VIEWS: 62 PAGES: 5 POSTED ON: 8/31/2011
................................................................................................................................ Clinical article Superficial digital flexor tendonitis in Thoroughbred race horses: outcome following non-surgical treatment and superior check desmotomy KT GIBSON, H M BURBIDGE and D U PFEIFFER Department of Veterinary Clinical Sciences, Massey University, Palmerston North, New Zealand Objective This study documents the results of non-surgical treatment and treatment by superior check desmotomy in Thoroughbred racehorses with superficial digital flexor (SDF) tendonitis. Design A prospective study was made of 124 thoroughbred racehorses with unilateral or bilateral SDF tendonitis. Procedure The flexor tendons were assessed by physical and ultrasonographic examination before treatment, and the lesions detected in affected tendons were characterised according to lesion type, length and cross-sectional area. Ninety three horses were managed non-surgically and 31 by superior check desmotomy. Recurrent or new injuries were defined as injuries affecting a previously injured superficial digital flexor tendon, the contralateral SDF tendon, or the suspensory ligament (interosseous muscle) in either forelimb. Results No statistically significant difference was found in ultrasonographic lesion severity between treatment groups. Horses managed by superior check desmotomy were 1.3 times more likely to complete five or more races than horses man- aged non-surgically (95% confidence limits 0.93-1.82). Horses treated surgically were 1.2 times more likely to develop recurrent or new injuries after returning to training than horses managed non-surgically (95% CL 0.95-155).Horses under- going superior check desmotomy were 5.5 times more likely to develop suspensory desmitis than horses treated non-sur- gically (95% CL 1.13-26.4). There was no difference in the time to recurrent or new injury between treatment groups. Conclusion There was no statistically significant difference between treatment groups in the proportions of horses able to complete five or more races after an episode of superficial digital flexor tendonitis. Superior check desmotomy did not appear to offer an advantage over non-surgical treatment in preventing recurrent or new injuries in Thoroughbred race- horses. Horses undergoing superior check desmotomy appeared to be at greater risk of developing suspensory ligament injuries than horses managed non-surgically. Aust Vet J 1997;75:631-635 Key words: Horse, tendonitis, superficial digital flexor tendon, Thoroughbred. SDF Superficial digital flexor Thoroughbred racehorses, and 69%’ to I njury to the S D F tendon in University Veterinary Clinic (MUVC) Thoroughbred racehorses is a 71 %’ for Standardbreds. Data are lack- between September 1391 and January common and career-threatening ing on success rates for Thoroughbred 1995, were included in a prospective injury. Recovery from such injury is racehorses managed by non-surgical study. Horses with injuries secondary prolonged, and recurrence frequent methods for SDF tendonitis. T h e pur- to lacerations, horses without after affected horses return to training pose of this study was to document the detectable SDF tendonitis, and those or racing.’ Recently, superior check results of non-surgical treatment and with inadequate follow-up were desmotomy (transection of the acces- treatment by superior check desmotomy excluded. sory ligament of the S D F muscle) has for SDF tendonitis in New Zealand Thoroughbred racehorses, and to History and physical examination been suggested as a treatment to reduce determine risk of recurrence of injury O n the day of examination, a this high recurrence rate.?’ ‘Success’ in these two groups of horses. detailed history was taken from the following treatment for tendonitis has owner or trainer, and horses examined been determined by the proportion of Materials and methods for lameness at the walk and trot. T h e treated horses able to compete in five One hundred twenty four tendinous structure affected, the pres- or more races postoperatively, without Thoroughbred racehorses with injuries ence and severity of lameness, and the recurrence of injury.6 Reported success to the S D F tendon in the metacarpal amount of swelling and pain on palpa- rates following superior check desmo- region, presented at the Massey tion were recorded. The presence of tomy vary from 52%’ to 67%” for Awt Vet/ Vol 75, No 9, September 1997 63 1 ................................................................................................................................ abnormalities unrelated to the flexor Table 1. Ultrasonographic grading system Zealand Racing Conference. Further used to describe core lesions in SDF tendon information was obtained by telephone tendons (for instance joint effusion) according to Genovese et ai? was also noted. interview with the owner or trainer o f Type 1 Lesion slightly hypoechoic compared the horses, and from medical records in Ultrnsonogr/lphirexamination to normal tendon. cases which were returned for re-exam- Before ultrasonographic examination, Lesion is halt echogenic, and halt ination. Follow-up intervals varied Type 2 the hair was clipped from the palmar anechoic. from 1.5 to 4.5 years from the time of surhce of the affected limbs, from the Lesion is mostly anechoic, represent- first examination. Type 3 accessory carpal bone to the ergot or, in ing significant fibre tearing. Horses were classified as: ( 1 ) retired some cases, the coronary band. The from racing without returning to train- Type 4 Lesion is totally anechoic, with almost skin over the site to be examined was total fibre tearing and haematoma ing, (2) retired for reasons unrelated to thoroughly wetted, and coupling gel formation. flexor tendon injury before completing applied. Ultrasonographic examina- five races, ( 3 ) did not race because of tion of the flexor tendons was per- recurrent or new injury to the flexor formed in sagittal and dorsal planes, addition to rest with or without a pro- tendons or suspensory ligament during using a linear 7.5 M H z micro-probe gram of controlled exercise. training, (4) raced one to four times, (Toshiba Ultrasound System Model Non-surgical management consisted but ceased training because of recur- SAL-55AS, with linear 7.5MHz micro- of rest, with the recommended area of rent or new injury, or (5) raced five or intra-operative probe IOE-702V. confinement being based on the sever- more times, with or without recurrent Toshiba, Japan). Images were recorded ity of the tendonitis. For horses with injury. on medical imaging film (Fuji Medical anechoic core lesions or severe diffuse Recurrent or new injuries were Imaging Film MI-NH, blue base, Fuji, tendonitis, stall rest with daily walking defined as recurring injury to the same Japan). in-hand was recommended. Despite SDF tendon, new injury to the con- Ultrasonographic findings were cate- this, most horses were paddock rested. tralateral S D F tendon, or injuries gorised according to lesion type, For horses with hypoechoic lesions, affecting the suspensory ligament in length, and the approximate cross- controlled exercise was recommended. either forelimb subsequent to a previ- sectional area of the tendon affected This consisted of walking and trotting ous SDF tendonitis. Horses that were (< 25%. 25-50%, > 50%). Limitations under saddle o n level ground. still racing without recurrent or new in equipment capability precluded Compliance with these recommenda- injury, but which had not yet complet- more accurate documentation of cross- tions was highly variable, and many ed five or more starts, were considered sectional area of the lesion. Core horses were not returned for re-exami- ‘untested’,h and were not used in calcu- lesions were graded on a scale of type I nation at the recommended times. lations. to 4, according to the method Superior check desmotomy was per- described by Genovese et al-(Table 1). formed on 31 horses using the proce- Statisticid analysis I n addition to these four grades, some dure described by Turner and Data were entered into a comput- lesions were described as diffuse. These McIlwraith.*All but two horses had the erised data base program (Microsoft lesions had a generalised anechoic or surgical procedure performed at Access for Windows 95 Version 7.0, hypoechoic appearance throughout the Massey University. Two horses, one of Microsoft Corporation, USA) and tendon, but were poorly defined, with which had unilateral, and the other analysed using a statistical software few or no short linear echoes present. bilateral surgery, were surgically treated package (NCSS for Windows Version Horses examined more than 3 months in Australia but recuperated in New 6.021, Number Cruncher Statistical after occurrence of the original tendon Zealand. Nine horses with unilateral Software, USA). During the statistical injury were assessed ultrasonographi- SDF injuries, treated early in the study analysis, the following parameters were cally for healing characteristics, but the (1991, 1992), were treated by unilateral used as outcome variables: recurrence lesions were not graded for compari- superior check desmotomy. Two horses of tendon injury, occurrence of new son. In these cases, the ultrasonograph- subsequently developed S D F ten- injury to the flexor tendons or suspen- ic assessment was used in recommend- donitis in the contralateral limb, and sory ligament, time to recurrence of ing the amount of exercise that the all since then have been treated with injury and time to occurrence of new horse should be permitted. In horses bilateral superior check desmotomy. injury. Categorical type risk factors, with acute or resolving injuries of the Surgery was performed with the horse including affected limb (left or right) SDF tendon, serial ultrasonographic in lateral recumbency in 29 cases, and and treatment category (surgical or re-examinations during the recovery in dorsal recumbency in 2 cases. non-surgical), and continuous type risk phase were recommended. factors, including length, type and Outcome approximate cross-sectional area of the fieatment Race records for horses returning to ultrasonographic lesion, were used as After ultrasonographic diagnosis, the racing were obtained from the New owner or trainer of each horse made a choice between non-surgical manage- Table 2. Age distribution ot 124 Thoroughbred horses with SDF tendonitis. ment (rest alone, or rest plus a program Age (years) 2 3 4 5 6 and over of controlled exercise) or surgical treat- Number (%) 9 (7.3%) 32 (26%) 29 (23.3%) 20 (16%) 34 (27.4%) ment (superior check desmotomy) in 632 ................................................................................................................................ independent variables in the following statistical analyses. Ultrasonographic lesion characteristics (length, type, cross-sectional area) were compared between treatment groups using the Mann-Whitney test and the chi-square test to assess if there was a difference between the sample of horses included in each treatment group. The statistical association between risk factors and each outcome variable was assessed using relative risk and its 95% confi- dence limits (95% CL), the Mann Whitney U test, or the Logrank test, depending on the scale of measure- 3 a al U C I al 35% 30% 25% 20% 15% 1Ph 5% 0% 1 n ment. Time to recurrence was analysed IPercentage of horses in each Category H Affected region of tendon using the Kaplan-Meier Product Limit Estimator. Figure 1. Distribution of lesions detected ultrasonographically in the SDF tendons of 124 Statistical significance was defined as Thoroughbred racehorses. The SDF tendon in the metacarpal region has been divided P < 0.05, 95% CL for relative risks into three zones.7 The middle and distal thirds of the tendon were most commonly which did not include 1, and statistical injured, although 29% of horses had injuries involving most of the length of the SDF tendon. power was estimated using P = 0.05. don is shown in Figure 1. Type 1 shown in Table 3. New or recurrent Results lesions were documented in 17 cases injuries occurred between the time the History and physical examination (1W o ) , type 2 in 13 cases (14%), type horse resumed training and 36 months Of the 124 Thoroughbred racehorses 3 in 13 (14%). type 4 in 38 limbs (50th race) after initial examination for included in the study, 33 were females, (42%), and i n 10 limbs (11%) the injury. 5 were entire males and 86 were geld- lesions were classified as diffuse. Superior check desmotomy (n= 31) - ings. The average age of the horses was Approximate cross-sectional area was Seventeen horses were treated by uni- 4.5 years (SD * 1.7 years), with a range documented in 88 horses with injuries lateral or bilateral superior check from 2 to 10 years. The age distribu- of < 3 months duration. In 44 cases desmotomy for a first episode of SDF tion of horses is shown in Table 2. (500/o), the lesion occupied less than tendonitis, and 14 horses were under- Thirty seven horses (30%) had not 25% of the cross-sectional area of the going treatment for recurrent episodes. raced at the time of initial presentation, tendon, in 28 cases 25-50% and in 16 N o horses in this category were 18 horses had raced between 1 and 4 cases over 50%. retired from racing before returning to times, 32 between 5 and 15 times, and training. Four horses were retired prior 37 horses had raced more than 15 Peiitment and outcome to completing five races for reasons times. In 91 cases the injury was < 3 Ninety three horses received non- unrelated to the flexor tendon injury, months old at the time of examination. surgical treatment and 31 were treated including exercise-induced pulmonary Fifteen horses were presented 3 to 6 by unilateral or bilateral superior check haemorrhage ( I ) , death associated with months after injury, and 17 were pre- desmotomy. colic (2) and an unspecified respiratory sented more than 6 months after the Non-surgical treatment (n=93) - condition (1). time of injury. The duration of injury From this group, 1 7 horses were retired T h e outcome for the 27 horses was unknown in one case. without returning to training for various which returned to training, and did not The left SDF tendon was affected in reasons including severity of injury; retire for unrelated reasons, is shown in 52 cases, the right in 53 cases and both advanced age of the horse, perceived Table 3. New or recurrent injury to the SDF tendons in 19 cases. O f 114 cases breeding value, concurrent injuries, flexor tendons or suspensory ligament in which the results of lameness exam- lack of ability, economic and other occurred between the time the horse ination were recorded, 43 horses (37%) unspecified reasons. Another 17 horses resumed training, and the 48th month were lame in the affected limb on the were retired before they had completed after examination for injury. day of presentation and 10 (8.7%) five races for reasons unrelated to flex- were lame in another limb. or tendon injury including: unrelated Relationship between outcome vari- lameness or injury (12), lack of ability ables and recurrence of injury Ultrasonographic examination (4) or death (1). Although some of T h e relative risk of recurrent or new T h e length of the ultrasonographi- these horses were successful in alterna- injuries in the right forelimb compared cally detected lesion in the worst affect- tive careers of dressage, show-jumping with the left forelimb was 1.19 (95% ed limb was 8 cm or less in 52 cases, or eventing, they were not considered CL 0.88 to 1.6). No statistically signif- between 10 and 16 cm in 30 cases, and in further analyses for this paper. icant difference was found when com- 20 to 24 cm in 9 cases. The distribu- The outcome for the remaining 59 paring ultrasonographic length of tion of the lesions within the SDF ten- horses which returned to training is lesion, lesion type, or cross-sectional Aust Vet/ Vol 75, No 9, September 1997 633 ................................................................................................................................ Table 3. Outcome for the horses that returned to training. demonstrated age-related alterations in the crimp morphology in the SDF Conservative management Superior check desmotomy n = 93 n = 31 tendon of wild horses which had not undergone training, and proposed that 'Untested' horses: training may accelerate the ageing Retired after diagnosis, or raced < 5 times 34 4 process within the tendon. for reasons unrelated to flexor tendon injury In order to compare the two treat- Outcome of horses returning to training: ments in this study, the groups of horses Did not race because of recurrent or 16 (27Yob) 6 (22Yob) were standardised as much as possible. new injurya occurring during training Although the owners made the deci- 10 (17%) 2 (7%) sion o n treatment for their horses, Raced 1-4 times but ceased training because of recurrent or new injury there was no significant difference in the ultrasonographic severity of lesion Raced 1-4 times, still racing without injury 1 (2%) - between treatment groups. Even so, Raced 5 or more times prior to, or without 32 (54%) 1 9 (70%) horses managed by non-surgical meth- recurrenffnew injury ods were subjected by their owners or Recurrent or new injuries: trainers to a wide variety of exercise Percentage of horses returned to trainingkacing 39 (66%) 21 (78%) regimens, and many of the horses were which developed recurrent or new injuries Developed suspensory desmitis after return 2 (3%) 5 (18%) not returned for reassessment of ten- to training don injury prior to returning to train- ing. Horses which did not return to aRecurrent injury to the previously affected SDF tendon, or new injury to the contralateral SDF tendon or a training, or which retired from training suspensory ligament. bPercentage of horses which returned to training, and did not retire for unrelated reasons. for reasons unrelated to tendon injury were not considered useful subjects for statistical comparison. Recurrent area between treatment categories (for Discussion injuries to the same tendon and length of lesion z = -0.91, P = 0.36; for Thoroughbred horses in New injuries to the contralateral tendon lesion type (2 = 0.31,2 df, P = 0.86; for Zealand race o n turf tracks, in a clock- were considered together because cross-sectional area (2 = 0.76, 3 df, P = wise or anti-clockwise direction owners and trainers sometimes had 0.86). N o difference was found in depending on the location of the race- difficulty recalling which limb was ultrasonographic characteristics of track. A relatively large proportion of originally affected, and they interpreted lesions in horses with or without re- horses race to the age of 6 years in flat a new tendon injury as failure of treat- injury using the Mann Whitney U test races, and some continue beyond that ment, whether it affected the originally (for length of lesion z = -0.60, P = 0.55; age in hurdle races and steeplechase injured limb or not. for lesion type z = 0.67, P = 0.50; for events. T h e age distribution of affected Injuries to the suspensory ligament cross-sectional area z = 0.23, p = 0.82). horses, and the equal proportion of left were considered to be related, new There was no statistically significant and right forelimbs affected by injury, injuries. T h e suspensory ligament sup- difference between the survival curves probably reflect the local racing popu- ports the fetlock joint in extension. (Figure 2) for time to recurrent or new lation and conditions. Where anti- Suspensory ligament injuries occurred injury between right and left forelimbs clockwise racing predominates, other in both treatment groups subsequent or between surgically and non-surgically authors have found that a large propor- treated horses (forelimb: (2 = 0.68, df = 1, tion of tendon injuries occur in the left P = 0.41; treatment: (2 = 1.29, df = 1, forelimb of Thoroughbred racehorses.",'" P = 0.26). Forty four per cent of horses were Horses undergoing superior check presented for evaluation of tendon desmotomy were 1.3 times more likely injury before their fifth race start. This to complete five or more races than group of horses sustained injury during horses treated non-surgically (95% CL preparation for racing, or during their 0.93 to 1.82; power = 0.28). There was first few starts. A similar observation no significant difference in the risk of was made by Rooney and Genovese" of recurrent or new injuries between the horses racing in the United States. two treatment groups, although the They concluded that fatigue due to relative risk of new injury was slightly lack of fitness may have contributed to higher (1.18) in the horses undergoing the large proportion of injured horses surgical treatment (95% CL 0.9 to in this group. In this study, however, a Months after injuly 1.55; power = 0.19). Horses treated by large proportion of affected horses were superior check desmotomy were 5.46 mature. Older horses may be predis- Figure 2. Kaplan-Meier survival curves times more likely to develop suspensory posed to SDF tendonitis by naturally showing proportion of horses still racing desmitis than those treated non-surgi- occurring, age-related changes in crimp without recurrent or new injury after cally (95% CL 1.13 to 26.4). formation in the central core of the surgical (solid line) and non-surgical SDF tendon.'*,I3Patterson-Kane et all2 (dotted line) treatment. 634 Aust Vet/ Vol75, No 9, September 1997 ................................................................................................................................ to an episode of S D F tendonitis, limited by the number of horses avail- study, and Vanessa Tilson for technical although were 5.5 times more likely to able for complete follow-up, however. assistance. occur in horses after superior check While a larger proportion of horses desmotomy. Hawkins and Ross4 were able to complete five or more References observed an apparently high incidence races after undergoing superior check 1. Genovese RL. Reef VB, Longo KL, Byrd JW. Davis WM. Superficial digital flexor tendonitis: of suspensory desmitis in a group of desniotomy, surgically treated horses Long term sonographic and clinical study of race- Standardbred racehorses which were appeared to be at greater risk of devel- horses. In: Rantanen NW. Hauser ML, editors. treated by superior check desmotomy oping recurrent or new injuries, Proceedings of the 1996 Dubai lnternational for SDF tendonitis, but were unable to including suspensory desniitis, than Equine Symposium: The equine athlete: tendon. ligament and soft tissue injuries. Veterinary Data, make comparisons with a group of horses managed by rest and controlled Wildomar, California, 1996:187-205. horses treated non-surgically. They exercise. 2. Bramlage LR. Superior check desmotomy as speculated that transection of the supe- In this study, horses were followed to a treatment for superficial digital flexor tendonitis: initial report. Proc Am Ass Equine Pract rior check ligament may result in the time of recurrent or new injury, or 1986;32:365-369. hyperextension of the carpus or fetlock in many cases, to the end of their rac- 3.Fulton IC. MacLean AA, OReilly JL, Church S. joint, and therefore predispose to sus- ing careers. Recurrent or new injuries Superior check ligament desmotomy for treat- pensory desmitis. Superior check occurred between the time the horse ment of superficial digital flexor tendonitis in Thoroughbred and Standardbred horses. Aust desmotomy on cadaver limbs resulted returned to training and 48 months Vet J 1994;71:233-235. in hyperextension of the fetlock joint after the time of first injury. Although 4. Hawkins JF, Ross MW. Transection of the and carpus, and increased strain on the many horses were able to complete five accessory ligament of the superficial digital flex- S D F tendon but not the deep digital or muscle for the treatment of superficial digital or more races after an episode of S D F flexor tendonitis in Standardbreds: 40 cases flexor tendon.14 Strain on the suspensory tendonitis, recurrent or new injuries to J (1988-1992)A m Vet Med Ass 1995;206:674-678. ligament was not measured in that the flexor tendons or suspensory liga- 5. Hogan PM. Bramlage LR. Transection of the study, but would be expected to ments limited the careers of a large pro- accessory ligament of the superficial digital flex- or tendon for treatment of tendinitis: long term increase with hyperextension of the portion of cases. results in 61 Standardbred racehorses (1 985- fetlock joint. Prospective studies involving clinical 1992).Equine Vet J 1995:27:221-226 No significant associations were cases suffer many important limita- 6. Bramlage LR, Rantanen NW, Genovese RL. found in this study between ultrasono- tions and this study was no exception. Page LE. Long-term effects of surgical treatment of superficial flexor tendinitis by superior check graphic features of the injury and the T h e number of horses available for desmotomy. Proc Am Ass Equine Pract likelihood of recurrence. This is in con- complete follow-up was limited, as 1988;34:655-656. trast to the findings of Genovese et al,' many horses developed unrelated 7. Genovese RL, Rantanen NW, Simpson BS. The use of ultrasonography in the diagnosis and who found that the greater the cross- injury or illness or were retired because management of injuries to the equine limb. Comp sectional area and length of the lesion, of lack of ability, thereby preventing Cont Educ Pract Vet 1987;9:945-955. the greater the likelihood of recurrence f d l evaluation of treatment for S D F 8. Turner AS, Mcllwraith CW. Superior check lig- of injury when the horse returned to tendonitis. It was difficult to ensure ament desmotomy (after Bramlage). In: Techniques in large animal surgery. 2nd edn. training. Similarly, Yovich et all" corre- compliance with recommended treat- Lea and Febiger. Philadelphia, 1989:147-149. lated increased lesion length with a ment strategies, with horses in both 9.Wilson JH. Robinson RA, Jensen RC, McArdle greater likelihood of re-injury. Cross- groups receiving widely varied exercise CJ. Equine soft tissue injuries associated with sectional area was not accurately docu- programs during recovery from injury. racing: descriptive statistics from American race- tracks. In: Rantanen NW, Hauser ML, editors. mented in this study, and many horses I n addition, owners had difficulty Proceedings of the 1996 Dubai lnternational were not presented for treatment until recalling exact details in cases of re- Equine Symposium: The equine athlete: tendon. several weeks to months after injury, injury, making it impossible to separate ligament and soft tissue injuries. Veterinary Data, Wildomar. California. 1996:l-21, thus affecting the ultrasonographic recurrent from new injuries in the 10. Yovich JV. Sawdon H. Booth T. Robertson type of lesion documented at the time analysis of results. Nevertheless, the ID. Correlation of ultrasonographic findings and of examination. results of this study suggest that, long term outcome in racehorses with superficial Superior check desmotomy has been despite treatment and apparent recov- digital flexor tendon injury. Aust Equine Vet 1996;13:88-92. suggested to reduce the risk of recur- ery, there is a high incidence of recur- 11. Rooney JR. Genovese RL. A survey and rence of injury in horses returning to rent or new injury in Thoroughbred analysis of bowed tendon in Thoroughbred race- training following recovery from S D F racehorses returning to training and horses. Equine Vet Sci 1981 $49-53. tendonitis.? According to the results of racing following an episode of SDF 12.Patterson-Kane JC. Firth EC, Goodship AE, Parry DAD. Age-related differences in collagen this study, the risk of recurrent and tendonitis. crimp patterns in the superficial digital flexor ten- new injury is not significantly different don core region of untrained horses. Aust Vet J in non-surgically treated horses and in Acknowledgments: 1997;75:39-44. 13. Wilmink J. Wilson AM, Goodship AE. horses undergoing superior check This project was funded by the New Functional significance of the morphology and desmotomy. There was no significant Zealand Equine Research Foundation. micromechanics of collagen fibres in relation to difference in the proportion of horses Thanks to: Tim Aldridge of the New partial rupture of the superficial digital flexor ten- able to complete five or more races Zealand Racing Conference, for pro- don in racehorses. Res Vet Sci 1992:53:354- 359. after either treatment, neither was there viding race records of Thoroughbred 14.Shoemaker RS, Bertone AL, Mohammad LN. a significant difference in the time of racehorses; D r Brian Anderson and Arms SW. Desmotomy of the accessory ligament recurrence between the two treatment Professor Elwyn Firth, who examined of the superficial digital flexor muscle in equine cadaver limbs. Vet Surg 1991 ;20:245-252. groups. T h e power of this study was some of the horses included in this (An-eprrdjor publictition I M a y 1997) Aust vet/ Vol75, No 9, September 1997 63 5
"Superficial digital flexor tendonitis in Thoroughbred race horses"