Approximately 8,400 yearling Thoroughbreds in the United States were offered at public auction annually between 2000 and 2015, according to the Jockey Club.1 Prepurchase examinations are an important aspect of the equine veterinary profession and are a source of professional liability.2 For yearling Thoroughbreds, prepurchase assessment at public auction often consists of brief visual inspection, observation of the horse walking in hand, endoscopic examination of the upper respiratory tract, and review of a standardized set of survey radiographs, as regulated by the specific sales company. Radiographic images are stored in repositories for review by prospective buyers and their veterinarian prior to the auction.
Given the results of the prepurchase examination, veterinarians are asked to make an accurate assessment of findings that may affect future racing or resale prognosis.3,4 The findings of prepurchase examinations and subsequent recommendations to clients should be based on objective, scientific evidence when possible. However, published objective data are lacking on the clinical importance of some common radiographic findings.
Researchers have attempted to identify normal and abnormal radiographic findings in yearling Thoroughbreds that could impact future racing performance. In research reported in 2003,3,4 radiographic data on 1,162 yearlings offered for public auction were analyzed, and the incidence of abnormalities as well as their correlations to racing performance were determined. Four abnormalities were significantly associated with a decreased incidence of horses starting a race as a 2- or 3-year-old: moderate to extreme supracondylar lysis of the third metacarpal bone, enthesophytosis of the proximal sesamoid bones, dorsomedial intercarpal joint disease, and distal intertarsal or tarsometatarsal osteophytosis or enthesophytosis. A significant decrease was also identified in the percentage of race starts in which horses placed, total earnings, and money earned per race start for horses with (vs without) proximal sesamoid bone enthesophytes. In contrast, in another study5 in which radiographs of 397 yearlings offered for sale in Kentucky were reviewed, only osteophytosis or enthesophytosis of the forelimb proximal sesamoid bones was significantly associated with a decreased likelihood of starting a race as a 2-year-old. Although both research groups found that osteochondral fragmentation (vs no fragmentation) of the proximal phalanx was associated with a decreased incidence of starting a race as a 2- or 3-year-old,4,5 this association was not significant.
An additional study6 involving 2-year-old Thoroughbreds in training revealed lower odds of starting a race when horses had (vs did not have) a diagnosis of dorsoproximal osteochondral fragmentation of the proximal phalanx, proximal sesamoid bone sesamoiditis or fracture, or wedge-shaped central or third tarsal bones. Other research showed poorer 2-and 3-year-old racing performance for yearlings with (vs without) radiographic evidence of sesamoiditis.7 Finally, in a study8 involving 348 yearling Thoroughbreds offered for sale in Texas, no single abnormal radiographic finding was significantly associated with the ability to start a race.
Fracture of the ACB in horses has been uncommonly identified in repository radiographs in previous studies.3,5,9,10 The overall prevalence of ACB fracture in all 4 studies3,5,9,10 combined was 4 of 2,176 (0.2%). Although this prevalence is low, ACB fracture does appear on repository radiographs, and to our knowledge, no peer-reviewed, objective data exist of its impact on future racing performance of Thoroughbreds.
The aim of the study reported here was to provide objective data regarding the prognostic value of identification of ACB fractures, specifically osteochondral fragments originating from the ACB, on the prepurchase radiographs of yearling Thoroughbreds. We hypothesized that there would be no significant difference in race performance between horses with ACB fracture diagnosed on yearling sale repository radiographs and their unaffected counterparts. Our second hypothesis was that, in horses with ACB fracture, concurrent osteoarthritic change (vs no such change) noted at the dorsal aspect of the affected carpus would be associated with poorer racing performance.
Materials and Methods
Study design
A retrospective matched cohort study was performed. Repository radiographic interpretation reports from the Keeneland September sale of yearlings in Kentucky (2005 through 2012) generated by any 1 of 3 experienced observers from 1 Kentucky-based private practicea or 1 experienced observer from a Florida-based private practiceb were searched. Experienced was defined as having reviewed a minimum of 1,000 sets of repository radiographs. The following radiographic images were included in the repository: metacarpophalangeal joints (dorsal 15° proximal-palmarodistal oblique, flexed lateromedial, standing lateromedial, dorsal 30° lateral 15° proximal-palmaromedial distal oblique, and dorsal 30° medial 15° proximal-palmarolateral distal oblique views), carpi (flexed lateromedial, dorsal 35° lateral-palmaromedial oblique, and dorsal 25° medial-palmarolateral oblique views), metatarsophalangeal joints (dorsal 15° proximal-plantarodistal oblique, dorsal 15° proximal 30° medial-plantarolateral oblique, dorsal 15° proximal 30° lateral-plantaromedial oblique, and standing lateromedial views), tarsi (dorsal 10° lateral-plantaromedial oblique, lateromedial, and dorsal 65° medial-plantarolateral oblique views), and stifle joints (lateromedial, caudal 20° lateral-craniomedial oblique, and caudal 15° proximal-craniodistal views).c
Animals
The exposed cohort was comprised of yearlings in which a fracture of 1 or both ACBs was noted on the radiographic interpretation report but that were presumed to be nonlame at the time (as per conditions of the sale). All horses with ACB fractures were initially included in this cohort; however, the cohort was further refined prior to analyses to include only those with osteochondral fragmentation specifically (referred to henceforth simply as ACB fracture). To select horses for the unexposed cohort, a 2:1 matching scheme was used by which 2 horses without any ACB fracture (1 yearling sold before and 1 sold after the matched horse with ACB) were included for every horse with ACB fracture. Specifically, unexposed horses were chosen from the yearlings with the closest hip number before and after the horse with the ACB fracture in the sale catalog, for which radiographic interpretation reports were available. This selection protocol was similar to that used in a previous study,9 except that the number of horses selected for the unexposed cohort was greater to increase statistical power while including representative individuals. This strategy allowed confirmation of the absence of ACB fracture in unexposed horses while providing age-matched counterparts with similar presale appraisal value and athletic potential for comparison.
At the Keeneland September sale, location in the sale catalog is based on presale analysis of pedigree power as well as appraisal of physical appearance and conformation performed by the sale company. For consistency, all horses selected for the unexposed cohort were identified from reports generated by 3 veterinarians at the Kentucky-based practice.a These horses were selected without replacement, so no unexposed horse was included in the data analyses more than once.
Categorization of lesions
Horses were first grouped for data analysis by exposure status (ACB fracture or no ACB fracture). The exposed cohort was then subdivided into yearlings with or without other radiographic findings previously identified in the literature as associated with a significant reduction in the likelihood of starting a race as a 2- or 3-year-old (ie, osteoarthritic change at the dorsal aspect of the carpus, moderate or severe tarsal osteoarthritis, forelimb proximal sesamoid bone enthesophytosis or osteophytosis, moderate or severe supracondylar lysis of the third metacarpal bone, or moderate to severe sesamoiditis).3,7 The exposed cohort was also subdivided into yearlings with ACB fracture and concurrent osteoarthritic change visible at the dorsal aspect of the affected carpus and those with ACB fracture and no such concurrent osteoarthritic change.
Radiographic findings considered to indicate osteoarthritis of the carpal joints included periarticular enthesophytosis or osteophytosis or remodeling at the dorsal aspect of the carpal bones.3 Likewise, radiographic findings consistent with osteoarthritis of the distal aspect of the tarsal joints included periarticular new bone formation at the dorsal aspect of the distal intertarsal or tarsometatarsal joint with or without a wedge shape to the central or third tarsal bones.3
Data regarding descriptors similar to “mild osteoarthritis of the lower hock joints” and “mild spavin” indicating a mild osteoarthritic change at the dorsal aspect of the distal tarsal joints were extracted as a separate variable and excluded from the definition of clinically important lesions because these findings were identified more commonly in horses without ACB fractures than in those with ACB fractures.
Racing performance analysis
Racing performance data were collected via an online third-party databasec of equine racing records and statistics, accessed on December 21, 2014, and May 15, 2016. The database was searched by the dam's name and horse's birth year. If no record existed, the horse in question was presumed to have failed to start a race. Collected performance data included the number of starts as well as earnings per year and earnings per start for the races performed as a 2- and 3-year-old. Data on career total starts, earnings, and mean earnings per start were also collected and assessed. Analyses of career performance data was performed, including only horses that had started at least 1 race and had not raced within 6 months before data collection.
Statistical analysis
Horses with bilateral lesions were included only once in the statistical analyses. The χ2 test of independence was used for comparisons of proportions of horses with versus without ACB fracture regarding certain binary characteristics (eg, raced vs did not race as a 2- or 3-year-old, presence vs absence of other radiographic lesions previously deemed important, and presence vs absence of radiographic evidence of carpal osteoarthritis). For 2- and 3-year-old horses, owing to the large number of 0 values for starts and therefore earnings, zero-inflated Poisson regression models and zero-inflated negative binomial regression models were used to compare horses with and without ACB fracture regarding the number of starts as well as earnings and earnings per start, respectively, while controlling for sex. The choice of zero-inflated model was based on a significant Vuong test result for comparisons of zero-inflated models with standard Poisson and negative binomial regression models. Because career analyses included only horses that had started a race (ie, no 0 values), standard linear regression modeling was used for starts, earnings, and earnings per start after logarithmic transformation of these variables.
The exposed cohort was then subdivided, and analyses were repeated to compare horses that had ACB fracture and other radiographic abnormalities previously deemed important with horses that had ACB fracture and none of these abnormalities. In a final analysis, horses with ACB fracture and concurrent carpal osteoarthritis were compared with those with ACB fracture and no such osteoarthritis. For all tests, values of P < 0.05 were considered significant. All statistical analyses were performed with statistical software.d
Results
Horses and fractures
During the study period, 52 ACB fractures were identified in 50 Thoroughbred yearlings. Two horses had unilateral vertical fractures in the frontal plane of the ACB and were subsequently excluded, and another horse was excluded because it would have shared the same matched horses without ACB fracture as another horse with ACB fracture, violating the assumption of independence of the horses. Consequently, 47 yearlings (2 with bilateral fractures; 27 [57%] males and 20 [43%] females) with 49 fractures (27 of the right ACB and 22 of the left ACB) were included in the exposed cohort, and 94 yearlings without ACB (55 [59%] males and 39 [41%] females) were included in the matched unexposed cohort.
Of the osteochondral fragments noted in the repository radiographic interpretation reports for horses with ACB fracture, most (29/49 [59%]) were in a dorsoproximal location. Other fragments were described as dorsal (5 [10%]), proximal (3 [6%]), dorsal nonarticular (2 [4%]), and dorsodistal articular, articular, axial, axial articular, and proximopalmar (1 [2%] each). Five (10%) osteochondral fragments had undescribed locations.
Twenty-four of the 47 (51%) horses with ACB fracture had concurrent osteoarthritis at the dorsal aspect of the affected carpus. In 16 of these horses, this concurrent carpal osteoarthritis would have been the only finding considered a clinically important radiographic abnormality given the predefined criteria. When osteoarthritic change in the carpus with the ACB fracture was removed from the definition of important radiographic abnormalities, there were 33 (70%) horses with ACB fracture but no other important radiographic abnormalities, whereas 68 of the 94 (72%) horses without ACB fracture were free of important radiographic abnormalities. These proportions were statistically similar (P = 0.79).
Overall, 54 of all 141 (38%) included horses had a diagnosis of mild tarsal osteoarthritis (11/47 [23%] with and 43/94 [46%] without ACB fracture). For 9 of 11 horses with ACB fracture and 33 of 43 horses without ACB fracture, this would have been the only clinically important lesion noted; therefore, given the predefined criteria, these horses were considered to have no radiographic abnormalities. Of the 47 horses with ACB fracture, 24 (51%) had concurrent carpal osteoarthritis at the dorsal aspect of the affected carpus and 23 (49%) had no such changes. For purposes of analyses, the carpus was viewed as a single unit; thus, horses were not separated on the basis of whether the concurrent carpal osteoarthritis was identified in the antebrachiocarpal joint (21/24), middle carpal joint (1/24), or both joints (2/24). Carpometacarpal joint disease was not reported for any horse. Of the horses with bilateral ACB fractures, one had a bilateral antebrachiocarpal joint osteoarthritic change and the other had unilateral antebrachiocarpal joint changes.
Performance
Of horses with ACB fracture as yearlings, 28% (13/47) started a race as a 2-year-old and 60% (28/47) started a race as a 3-year-old. Six (13%) horses with ACB fracture never started a race. Of horses without ACB fracture, 37% (35/94) started a race as a 2-year-old and 73% (69/94) started a race as a 3-year-old. The incidence of a horse starting at least 1 race during either of these years did not differ significantly between the 2 cohorts (Table 1). Additionally, 66% (31/47) of horses with ACB fracture and 79% (74/94) of horses without ACB fracture started a race at some point in their careers, and the difference in these proportions was not significant (Table 2). No significant differences were identified in the number of races started, earnings, or earnings per start as a 2- or 3-year-old between horses with and without ACB fracture.
Comparison of race performance as a 2- or 3-year-old between Thoroughbred yearlings with (n = 47) and without (94) osteochondral fragmentation of the ACB (ACB fracture) diagnosed on yearling sale repository radiographs.
| 2-year-old | 3-year-old | |||||
|---|---|---|---|---|---|---|
| Performance variable | With ACB fracture | Without ACB fracture | P value | With ACB fracture | Without ACB fracture | P value |
| No. (%) that started ≥ 1 race | 13 (28) | 35 (37) | 0.26* | 28 (60) | 69 (73) | 0.10* |
| Median (range) No. of starts, if raced | 3 (1–8) | 2 (1–8) | 0.17† | 5 (1–14) | 5 (1–20) | 0.66† |
| Median (range) annual earnings ($), if raced | 7,908 (0–157,267) | 9,318 (200–126,000) | 0.94‡ | 20,095 (480–195,711) | 15,303 (150–538,836) | 0.90‡ |
| Median (range) earnings per start ($) | 2,575 (0–22,467) | 3,106 (173–63,000) | 0.23‡ | 3,150 (120–27,958) | 2,509 (101–59,870) | 0.88‡ |
Value derived from the χ2 test of independence.
Value derived from a zero-inflated Poisson regression model adjusted for horse sex.
Value derived from a zero-inflated negative binomial regression model adjusted for horse sex.
Of the 47 horses with ACB fractures, 24 were included in the career performance analyses because they had successfully started a race but had not started a race within 6 months prior to the date of data collection. Of the 94 horses without ACB fractures, the same criteria resulted in 53 horses being included in the career performance analyses. Sixteen horses with ACB fracture and 20 without ACB fracture failed to start a race, and another 7 horses with ACB fracture and 21 without ACB fracture were excluded from career performance analyses because of having raced within 6 months prior to the date of data collection. These exclusions resulted in inclusion of fewer horses in the career data analyses than the number that raced as a 2- or 3-year-old. No significant differences were identified between cohorts in total career starts, earnings, or earnings per start (Table 2). However, a post hoc power analysis indicated that to have 80% power to detect a significant difference (α = 0.05) between the observed proportions of horses that raced, the exposed cohort would have required 136 horses and the unexposed cohort 272 horses.
Comparison of career race performance between Thoroughbreds with (n = 47) and without (94) ACB fracture diagnosed on yearling sale repository radiographs.
| Performance variable | With ACB fracture | Without ACB fracture | P value |
|---|---|---|---|
| No. (%) that started ≥ 1 race | 31 (66) | 74 (79) | 0.10* |
| Median (range) No. of starts | 13 (1–48) | 15 (2–43) | 0.62† |
| Median (range) earnings ($) | 23,967 (3,153–1,719,621) | 34,130 (680–2,015,893) | 0.53† |
| Median (range) earnings per start ($) | 3,095 (978–101,154) | 3,095 (171–95,995) | 0.20† |
Value derived from the χ2 test of independence.
Value derived from a linear regression model with a logarithmically transformed independent variable, adjusted for horse sex.
For number of starts, earnings, and earnings per start, horses that had never raced or were suspected to be actively racing at the time of data acquisition were excluded, thereby leaving 24 horses with ACB fracture and 53 horses without ACB fracture in those analyses.
Statistical analyses revealed no association between concurrent carpal osteoarthritis in horses with ACB fracture and any performance variable. To further isolate the effect of the ACB fracture, the racing performance of horses with ACB fracture and other clinically important radiographic lesions was compared with the performance of horses with ACB fracture and radiographic lesions limited to the fracture (with horses with concurrent osteoarthritis of the affected carpi excluded), again revealing no significant differences between groups (Table 3).
Comparison of race performance as a 2- and 3-year-old between Thoroughbreds with ACB fracture with versus without concurrent carpal osteoarthritis or with versus without other clinically important radiographic lesions.
| Performance variable | With carpal osteoarthritis (n = 24) | Without carpal osteoarthritis (n = 23) | P value | ACB fracture with other lesions (n = 14) | ACB fracture without other lesions (n = 17) | P value |
|---|---|---|---|---|---|---|
| 2-year-old | ||||||
| Median (range) No. of starts | 0 (0–7) | 0 (0–8) | 0.93* | 0 (0–6) | 0 (0–8) | 1.00* |
| No. (%) that started ≥ 1 race | 6 (25) | 7 (30) | 0.68† | 5 (36) | 4 (24) | 0.46† |
| Median (range) No. of starts, if raced | 3.5 (3–7) | 3 (1–8) | 0.38* | 3 (1–6) | 2.5 (1–8) | 0.95* |
| Median (range) earnings ($) | 10,938 | 7,908 | 0.20‡ | 7,908 | 6,854 | 0.08‡ |
| (1,189–157,267) | (0–14,480) | (0–32,760) | (4,600–11,252) | |||
| Median (range) earnings per start ($) | 3,057 | 2,475 | 0.34‡ | 1,789 | 2,697 | 0.49‡ |
| (396–22,467) | (0–4,827) | (0–6,552) | (1,407–4,600) | |||
| 3-year-old | ||||||
| Median (range) No. of starts | 2.5 (0–14) | 2 (0–10) | 0.63* | 2.5 (0–14) | 2 (0–10) | 0.47* |
| No. (%) that started ≥ 1 race | 14 (58) | 14 (61) | 0.86† | 10 (71) | 9 (53%) | 0.29† |
| Median (range) No. of starts, if raced | 5 (1–14) | 4 (2–10) | 0.08* | 6 (1–14) | 4 (2–10) | 0.50* |
| Median (range) earnings ($) | 22,635 | 15,408 | 0.91‡ | 15,408 | 33,876 | 0.29‡ |
| (732–195,711) | (480–170,433) | (2,765–195,711) | (480–170,433) | |||
| Median (range) earnings per start ($) | 3,507 | 2,844 | 0.61‡ | 2,658 | 6,775 | 0.17‡ |
| (366–27,958) | (120–20,300) | (1,128–27,958) | (120–20,300) | |||
| Career | ||||||
| Median (range) starts, if raced | 14 (1–48) | 12 (1–42) | 0.82§ | 8 (1–48) | 19 (1–42) | 0.20§ |
| Median (range) career earnings ($), if raced | 21,658 | 26,275 | 0.92§ | 21,170 | 164,262 | 0.18§ |
| (3,153–1,719,621) | (3,383–758,338) | (3,153–1,719,621) | (4,600–758,338) | |||
| Median (range) career earnings per start ($) | 3,037 | 3,903 | 0.96§ | 2,628 | 5,416 | 0.34§ |
| (1,547–101,154) | (978–18,158) | (1,128–101,154) | (978–18,158) | |||
Value derived from a zero-inflated Poisson regression model adjusted for horse sex.
Value derived from the χ2 test of independence.
Value derived from a zero-inflated negative binomial regression model adjusted for horse sex.
Value derived from a linear regression model with a logarithmically transformed independent variable, adjusted for horse sex.
Career data were unavailable for horses that were still racing or that had never started a race.
Discussion
Findings of the present study indicated that nonlame Thoroughbred yearlings with ACB fractures, specifically osteochondral fragmentation, diagnosed on prepurchase radiographs could achieve a level of race performance similar to their matched, unaffected counterparts, supporting our first hypothesis that there would be no difference between these 2 cohorts. No significant difference was identified in the likelihood of a horse starting a race or any other performance variable assessed, suggesting that although the ability to race successfully was influenced by no single factor, the identification of ACB fracture in yearling Thoroughbreds sold for racing did not appear to adversely impact their subsequent likelihood of racing.
The ACB is 1 of 7 carpal bones comprising the carpus. Although not directly weight bearing, the ACB acts as a sesamoid bone that serves as the attachment site for several short ligaments and the tendons of insertion of the flexor carpi ulnaris and ulnaris lateralis muscles. The ACB has dorsal articular surfaces that form a portion of the palmar aspect of the antebrachiocarpal joint, and the medial surface of the ACB is intimately involved in the carpal canal.11–13
Fractures of the ACB can differ in morphology, although those described herein were osteochondral fragments. The most common form of ACB fracture reported in the veterinary literature is a complete, mildly displaced fracture in the frontal plane. This type of ACB fracture is often diagnosed in jumping or steeplechase horses that are evaluated for lameness, commonly after a traumatic event.13,14 The horses included in the present study were those brought for routine prepurchase examination at public auction, and in accordance with yearling sales protocol, yearlings brought for sale are expected to be free of lameness at the walk.
Clinical outcomes of ACB fractures of various morphologies have been reported and include so-called carpal canal syndrome15 and carpal osteoarthritis. However, dorsoproximal fragments and vertical fractures can carry a good prognosis when treated conservatively, and secondary osteoarthritis of the antebrachiocarpal joint is seemingly uncommon.14 Although uncommon in sport horses following ACB fracture, the potential for development of antebrachiocarpal synovitis or osteoarthritis following osteochondral fragmentation of the ACB is a concern for Thoroughbred racehorse prospects.13,14 Osteochondral fragments arising from the palmar aspects of the cuboidal carpal bones in the antebrachiocarpal and middle carpal joints secondary to trauma can carry a poorer prognosis for return to function than dorsal carpal osteochondral fragments.16 This poorer prognosis is attributable to the development of carpal osteoarthritis or degenerative joint disease. At the time of sale, 23 of 47 (49%) yearlings with osteochondral fragmentation of the ACB in the present study had radiographic evidence of carpal osteoarthritis; however, its presence was not associated with poorer performance and thus our second hypothesis was rejected. However, results of previous studies3,e indicating poorer athletic performance in horses with (vs without) ACB fracture and concurrent antebrachiocarpal joint osteoarthrosis as well as in horses with (vs without) dorsal middle carpal disease alone illustrate the importance of carpal health in racehorses.
The results of the present study regarding similar racing performance of horses that have ACB fracture as yearlings with or without concurrent radiographic evidence of concurrent carpal osteoarthritis conflict with preliminary results in another study.e Those data indicate lower earnings per start and total earnings as a 2- and 3-year-old for yearlings with osteochondral fragments arising from the ACB with versus without corresponding dorsal antebrachiocarpal joint disease. That studye was based on repository radiographic evaluation of a similar group of yearlings from 2004 through 2007 and maternal sibling race records for comparison. Because both studies had limited statistical power given the low numbers of included horses, analysis of a larger cohort of horses with ACB fracture might yield different results. Furthermore, differences in the unexposed cohorts used in the 2 studies could have influenced the results. Recognizing that there is a conceivable association between fracture of the ACB with subsequent antebrachiocarpal arthritis and potential progression of disease once these yearlings enter race training, we believe our data indicated that in horses with osteochondral fragmentation of the ACB identified on prepurchase radiographs, the incidence of starting a race is not significantly lower than that for horses without such fractures.
Although no differences in race performance were identified in the present study between horses with ACB fracture identified on yearling repository radiographs and their unaffected counterparts, in contrast to previous findings,e we assert that this discrepancy is validating evidence that radiographic examination of a yearling has limited value in predicting its future athletic potential. In all prepurchase examinations, emphasis should be made that the findings indicate the horse's condition at the time of examination. Future racing performance depends on a multitude of influences beyond the veterinarian's control and unknown at the time of examination, including but not limited to individual athletic capability, future injury, training differences, owner investment, and surgical interventions.
Because of the retrospective nature of the present study, the cause of the identified fractures could not be established. Our reliance on radiographic interpretation reports prevented full characterization of the location of these fragments in some horses. Because the small sample size was recognized as a further limitation, horses with osteochondral fragments originating from the ACB were grouped together regardless of fragment location. Had the radiographs rather than the reports been available, more definitive descriptions could have been obtained; however, the original observers assessed the descriptions as adequate at the time of interpretation for the intended purpose, which was to assess the horse for radiographic findings that might be relevant to future racing or resale value.
The approach used in the present study contributed a further degree of separation between our data and previously reported preliminary data that were derived from analysis of horses with the specific diagnosis of dorsoproximal fragmentation of the ACB.e Small dorsoproximal osteochondral fragments from the ACB are often considered similar to osteochondrosis dissecans lesions or potentially avulsion fractures, whereas fractures in the frontal or horizontal planes are consistent with traumatic events.14 It is also common for ACB fractures to heal with fibrous or fibrocartilaginous unions with medical management, and this may result in observation of persistent fracture lines on radiographs obtained after healing has completed.17 However, the lack of medical history available for the horses included in the present study and the lack of lameness at the time of examination prevented the prognostic application of our findings to horses evaluated clinically for lameness associated with acute fracture of the ACB.
An additional limitation was that no follow-up medical information was available for the included horses to establish which, if any, of the ACB fractures required treatment after horses were sold. It is conceivable that some horses may have had these ACB fragments surgically removed or had other treatments to address the potential synovitis resulting from the presence of the osteochondral fragment.18,19
In the study reported here, horses without ACB fracture were not purposefully matched with those with osteochondral fragmentation of the ACB on the basis of sex; however, sex distributions were similar between these cohorts. At the Keeneland September sale of yearlings, catalog placement involves no accounting for horse sex because sire power and racing prowess of the dam carry more weight as analyzed by the Jockey Club. In the present study, the median distance in hip number (identification method denoting sale order and location in the sale catalog) between horses with ACB and their respective counterparts in the study population was 1, leading to the conclusion that in a sale that averaged 4,919 catalog entries annually during the study period, these horses were fairly closely matched. Our study was based on catalog entries, and no effort was made to assess price obtained in the sale ring nor whether horses were withdrawn from the sale, deferring to presale appraisal as the basis for determining peer relationships between horses with and without ACB fracture.
The prevalence of mild tarsal osteoarthritis in both cohorts of horses in the present study was higher than expected given findings of previous reports2,20,21 indicating a prevalence of 9.9% to 27% for tarsal osteoarthritis in yearling Thoroughbred race prospects; however, it should also be recognized that the statistical effect of this radiographic finding on subsequent racing performance was ascertained in only 1 previous study.2 Given the higher prevalence of this finding, compared with expected prevalence, as well as the higher prevalence in the nonexposed versus exposed cohorts in the present study, only the findings of moderate or severe distal tarsal osteoarthritis were included in the classification of clinically important radiographic lesions.
Reduction of bias and identification of confounders are common challenges in retrospective investigations. In the present study, which was based on radiographic interpretation reports, we believe the bias associated with the radiographic findings was minimal because none of the observers had knowledge of the study at the time of interpretation. We also believe that these reports were primarily generated for potential buyers, rather than consignors, which might have led to exaggeration of radiographic findings, if any bias existed. However, this bias would have been similar between cohorts. Additionally, although we recognize that substantial bias may have existed in the order of placement of these horses in the sale catalog, the inclusion of 2 horses without ACB fracture (1 horse on either side in the catalog) for each horse with ACB fracture would have minimized this bias.
The clinical importance of radiographic findings in young, untrained Thoroughbreds can be difficult to ascertain, despite multiple studies having been conducted to investigate correlations with future performance. Veterinarians are expected to make assessments and predictions of the importance of their examination findings on the basis of the available literature as well as their own experiences. We found no association between the radiographic identification of osteochondral fragmentation of the ACB in yearling Thoroughbreds and the incidence of starting a race as a 2- or 3-year-old in the present study.
Acknowledgments
The authors declare that there were no financial or ethical conflicts of interest.
The authors thank Drs. Patrick Worden, Michael Prichard, Jeffrey Berk, and Elizabeth Santschi for their contribution of medical records to the study.
ABBREVIATIONS
| ACB | Accessory carpal bone |
Footnotes
Drs. Michael Prichard, Elizabeth Santschi, and Jeffrey Berk, Equine Medical Associates PSC, Lexington, Ky.
Dr. Patrick Worden, Equine Medical Center of Ocala, Ocala, Fla.
Equibase [database online], Lexington, Ky: Equibase Company LLC, Lexington, Ky, 2014. Available at: www.equibase.com. Accessed Dec 21, 2014.
Stata SE, version 14.2, StataCorp, College Station, Tex.
Higgins JL, Spike-Pierce DL, Bramlage LR. Racing prognosis of Thoroughbred yearlings with dorsal proximal accessory carpal bone fragments (abstr), in Proceedings. 52nd Annu Conv Am Assoc Equine Pract 2010;402.
References
1. The Jockey Club. 2017 Fact Book. Available at: www.jockeyclub.com/default.asp?section=Resources&area=11. Accessed Jan 2, 2015.
2. Meagher DM. A review of equine malpractice claims, in Proceedings. 51st Annu Conv Am Assoc Equine Pract 2005;508–514.
3. Kane AJ, Park RD, McIlwraith CW, et al. Radiographic changes in Thoroughbred yearlings. Part 1: prevalence at the time of yearling sales. Equine Vet J 2003;35:354–365.
4. Kane AJ, McIlwraith CW, Park RD, et al. Radiographic changes in Thoroughbred yearlings. Part 2: associations with racing performance. Equine Vet J 2003;35:366–374.
5. Preston SA, Brown MP, Trumble TN, et al. Effects of various presale radiographic findings for yearling Thoroughbreds on 2-year-old racing performance. J Am Vet Med Assoc 2012;241:1505–1513.
6. Meagher DM, Bromberek JL, Meagher DT, et al. Prevalence of abnormal radiographic findings in 2-year-old Thoroughbreds at in-training sales and associations with racing performance. J Am Vet Med Assoc 2013;242:969–976.
7. Spike-Pierce DL, Bramlage LR. Correlation of racing performance with radiographic changes in the proximal sesamoid bones of 487 Thoroughbred yearlings. Equine Vet J 2003;35:350–353.
8. Cohen ND, Carter GK, Watkins JP, et al. Association of racing performance with specific abnormal radiographic findings in Thoroughbred yearlings sold in Texas. J Equine Vet Sci 2006;26:462–474.
9. Furniss C, Carstens A, van den Berg SS. Radiographic changes in Thoroughbred yearlings in South Africa. J S Afr Vet Assoc 2011;82:194–204.
10. Whitman JL, Prichard MA, Hance SJ, et al. Radiographic lucencies in the medial femoral condyle of Thoroughbred sale yearlings: a preliminary investigation of the effect on race records, in Proceedings. 52nd Annu Conv Am Assoc Equine Pract 2006;416–419.
11. Budras KD, Sack WO, Sabine R. Thoracic limb. In: Anatomy of the horse. 4th ed. Frankfurt, Germany: Manson Publishing, 2003;2–8.
12. Sisson S. The anatomy of domestic animals. 4th ed. Philadelphia: Saunders, 1955;92–96.
13. Minshall GJ, Wright IM. Frontal plane fractures of the accessory carpal bone and implications for the carpal sheath of the digital flexor tendons. Equine Vet J 2014;46:579–584.
14. Dyson SJ. Fractures of the accessory carpal bone. Equine Vet J 1990;2:188–190.
15. Radue P. Carpal tunnel syndrome due to fracture of the accessory carpal bone. Equine Pract 1981;3:8–16.
16. Getman LM, Southwood LL, Richardson DW. Palmar carpal osteochondral fragments in racehorses: 31 cases (1994–2004). J Am Vet Med Assoc 2006;228:1551–1558.
17. Barr ARS, Sinnott MJA, Denny HR. Fractures of the accessory carpal bone in the horse. Vet Rec 1990;126:432–434.
18. Cheetham J, Nixon AJ. Arthroscopic approaches to the palmar aspect of the equine carpus. Vet Surg 2006;35:227–231.
19. Goodrich LR, Nixon AJ. Medical treatment of osteoarthritis in the horse - a review. Vet J 2006;171:51–69.
20. Oliver LJ, Baird DK, Baird AN, et al. Prevalence and distribution of radiographically evident lesions on repository films in the hock and stifle joints of yearling Thoroughbred horses in New Zealand. N Z Vet J 2008;56:202–209.
21. Preston SA, Zimmel DN, Chmielewski TL, et al. Prevalence of various presale radiographic findings and association with sales price in Thoroughbred yearlings sold in Kentucky. J Am Vet Med Assoc 2010;236:440–445.
