that have used both 30- (Verheyen et al. 2003) and 60-day (Estberg et al. 1995, Estberg et al. 1996a, Estberg et al. 1997, Estberg et al. 1998a) exposure windows. The smaller windows were used in this study for two reasons. Firstly, the injuries considered in this analysis were low grade injuries and it was hypothesised that exposure to high speed exercise over shorter time intervals was an important risk factors. Secondly, use of larger windows, in particular the 60 day window, would have resulted in the exclusion of large amounts of data as more then 50% of preparations were less than 60 days in duration. The training variables were lagged by seven days because their value may have been dependent on the injury status of the horse. For example, a horse with an MS injury may be rested in a box prior to removal from the stable for a number of reasons, such as the need for intensive medical treatment. The decision to lag training variables by seven days was arbitrary and it is important to note that for some injuries in training may have been modified for more than seven days before the horse is removed from the stables. While, in some instances horses may have been removed from the stable as soon as the MS injury was detected and as such some relevant exposure information has been lost. Despite these issues it is the author s opinion that the inclusion of the seven day lag strengthens the argument that the association between covariates and the outcome is causal. 31

43 Risk factors Sex The results showed that compared to female horses male horses were at 1.77 times more risk of shin soreness. This increased risk could represent a true difference or be confounded by factors such as body size and shape or differences in training factors. For example one trainer in the study reported cantering male horses twice around the sand to keep the weight off them. In contrast, the risk of joint injuries did not differ between male and female horses. Age at the commencement of training Age at the time of enrolment in the study was identified as a risk factor for joint injuries but not for shin soreness. The results indicated that horses enrolled in the study at 24 months were 46% less likely than horses that were enrolled at < 24 months to sustain a joint injury. Owing to the design of the study enrolment corresponded to the commencement of training. Hence these results suggest that the incidence of joint injuries may be reduced if horses do not commence training until they are twoyears of age. However, further research is required as there has been some suggestion that training at a young age may aid in the bone development (Smith et al. 1999). Exposure to high-speed exercise The hazard of forelimb joint injuries was significantly associated with cumulative distance trained at high-speed. The results showed that compared to horses that were not exposure to high-speed exercise, the hazard of joint injury was 4.13 times higher in horses that had accumulated 2000 meters of high-speed exercise in a seven day period. Therefore, limiting the exposure to high-speed exercise should reduce the incidence of joint injuries in two- and three-year-old Thoroughbred racehorses. The hazard of shin soreness was also associated with the cumulative distance trained at high-speed. Compared to horses that had not been exposed to high-speed the hazard of shin soreness was 4.06 (95% CI = ), 7.11 (95% CI = ) and 6.92 (95% CI = ) times higher, respectively, in horses that had accumulated between 1 and 999 metres, between 1000 and 1999 metres and 2000 metres. The increased risk of shin soreness associated with exposure to high-speed exercise and shin soreness is supported by Verheyen et al. (2004). It is worth noting that the results of the current study and Verhyen et al. (2005) contradict Boston and Nunamaker (2000) who reported a protective effect of exposure to speeds 900 m/minute. Verhyen et al. (2005) suggest that the most likely reason for this is that Boston and Nunamaker determined average distances trained in each week by dividing the cumulative sum of distance at each speed during the whole training period by the time at risk. This is supported by results presented in their paper that showed when exposure to high-speed exercise was quantified in this manner the risk of dorsometcarpal disease decreased with increasing average weekly distances. In contrast when exposure was quantified using shorter time intervals of one week, two-weeks and one month the risk of dorsometcarpal disease increased as the distance trained at high-speed increase. Other possible reasons for the difference between the current study and Verhyen et al. (2005) and Boston and Nunamaker include methodological differences. In the both the current study and Verheyen et al. (2005) all the data was collected prospectively. In contrast, Boston and Nunamaker (2000) used a combination of prospective and retrospective data collection, and in some cases the retrospective records were 20-years old. Over time the accuracy of the record keeping and the extrinsic risk factors could have changed. In addition the commencement of the study may have caused reporting to be altered. The results of the current study and Verhyen et al. (2005) support a hypothesis that shin soreness and joint injuries are due to cumulative exposure to high-speed exercise that causes accelerated bone remodelling. Once accelerated remodeling has commenced continual loading of the bone will increase the size of the resorptive cavities, resulting in the appearance of micro fractures that extend into the 32

44 cortex and cause marked reduction in bone strength (Riggs and Evans 1990, Nunamaker 1996). At a microscopic level, the first signs of accelerated remodelling are vascular congestion, thrombosis and resportion of the bone tissue (Brunker et al. 1999, pp 8-9). Conclusion The results showed that exposure to high speed exercise increased the risk of both shins soreness and joint injuries. These results support a hypothesis that these injuries are the result of cumulative exposure to high-speed exercise that results in accelerated remodelling. Unfortunately, it is not possible to prepare Thoroughbred racehorses for racing without exposure to high-speed exercise. Therefore, future research should focus on developing strategies that would reduce the risk associated with exposure to high-speed exercise. Development of such strategies may reduce the incidence of shin soreness and joint injuries. 33

45 General discussion This research is the first large scale study of risk factors for training related injuries in Australian Thoroughbred racehorses. The major strength of the current study was that the data were collected prospectively, from a relatively large number of horses over an extended period of time. During the study period, 451 horses and 14 trainers contributed 1,272 preparations and 821 complete spell periods. This study focused on MS injuries that resulted in the horse leaving the stable for more than seven days. This differs from previous studies that included MS injuries that resulted in death (Estberg et al. 1995, Estberg et al. 1996a, Estberg et al. 1996b, Carrier et al. 1998, Estberg et al. 1998a), an extended break from racing (Hill et al. 1986, Mohammed et al. 1991, Mohammed et al. 1992, Peloso et al. 1994, Wilson et al. 1996, Cohen et al. 1997, Wilson et al. 1997, Bailey et al. 1998, Estberg et al. 1998b, Hill et al. 2001, Parkin and Clegg 2004) or modified training (Rossdale 1989, Bailey et al. 1999). A modified training definition as applied by Bailey et al (1999) was not used in this study because it was felt that modified training was subjective. In contrast, a horse being removed from the stable was a clearly defined event. However, there was likely to have been variability between the trainers in the severity of injury that resulted in a horse being removed from the stable. Frequency of MS injuries During the study period 428 MS injuries were recorded in 248 horses. The most common site of injury was the third metacarpal bone and the most common injury at the site was shin soreness. This is consistent with the results of previous research by Mason and Bourke (1973), Bailey et al. (1999) and Perkins et al. (2004a). The incidence rates (IR) of MS injuries was higher in two-year-olds than three-year-olds. When specific types of MS injuries were considered the IR for all categories of injuries, except for tendon injuries, were higher in two-year-olds than three-year-olds. However, only the differences for injuries involving the third metacarpal bone were statistically significant. This is consistent with a study by Perkins et al. (2004a) who reported that the rate of shin soreness was highest in two-year-olds and decreased with increasing age and that the IR for soft tissue injuries were highest in horses more than five years of age. The results of the current study and of Perkins et al. (2004a) would seem to suggest that young horses are at increased risk of shin soreness and older horses are at increased risk of soft tissue injury. The increased risk of soft tissue injuries in older horses may be the result of long term exposure to training rather than an effect of age. As the majority of horses commence training as twoyear-olds, the increased risk of shin soreness in this age group could be due to the onset of training. This is supported by anecdotal evidence that older horses commencing training will also develop shin soreness (Buckingham and Jeffcott 1990). Therefore, it is not possible to conclude that delaying the commencement of training will reduce the incidence of MS injuries. Impact of MS injuries The impact of MS injuries was assessed at both the horse and preparation/spell level. The two horse level outcomes were: (i) time to first start and (ii) time to recovery. These outcomes were chosen as they represent the time taken before owner(s) have an opportunity to receive a return on their investment. Several performance outcomes were considered when examining the impact of MS injuries at the level of the preparation/spell. The outcomes were: 1. Duration of preparations. 2. Days in the preparation until the first start in either a race or barrier trial. 3. Days from the first start in either a race or barrier trial and the end of the preparation. 4. Number of starts per 100-training days. 5. Duration of spell periods. 34

46 Horse-level measures of impact The presence of an MS injury greatly increased the time to first start. Horses that did not sustain an MS injury had a median of approximately 19 weeks until their first start. The time to first start in horses that sustained an MS injury was approximately 33 weeks. This means that horses that sustained an MS injury were unproductive for 14 more weeks than those that did not sustain an MS injury. This represents a substantial increase in costs to the owner in form of lost opportunity, additional training and adjustment fees and any medical expenses associated with the injury. The significant variation between trainers in the time to first start, adjusted for injury status, suggests that factors other than the presence of MS injury play a role in the horses not starting in races. These may include management decisions to commence training but to delay the commencement of racing for a variety of reasons. It is also possible that time to first start was delayed due to the presence of health problems other than MS injury, such as respiratory disease. Of the 248 horses that sustained an MS injury, 70% returned to training after their first MS injury. The median time to recovery was 5.5 months. Time to recovery was significantly associated with the intensity of exercise prior to the onset of MS injury, with the time to recovery decreasing as the intensity of exercise prior to injury increased. This may be a true association or it may be biased by a number of factors. For example it may reflect a decision by owners and/or trainers to spell a horse for a longer period if it sustained an MS injury at low exercise intensity as they require time to grow, thereby increasing the duration of recovery. Preparation and spell level measures of impact After controlling for confounders the presence of an MS injury at the end of a preparation increased the duration of the spell period. The median duration of spell periods that were associated with MS injury was 11 weeks, while those not associated with an MS injury had a median duration of 10 weeks. Incomplete spell periods were excluded from the analysis. This may have resulted in an underestimation of the duration of spell periods because it was considered more likely that longer spell periods would be incomplete. When interpreting the results it is important to note that spells not associated with an MS injury may have been associated with other health problems such as respiratory disease. For the 70% (n = 173) horses that returned to training after an MS injury the duration of preparations, time to first start, duration of the interval from first start until the end of the preparation and the number of starts per 100 training days were not adversely affected by previous MS injury. This suggests that MS injuries in young horses may not have a long term impact on racing performance. Risk factors for MS injuries Exposure to high-speed exercise In the first fast preparation the odds of MS injury involving structures in the lower forelimb increased as the total distance trained at high speed increased. The results presented in this report showed that the risk of MS injury varied depending on the maximum gallop speed in the preparation. This suggested that the risk associated with cumulative exercise at medium-speed, fast-speed and starts may differ. This was not explored further. The hazard of forelimb joint injuries was significantly associated with cumulative distance trained at high-speed. The results showed that compared to horses that were not exposure to high-speed exercise, the hazard of joint injury was 4.13 times higher in horses that had accumulated 2000 meters of high-speed exercise in a seven day period. Therefore, limiting the exposure to high-speed exercise should reduce the incidence of joint injuries in two- and three-year-old Thoroughbred racehorses. The hazard of shin soreness was also associated with the cumulative distance trained at high-speed. Compared to horses that had not been exposed to high-speed the hazard of shin soreness was

47 (95% CI = ), 7.11 (95% CI = ) and 6.92 (95% CI = ) times higher, respectively, in horses that had accumulated between 1 and 999 metres, between 1000 and 1999 metres and 2000 metres. The increased risk of shin soreness associated with exposure to high-speed exercise and shin soreness is supported by Verheyen et al. (2004). The results of the current study and Verhyen et al. (2005) support a hypothesis that shin soreness and joint injuries are due to cumulative exposure to high-speed exercise that causes accelerated bone remodelling. Once accelerate remodelling has commenced continual loading of the bone will increase the size of the resorptive cavities, resulting in the appearance of micro fractures that extend into the cortex and cause marked reduction in bone strength (Riggs and Evans 1990, Nunamaker 1996). At a microscopic level, the first signs of accelerated remodelling are vascular congestion, thrombosis and resportion of the bone tissue (Brunker et al. 1999). The results of the current study and Verhyen et al. (2005) support a hypothesis that shin soreness is due to cumulative exposure to high-speed exercise that causes accelerated bone remodelling. Once accelerated remodelling has commenced continual loading of the bone will increase the size of the resorptive cavities, resulting in the appearance of micro fractures that extend into the cortex and cause marked reduction in bone strength (Riggs and Evans 1990, Nunamaker 1996). At a microscopic level, the first signs of accelerated remodelling are vascular congestion, thrombosis and resportion of the bone tissue (Brunker et al. 1999). Previous MS injury In the survival analysis the event of interest was the first or incident MS injury of a particular type. Therefore the impact of previous MS injury was not investigated in these analyses. The results of logistic regression showed that horses that had previously sustained an MS injury were at increased risk of injury. This is supported by research that found pre-existing subclinical or mild suspensory apparatus injury was a risk factor for all types of MS injuries and suspensory apparatus injury (Hill et al. 2001). Age There were conflicting results regarding the association between age and risk of injury. In the multivariable logistic analysis of risk factors for MS injury involving the forelimb in the first training preparation to include exposure to high-speed exercise, the effect of age was investigated using age in months at the commencement of training, at either the start of the two-year-old season or at the start of the preparation. In the multivariable logistic model there was no significant association between age and the outcome variable. This suggests that delaying the age, in months, at which a horse was first exposed to HS, did not alter the risk of an MS injury. The results were consistent with those reported by Wilson et al. (1996) who found that two-year-olds with their first start early in the twoyear-olds season were at no greater risk of MS injury than those horses with their first start at the end of the season. Age at the time of enrolment in the study was identified as a risk factor for joint injuries but not for shin soreness. The results indicated that horses enrolled in the study at 24 months were 0.54 times less likely than horses that were enrolled at < 24 months to sustain a joint injury. Owing to the design of the study enrolment corresponded to the commencement of training. Hence these results suggest that the incidence of joint injuries may be reduced if horses do not commence training until they are two-years of age. Sex The risk of shin soreness was 1.77 times more in males than females. This may represent a true biological difference. Alternatively, the results could be biased by confounders such as body size and shape or differences in training not included in this model. For example one trainer in the study reported cantering male horses twice around the sand to keep the weight off them. The risk of fetlock joint and carpal joint problems did not differ between male and female horses. 36

48 Recommendations for future research The results of the current study and those of Bailey (1999) and Mason and Bourke (1973) have contributed to our understanding of the frequency and impact of MS injuries in Australian two- and three-year-old Thoroughbred racehorses. However, there is a distinct lack of information regarding the epidemiology of MS injuries in older horses. Results from Perkins et al. (2004 a,b) indicated that the type of MS injuries and impact of these injuries vary between age groups. Therefore future research in Australia should describe the incidence and impact of MS injuries in horses more than three years old. The results of this study show that exposure to high-speed exercise is associated with an increased risk of injury. Unfortunately, Thoroughbred racehorses are unlikely to compete successfully unless exposed to high-speed exercise. Therefore, future research should focus on identification of ways in which high-speed exercise can safely be introduced into a training period. The results of Verhyen et al. (2005) also indicate that distances trained at canter may be risk factors for MS injuries. Therefore studies should collect information relating to distances trained at slower speeds. Ideally, these future studies should collect this information using less subjective measures than in the current study as the results of studies in New Zealand have suggested that use of subjective gait as a true measure of speed may be problematic (Rogers and Firth 2004). Recent advances in microelectronics and GPS technology mean that similar systems could be implemented in other countries with relative ease. Future studies need also to consider the role of track-related factors needs to be considered. The first step in such a study would be determine appropriate techniques that can be used to monitor the physical characteristics of racetracks, because such techniques do not presently exist (Stubbs 2004). A longitudinal study could be undertaken to determine variables relating to track performance that, after adjusting for training, are associated with MS injury. When designing studies to investigate safer ways to expose horses to high-speed exercise and trackrelated factors it is important to note that the results of the current study have shown that risk factors for different types of MS injuries do vary. Therefore, analyses need to consider the risk factors for different injuries separately. However, this does not mean that the studies need to examine only one outcome. On the contrary, to rationalise the resources it would be advisable to undertake prospective studies that investigate a number of injuries (Samet and Munoz 1998b, Szklo 1998). It is the authors belief that training information is the most difficult to collect. Thus collecting information on a number of different outcomes creates a more cost effective study. Furthermore, it did not appear that collection of a number of outcomes impacted negatively on the quality of injury information. Conclusion The research presented in this report confirmed previous research that MS injuries are a common problem in two- and three-year-old Thoroughbred racehorses. The results showed that injury prior to the first start had considerable impact on time to first start. However, preparations after the first MS injury do not appear to be adversely affected. The investigation of risk factors showed that increasing distance galloped at high speed increased the risk of MS injury. Future studies need to use more accurate measures of training to examine the relationship between distance galloped and other training related factors, and risk of MS injury. Finally, when considering future research the importance of country, and even region, specific studies can not be emphasised enough as differences in typical training and management strategies make it difficult to generalise the results. 37

49 References Aida, Sakamoto, Tomita and Asai (2001). 'Research on locomotive diseases in thoroughbred horses under training'. Journal of Equine Science. 12: 68. Axelsson, Bjornsdottir, Eksell, Haggstrom, Sigurdsson and Carlsten (2001). 'Risk factors associated with hindlimb lameness and degenerative joint disease in the distal tarsus of Icelandic horses'. Equine Veterinary Journal. 33: Bailey, Reid, Hodgson, Suann and Rose (1997). 'Risk factors associated with musculoskeletal injuries in Australian Thoroughbred racehorses'. Preventive Veterinary Medicine. 32: Bailey (1998). Wastage in the Australian Thoroughbred racing industry. Rural Industries Research and Development Corporation, Canberra. Bailey, Reid, Hodgson, Bourke and Rose (1998). 'Flat, hurdle and steeple racing: risk factors for musculoskeletal injury'. Equine Veterinary Journal. 30: Bailey, Reid, Hodgson and Rose (1999). 'Impact of injuries and disease on a cohort of two- and threeyear-old Thoroughbreds in training'. The Veterinary Record. 145: Boston and Nunamaker (2000). 'Gait and speed as exercise components of risk factors associated with onset of fatigue injury of the third metacarpal bone in 2-year-old Thoroughbred racehorses'. American Journal of Veterinary Research. 61: Brunker, Bennell and Matheson (1999). Stress Fractures. Carlton, Victoria, Blackwell Science Asia. pp Buckingham and Jeffcott (1990). 'Shin soreness: a survey of Thoroughbred trainers and racetrack veterinarians'. Australian Equine Veterinarian. 8: Buckingham, McCarthy, Savage and Jeffcott (1990). 'Clinical applications of transmission ultrasound velocity measurement in equine bone'. Australian Veterinary Journal. 67: 380. Bushe, Turner, Poulos and Harwell (1990). 'The effect of hoof angle on coffin, pastern and fetlock joint angles'. Annual meeting of the American Association of Equine Practitioners. Carrier, Estberg, Stover, Gardner, Johnson, Read and Ardans (1998). 'Association between long periods without high-speed workouts and risk of complete humeral or pelvic fracture in thoroughbred racehorses: 54 cases ( )'. Journal of the American Veterinary Medical Association. 212: Cohen, Peloso, Mundy, Fisher, Holland, Little, Misheff, Watkins, Honnas and Moyer (1997). 'Racingrelated factors and results of prerace physical inspection and their association with musculoskeletal injuries incurred in thoroughbreds during races'. Journal of the American Veterinary Medical Association. 211: Cohen, Dresser, Peloso, Mundy and Woods (1999a). 'Frequency of musculoskeletal injuries and risk factors associated with injuries incurred in Quarter Horses during races'. Journal of the American Veterinary Medical Association. 215: Cohen, Mundy, Peloso, Carey and Amend (1999b). 'Results of physical inspection before races and race-related characteristics and their association with musculoskeletal injuries in Thoroughbreds during races'. Journal of the American Veterinary Medical Association. 215: Cohen, Berry, Peloso and Mundy (2000). 'Thoroughbred racehorses that sustain injury accumulate less high speed exercise compared to horses without injury in Kentucky'. Annual meeting of the American Association of Equine Practitioners pp Davies (1996). 'The effects of different exercise conditions on metacarpal bone strains in Thoroughbred racehorses'. Pferdeheilkunde. 12:

50 Davies, Gale and Baker (1999). 'Radiographic measures of bone shape in young thoroughbreds during training for racing'. Equine Veterinary Journal. Supplement. 30: Estberg, Gardner, Stiver, Johnson, Case and Ardans (1995). 'Cumulative racing-speed exercise distance cluster as a risk factor for fatal musculoskeletal injury in Thoroughbred racehorses in California'. Preventive Veterinary Medicine. 24: Estberg, Stover, Gardner, Drake, Johnson and Ardans (1996a). 'High-speed exercise history and catastrophic racing fracture in thoroughbreds'. American Journal of Veterinary Research. 57: Estberg, Stover, Gardner, Johnson, Case, Ardans, Read, Anderson, Barr, Daft, Kinde, Moore, Stoltz and Woods (1996b). 'Fatal musculoskeletal injuries incurred during racing and training in Thoroughbreds'. Journal of the American Veterinary Medicine Association. 208: Estberg, Gardner, Stover and Johnson (1997). 'Intensive exercise schedules and risk of catastrophic musculoskeletal injury and lay-up in California Thoroughbred racehorses'. Annual meeting of the American Association of Equine Practitioners: Estberg, Gardner, Stover and Johnson (1998a). 'A case-crossover study of intensive racing and training schedules and risk of catastrophic musculoskeletal injury and lay-up in California thoroughbred racehorses'. Preventive Veterinary Medicine. 33: Estberg, Stover, Gardner, Johnson, Jack, Case, Ardans, Read, Anderson, Barr, Daft, Kinde, Moore, Stoltz and Woods (1998b). 'Relationship between race start characteristics and risk of catastrophic injury in thoroughbreds: 78 cases (1992)'. Journal of the American Veterinary Medical Association. 212: Evans and Walsh (1997). 'Effect of increasing the banking of a racetrack on the occurrence of injury and lameness in Standardbred horses'. Australian Veterinary Journal. 75: Firth (2004). 'Problems in quantifying bone response to exercise in horses: a review'. New Zealand Veterinary Journal. 52: Gabel, Spencer and Pipers (1977). 'A study of correlation of closure of the distal radial physis with performance and injury in the Standardbred'. Journal of the American Veterinary Medical Association. 170: Gaustad, Kjaersgaad and Dolvik (1995). 'Lameness in three-year-old Standardbred trotters - influence of parameters determined during the first year of life'. Journal of Equine Veterinary Science. 15: Goodship and Birch (2001). Exercise effects on the skeletal tissues. Equine locomotion. W. Black and H.M. Clayton, Ed(s). London, W.B Saunders. pp Hernandez, Hawkins and Scollay (2001). 'Race-start characteristics and risk of catastrophic musculoskeletal injury in Thoroughbred racehorses'. Journal of the American Veterinary Medical Association. 218: Hill, Carmichael, Maylin and Krook (1986). 'Track condition and racing injuries in thoroughbred horses'. Cornell Veterinarian. 76: Hill, Stover, Gardner, Kane, Whitcomb and Emerson (2001). 'Risk factors for and outcomes of noncatastrophic suspensory apparatus injury in Thoroughbred racehorses'. Journal of the American Veterinary Medical Association. 218: Hosmer and Lemenshow (1999). Applied survival analysis. New York, Wiley-Interscience. pp Jeffcott, Rossdale, Freestone, Frank and Towers-Clark (1982). 'An assessment of wastage in thoroughbred racing from conception to 4 years of age'. Equine Veterinary Journal. 14: Johnson, Stover, Daft, Kinde, Read, Barr, Anderson, Moore, Woods and Stoltz (1994). 'Causes of death in racehorses over a 2 year period'. Equine Veterinary Journal. 26:

51 JRA (1991). Preventing accident to racehorses: studies and measures taken by the Japan Racing Association. Report of the committee on the prevention of accidents to racehorse. Japan Racing Association. Kane, Stover, Gardner, Bock, Case, Johnson, Anderson, Barr, Daft, Kinde, Larochelle, Moore, Jagannatha, Stoltz, Woods, Read and Ardans (1998). 'Hoof size, shape, and balance as possible risk factors for catastrophic musculoskeletal injury of Thoroughbred racehorses'. American Journal of Veterinary Research. 59: Katayama, Ishida, Kaneko, Yamaoka and Oikawa (2001). 'The influence of exercise intensity on bucked shin complex in horses'. Journal of Equine Science. 12: Knutzen and Hart (1996). Running. Epidemiology of sports injuries. D.J. Caine, C.G. Caine and K.J. Lindner, Ed(s). Champaign, USA, Human Kinetics. pp 1-13 Kobluk, Robinson, Clanton, Trent, Ames and Gordon (1990a). 'Comparison of the exercise level and problem rate of 95 Thoroughbred horses: A cohort study'. Annual meeting of the American Association of Equine Practitioners: Kobluk, Robinson, Gordon, Clanton, Trent and Ames (1990b). 'The effect of conformation and shoeing: a cohort study of 95 Thoroughbred racehorses'. Annual meeting of the American Association of Equine Practitioners: Leach and Sprigings (1979). 'Gait fatigue in the racing Thoroughbred'. Journal of Equine Medicine & Surgery. 3: Lindner and Dingerkus (1993). 'Incidence of training failure among Thoroughbred horses at Cologne, Germany'. Preventive Veterinary Medicine. 16: Mason and Bourke (1973). 'Closure of the distal radial epiphysis and its relationship to unsoundness in two year old thoroughbreds'. Australian Veterinary Journal. 49: McKee (1995). 'An update on racing fatalities in the UK'. Equine Veterinary Education. 7: Mohammed, Hill and Lowe (1991). 'Risk factors associated with injuries in Thoroughbred horses'. Equine Veterinary Journal. 23: Mohammed, Hill and Lowe (1992). 'The risk of severity of limb injuries in racing Thoroughbred horses'. Cornell Veterinarian. 82: More (1999). 'A longitudinal study of racing thoroughbreds: performance during the first years of racing'. Australian Veterinary Journal. 77: Moyer, Spencer and Kallish (1991). 'Relative incidence of dorsal metacarpal disease in young Thoroughbred racehorses training on two different surfaces'. Equine Veterinary Journal. 23: Moyer and Fisher (1992). 'Bucked Shins: Effects of differing track surfaces and proposed training regimes'. Annual meeting of the American Association of Equine Practitioners pp Nielsen, Potter, Morris, Odom, Senor, Reynolds, Smith and Martin (1997). 'Changes in the third metacarpal bone and frequency of bone injuries in young quarter horses during race-trainingobservations and theoretical considerations'. Journal of Equine Veterinary Science. 17: Nunamaker (1996). 'Stress fractures in Thoroughbred racehorses'. Surgery Forum: 1-4. Oikawa, Ueda, Inada, Tsuchikawa, Kusano and Takeda (1994). 'Effect of restructuring of a racetrack on the occurrence of racing injuries in Thoroughbred horses'. Journal of Equine Veterinary Science. 14: Parkin and Clegg (2004). 'Risk of fatal distal limb fractures among Thororoughbreds involved in five types of racing in the United Kingdom'. Veterinary Record. 154:

52 Peloso, Mundy and Cohen (1994). 'Prevalence of, and factors associated with, musculoskeletal racing injuries of thoroughbreds'. Journal of the American Veterinary Medical Association. 204: Perkins, Reid and Morris (2004a). 'Profiling the New Zealand Thoroughbred racing industry 2. Conditions interfering with training and racing'. New Zealand Veterinary Journal. 53: Perkins, Reid and Morris (2004b). 'Profiling the New Zealand Thoroughbred racing industry 1. Training, racing and general health patterns'. New Zealand Veterinary Journal. 53: Perkins, Reid and Morris (2005). 'Risk factors for superficial digital flexor tendon and suspensory apparatus in Thoroughbred racehorses'. New Zealand Veterinary Journal. 53: Riggs and Evans (1990). 'The microstructural basis of the mechanical properties of equine bone'. Equine Veterinary Education. 2: Robinson, Kobluk, Clanton, Martin, Gordon, Ames, Trent and Ruth (1988). 'Epidemiology studies of musculoskeletal racing and training injuries in Thoroughbred horses, Minnesota, USA'. Acta Veterinaria Scandinavica. Supplment: Rogers and Firth (2004). 'Musculoskeletal responses of two-year-old Thoroughbred horses to early training 2. Measurement error and effect of training stage on the relationship between objective and subjective criteria of training workload'. New Zealand Veterinary Journal. 52: Rooney (1983a). 'Track condition in relationship to fatigue and lameness in Thoroughbred racehorses'. Equine Veterinary Data. 4: Rooney (1983b). 'The relationship of sex to breakdown in Thoroughbred racehorses'. Equine Veterinary Data. 4: 132. Ross and Kaneene (1996a). 'An individual-animal-level prospective study of risk factors associated with the occurrence of lameness in the Michigan (USA) equine population'. Preventive Veterinary Medicine. 29: Ross and Kaneene (1996b). 'An operation-level prospective study of risk factors associated with the incidence density of lameness in Michigan (USA) equine operations'. Preventive Veterinary Medicine. 28: Rossdale, Hopes, Digby and Offord (1985). 'Epidemiological study of wastage among racehorses 1982 and 1983'. Veterinary Record. 116: Rossdale (1989). 'The epidemiology of wastage in thoroughbred horses'. Pro Veterinario. 9: 5-6. Samet and Munoz (1998a). 'Perspective: cohort studies'. Epidemiologic Reviews. 20: Samet and Munoz (1998b). 'Evolution of the cohort study'. Epidemiologic Reviews. 20: Schamhardt, Merkens and van Osch (1991). 'Ground reaction force analysis of horses ridden at the walk and trot'. Proceedings of the 3rd Equine Exercise Physiology Conference: Smith, Birch, Patterson-Kane, Firth, Williams, Cherdchutham, van Weeren and Goodship (1999). 'Should equine athletes commence training during skeletal development?: changes in tendon matrix associated with development, ageing, function and exercise'. Equine Veterinary Journal. Supplement. 30: Stover, Johnson, Daft, Read, Anderson, Barr, Kinde, Moore, Stoltz and Ardans (1992). 'An association between complete and incomplete stress fractures of the humerus in racehorses'. Equine Veterinary Journal. 24: Stover, Ardans, Read, Johnson, Barr, Daft, Kindu, Anderson, Woods, Moore, Stoltz and Pool (1993). 'Patterns of stress fractures associated with complete bone fractures in racehorses'. Annual meeting of the American Association of Equine Practioners:

53 Stubbs (2004). Racetrack design and preformance. Rural Industries Research and Development Corporation, Szklo (1998). 'Population-based cohort studies'. Epidemiologic Reviews. 20: Thrusfield (1995). Veterinary Epidemiology. Oxford, UK, Blackwell Science Ltd. pp 23, 180, 223. Verheyen, Newton and Wood (2003). 'A case-control investigating factors associated with pelvic and tibial stress fractures in Thoroughbred racehorses in training'. Society for Veterinary Epidemiology and Preventive Medicine: Verheyen and Wood (2004). 'Descriptive epidemiology of fractures occuring in British Thoroughbred racehorses in training'. Equine Veterinary Journal. 36: Williams, Harkins, Hammond and Wood (2001). 'Racehorse injuries, clinical problems and fatalities recorded on British racecourses from flat racing and National Hunt racing during 1996, 1997 and 1998'. Equine Veterinary Journal. 33: Wilson, Jensen and Robinson (1996). 'Racing Injuries of two-year-old Thoroughbreds and Quarter Horses'. Pferdeheilkunde. 12: Wilson, Shaw and King (1997). 'Thoroughbred racing injury rates are inversely associated with trainer success'. Annual meeting of the American Association of Equine Practitioners p

54 Risk Factors for Injuries in Thoroughbred Racehorses by N. Cogger et al RIRDC Publication No. 06/050 Lameness, or Musculoskeletal (MS) injuries, are the most common health problem in Thoroughbred racehorses and represent a economic cost to industry in terms of treatment cost and lost opportunity to race. In addition the high number of MS injuries in racehorses, particularly in two-year-olds, has also raised welfare concerns. In May 2000 a 27-month a RIRDC funded study investigating risk factors for MS injuries commenced. Multivariate statistical models were used to explore risk factors for MS injuries. The results suggest that MS injuries involving structures in the lower forelimb (carpus to fetlock inclusive) could be reduced by limiting exposure to high-speed exercise, in particular gallops at fast speed ( 890 m/minute). This supports the proposition that training injuries are caused by the accumulation of micro damage. The results also suggest there are a number of other factors that vary at the trainer level that may be risk factors for injuries, in particular joint injuries. These include unmeasured variables such as the rate of increase in distance galloped at fast-speed, conformation of the horse, skill of the riders and farrier and veterinary involvement. This publication can be downloaded from our website All RIRDC books can be purchased from:. Contact RIRDC: Level 2 15 National Circuit Barton ACT 2600 PO Box 4776 Kingston ACT 2604 RIRDCInnovation for rural Australia