Enostosis-like lesions in horses were first described as unusual focal or multifocal sclerotic lesions within the medullary cavity of long bones.1 Enostosis-like lesions can be differentiated from stress fractures by their intramedullary location, as determined either scintigraphically or radiographically by means of orthogonal views. In horses, ELLs have most commonly been diagnosed by means of nuclear scintigraphy and further characterized by means of radiography. Enostosis-like lesions have been identified as a cause of lameness in horses and caused lameness in 5 of the 10 horses in the original case series1; in each of those cases, the lameness subsequently resolved with rest. In a larger, more recent retrospective case series2 of 21 horses with ELLs, lameness was attributed to the ELLs in 15 (71%) cases. Furthermore, those authors concluded that ELLs had no long-term negative effect on athletic performance in the affected horses.2 Other case reports3,4 of horses with ELLs in various long bones have been published, but a study with a large number of horses containing further clinical and epidemiological information is required.
Our clinical impression was that ELLs in equids were more common than previously recognized and that, in recent years, the lameness attributable to ELLs has increased. The purpose of the study reported here was to retrospectively document the clinical and epidemiological features of horses with ELLs.
Materials and Methods
Criteria for selection of cases—Medical records of all equids admitted to the New Bolton Center of the University of Pennsylvania between March 1997 and March 2009 for bone scintigraphy were reviewed. All equids with a scintigraphic diagnosis of an ELL were included in the study. Data from horses and ponies that underwent a scintigraphic examination during that period and did not have an ELL were used for epidemiological comparison of breed, sex, and age between affected and unaffected equids. Data obtained from the medical records of equids with ELLs included signalment, lameness history, relevant clinical examination information (gait characteristics, lameness grade, limb or limbs affected, and results of diagnostic anesthetic evaluation if available), final diagnosis including determination of whether pain was associated with an ELL, location and number of ELLs, and radiographic characteristics if available. When management recommendations were included in the medical record, these were also reviewed. For racehorses, the interval from diagnosis to return to racing was calculated on the basis of race records. Outcome was assessed by evaluation of race records and telephone follow-up interviews of the owner or trainer of the Thoroughbred racehorses. Previous drug administration, previous or subsequent episodes of lameness, training track surface, and training techniques were queried in an attempt to find a common factor.
Scintigraphic examination—Nuclear scintigraphy was performed routinely as previously reported.1 Images were evaluated visually for anatomic location, intensity, and definition (diffuse or focal) of radiopharmaceutical uptake and graded with a modified version of a scoring system (Appendix 1).5 An ELL was diagnosed when focal or multifocal areas of increased radiopharmaceutical uptake were evident within a medullary cavity of a long bone on orthogonal views (Figure 1). Equids in which an area of increased radiopharmaceutical uptake could not be conclusively localized to a medullary cavity of a long bone and those with scintigraphic and radiographic findings consistent with a stress fracture were excluded from the study.
Representative paired orthogonal scintigraphic images of right humeral (A), left femoral (B), and right tibial (C) ELLs in 3 horses. Lateral views are on the left and cranial (humeral) and caudal (humeral and femoral) views are on the right. LH = Left hind limb. RH = Right hind limb.
Citation: Journal of the American Veterinary Medical Association 245, 9; 10.2460/javma.245.9.1042
Radiographic examination—Radiography was performed at the discretion of the attending clinician. When available, images of orthogonal radiographic views were retrieved and abnormalities were scored in a blinded manner on the basis of a modified version of a scoring system (Appendix 2).5
Lameness examination—All equids that underwent scintigraphy were examined for lameness and received complete routine musculoskeletal palpation. Lameness was evaluated at the time of scintigraphic examination when the equids were walking and trotting in hand and was graded on a scale of 0 (not lame) to 5 (non–weight bearing).6 Flexion and manipulation tests as well as intra-articular or perineural anesthesia were performed on a selected basis. When possible, the source of pain was localized on the basis of findings of diagnostic anesthetic testing. In numerous horses, ELLs involved sites from which signs of pain were difficult or impossible to abolish by use of routine diagnostic anesthetic testing, and signs of pain in these horses were ascribed to the site of an ELL if other potential sources of pain had been ruled out on the basis of all available clinical and imaging information. For purposes of statistical analysis, the cause of lameness was considered to be ELL-associated pain when lameness was identified in the ELL-affected limb and other sources of pain were either ruled out or not detected. A small number of horses were admitted to the hospital on separate occasions, each time with lameness in a different limb. Each admission was evaluated separately and included in the analysis.
Race records—Race records were obtained from Bloodstock Research Information Services, Lexington, Ky, and the US Trotting Association, Columbus, Ohio, for Thoroughbred and Standardbred racehorses, respectively. Information was collected pertaining to whether the horses had raced before and after ELL diagnosis. The date of the first start after the diagnosis of an ELL had been made was noted.
Statistical analysis—Multivariable logistic regression analyses were used to evaluate for potential relationships of age, breed, sex (female vs gelding vs sexually intact male), anatomic location of the ELL, lameness grade, and racing performance with each other. Lameness scores were modeled on the basis of Somers D evaluations. Kruskal-Wallis tests for nonparametric data analysis were used when appropriate. Values of P < 0.05 were considered significant. All analyses were performed with commercially available statistical software.a Data are reported as mean ± SD throughout unless otherwise stated.
Results
During the study period, 4,992 horses and ponies underwent bone scintigraphic examinations and ELLs were diagnosed in 79 (1.6%) equids. In 4 of the 79 (5.1%) animals, anatomically distinct ELLs were diagnosed on more than 1 occasion, causing lameness in different limbs. In 3 Thoroughbreds, a diagnosis of an ELL was made on 2 occasions, and in a warmblood, a diagnosis of an ELL was made on 4 occasions. Thus, there were 85 ELL-related admissions to the hospital (all of which were included in the epidemiological analysis).
Of the 85 ELL-related admissions, 55 (64.7%) involved Thoroughbreds, 19 (22.4%) involved warmbloods, and 6 (7.1%) involved Standardbreds; the remaining 5 (5.9%) admissions involved 2 Hanoverians, 2 Quarter Horses, and a pony. Among the 79 affected equids, the mean age was 5.6 years (range, 2 to 14 years) for Thoroughbreds, 10.5 years (range, 6 to 16 years) for Standardbreds, 6 years (range, 4 to 7 years) for warmbloods, and 5.4 years (range, 3 to 9 years) for the remaining 4 equids. Of the 78 horses and 1 pony with ELLs, 23 (29.1%) were mares, 38 (48.1%) were geldings, and 18 (22.8%) were stallions.
On the basis of the general population of equids undergoing scintigraphic examination (excluding horses with ELLs, n = 4,913) during the study period, Thoroughbreds were more likely (P = 0.021) and Standardbreds were less likely (P = 0.024) to have ELLs (Figure 2). Thoroughbreds (P = 0.048), Standardbreds (P = 0.049), and warmbloods (P = 0.001) with ELLs were older than the horses of those breeds without ELLs (Figure 3). There was no significant association between sex and diagnosis of an ELL.
Breed distribution of 79 equids with detectable ELLs (gray bars), compared with the distribution of breeds among equids that underwent bone scintigraphy (ie, 4,913 horses excluding those with ELLs [white bars]), at a veterinary teaching hospital during the period March 1997 to March 2009. Numbers in each bar represent the number of equids. Within each breed grouping, a horizontal bar denotes a significant (P < 0.05) difference.
Citation: Journal of the American Veterinary Medical Association 245, 9; 10.2460/javma.245.9.1042
Mean ± SD age of Thoroughbred, warmblood, and Standardbred horses with detectable ELLs (n = 79 [gray bars]), compared with the total population of equids that underwent bone scintigraphy (excluding equids with ELLs, n = 4,913 [white bars]), at a veterinary teaching hospital during the period March 1997 to March 2009. Numbers in each bar represent the number of equids. Within each paired gouping, a horizontal bar denotes a significant (P < 0.05) difference between categories.
Citation: Journal of the American Veterinary Medical Association 245, 9; 10.2460/javma.245.9.1042
Significantly more horses with ELLs were documented in 2008 and 2009, compared with findings for previous years. The breed, age, and sex distributions of equids with an ELL in 2008 and 2009 were not significantly different from the distributions in previous years.
Among the 79 equids (85 ELL-related admissions), 157 ELLs were diagnosed. There was no difference in the distribution of ELLs between right and left limbs in both the fore- and hind limbs, and as such data were combined (eg, the left and right radial ELLs were combined as the ELLS in a radial location for analysis). Additionally, there was no difference in the distribution of ELLs between the third metacarpal and third metatarsal bones; those data were also combined for analysis.
Increasing intensity of radiopharmaceutical uptake within a medullary cavity of a long bone was associated with increasing severity of lameness (P < 0.001). Diffuse radiopharmaceutical uptake was associated with a greater lameness grade, compared with that associated with focal uptake (P = 0.017). The humerus was significantly (P < 0.05) more likely to have greater radiopharmaceutical uptake than the tibia. However, there was no significant association between the anatomic location of the ELL and the intensity of radiopharmaceutical uptake. Radiographic score was not significantly correlated with intensity of radiopharmaceutical uptake or lameness grade.
The distribution of the 157 ELLs differed significantly among the affected bones. The most common anatomic locations for ELLs were the tibia (62 ELLs) and the radius (46 ELLs); the number of ELLs did not differ significantly between these 2 locations. Significantly fewer ELLs were detected in the third metacarpal and third metatarsal bones (25 ELLs) and the humerus (20 ELLs); the number of ELLs did not differ between these 2 locations. The significantly least likely location was the femur (4 ELLs).
In 42 of the 85 (49.4%) admissions, lameness was attributed to a scintigraphically evident ELL by a thorough process of exclusion. Likelihood of lameness associated with an ELL did not differ (P = 0.769) among breeds. Enostosis-like lesions in the radius and tibia were significantly (P < 0.001) less likely to cause lameness, compared with those in the femur or humerus (Figure 4). Lameness associated with an ELL was significantly (P < 0.05) more severe in the humerus (mean ± SD score, 1.62 ± 0.26) than in the third metacarpal and third metatarsal bones (mean score, 0.5 ± 0.24), radius (mean score, 0.38 ± 0.12), or tibia (mean score, 0.47 ± 0.14). Severity of lameness involving the femur (mean score, 1.5 ± 0.29) was similar to severity of lameness involving the humerus; however, lameness involving the femur did not differ significantly from that associated with an ELL at other locations owing to the small number of femur-located lesions (n = 4).
Percentage of 157 ELLs that did (black bars) or did not (gray bars) cause lameness in 85 ELL-related equine admissions to a veterinary teaching hospital during the period March 1997 to March 2009. a,bValues with different letters are significantly (P < 0.05) different. McIII+MtIII = Third metacarpal and third metatarsal bones.
Citation: Journal of the American Veterinary Medical Association 245, 9; 10.2460/javma.245.9.1042
As the number of ELLs detected at a single admission increased, the likelihood of a racehorse returning to racing after diagnosis decreased significantly. Racehorses with a single ELL were 3.9 (P = 0.005) and 8.4 (P = 0.001) times as likely to race after diagnosis as were horses with 2 or 3 ELLs at admission, respectively. Of the 3 Thoroughbreds with 2 ELLs diagnosed on 2 occasions, 2 were racehorses and both raced after the first ELL diagnosis. After the second diagnosis, one returned to racing briefly and the other did not.
Thirty-seven of the 52 (71.2%) Thoroughbreds and all 6 Standardbreds had race records available for review. Twenty-three of the 37 (62.2%) Thoroughbreds and 5 of the 6 Standardbreds returned to racing. Of the 43 racehorses, 28 (65.1%) returned to racing and 15 (34.9%) did not after ELL diagnosis. Seven of the 37 (18.9%) Thoroughbreds had not raced prior to ELL diagnosis, but all 7 horses subsequently raced
The mean recommended rest period for racehorses after diagnosis of an ELL was 83.7 ± 7.2 days (mode, 120 days). The mean time from ELL diagnosis to return to racing was 184.3 ± 21.7 days. The most commonly recommended rehabilitation program was 4 weeks of strict stall rest, followed by 4 weeks of stall rest with hand walking before 8 weeks of paddock turnout with subsequent return to training. The mean recommended rest period for warmbloods was 56.8 ± 12.9 days.
Of the 56 owners and trainers of Thoroughbred racehorses, 26 (46%) were able to be contacted for telephone interviews but no common history was evident to determine a predisposing cause for the ELLs, especially in the later years.
Discussion
In the present study, ELLs were found to be uncommon, diagnosed in < 2% of equids undergoing scintigraphic examination over a 12-year period, and found to cause lameness in approximately 50% of equids in which an ELL was diagnosed. This is similar to preliminary findings published previously.1 Enostosis-like lesions caused pronounced, recurrent lameness within the same animal but in different locations, suggesting that once ELL-associated lameness resolves, horses and ponies are not necessarily at risk for recurrence of an ELL at that site but are at risk for lameness from the same condition at a distant site. From the study data obtained, we were not able to determine the cause or risk factors for development of ELLs, but given their intramedullary and disparate locations, ELLs are unlikely to develop as a result of factors related to conformation or other anatomic predisposing factors and are clearly a clinical entity different from that of stress fractures.
In the present study, an ELL diagnosis was made in a significantly greater percentage of equids in 2008 and 2009, confirming our suspicions of an increased incidence, yet no readily apparent reason could be established. Equids in which ELLs were diagnosed in 2008 and 2009 were not significantly different from other horses included in this study with respect to age, breed, sex, or anatomic distribution of ELLs. Follow-up communication with owners and trainers failed to identify a common factor within this subgroup of animals. Because the imaging process and personnel involved or interpretation of the images during this period was not altered, it was unlikely to be an operator-dependent change. Anecdotally, in our practice area, the increased incidence of ELLs prompted speculation that there may be an association between ELLs in Thoroughbred racehorses and newly applied synthetic track surfaces, but we were unable to establish this causal relationship; furthermore, in 2010, the percentage of ELLs diagnosed was similar to the percentages recorded in the years prior to 2008 (unpublished data).
Among all equids undergoing bone scintigraphic examination during the period of the present study, Thoroughbreds were more likely and Standardbreds less likely to develop ELLs. Compared with Standardbreds, Thoroughbreds are exposed to higher concussive loading during training and have a higher incidence of stress fractures and nonadaptive bone disease. Given the Thoroughbred predisposition to ELLs in the present study, it may be that high concussive loads predispose racehorses to the formation of ELLs. As a result, training modifications to reduce concussion may reduce the incidence of this disease or be useful when developing treatment protocols. Examination of bone biopsy specimens from equids with ELLs and use of advanced imaging methods such as CT and MRI in affected animals would likely help to establish the etiopathogenesis of ELLs. We are aware of a warmblood with right forelimb lameness for which examination of a bone biopsy specimen from a distal humeral ELL revealed changes in the medullary cavity similar to those in horses with stress fractures of adjacent cortical bone.b Bone reacts similarly to numerous insults, so a stress-related intramedullary event would be difficult to differentiate.
In the equids included in the present study, lesions were detected in the tibia and radius significantly more commonly than they were in the humerus and the third metacarpal and third metatarsal bones. Lesions were least common in the femur. As has been previously reported for horses with ELLs, localizing pain can be difficult because the more proximal limb sites with ELLs, such as the mid to distal diaphysis of the humerus and midfemoral diaphysis, are locations not amenable to nerve blocks.7 In the study reported here, lameness was only attributed to an ELL when reasonable efforts had been made to localize the lameness by means of diagnostic anesthetic testing. Lameness was more severe in equids with an ELL in the humerus than in those with ELLs in other bones. Equids in which an ELL was diagnosed in a femur had pronounced lameness, nearly equal in lameness score to those with humeral ELLs, but lameness score was not significantly different from that found in horses with ELLs in other bone locations. The reason for these anatomic predilections and associated differences in lameness is unknown. Interestingly, the intensity of radiopharmaceutical uptake evident in the scintigraphic examinations was related to the degree of lameness. Severe lameness associated with lesions of high intense scintigraphic intensity may indicate a more acute stage of the disease process. As these lesions become less distinct, the horses became increasingly less lame. Radiographic findings were not associated with lameness or the scintigraphic intensity of an ELL. This is likely because radiography is not a sensitive method for detection of acute bone injury; often, a period of bone remodeling is required before lesions become radiographically evident.
Stress fractures are a common cause of lameness in equids and can be mistaken for ELLs. Tibial stress fractures occur predominantly in 2-year-old horses (44/62 [71%] horses), and humeral stress fractures occur predominantly in 3-year-old horses (21/46 [46%]).8 Furthermore, 43 of 61 (70%) horses with tibial stress fractures had not been raced at the time of diagnosis.8 Comparatively, in the present study, ELLs occurred more commonly in older Thoroughbreds (5.6 years), and 30 of 37 (81%) Thoroughbreds for which there were race records had raced prior to ELL diagnosis. In another study,8 49 of 61 (80%) horses with tibial stress fractures and 30 of 39 (77%) horses with humeral stress fractures returned to racing after a mean interval from the time of diagnosis of 225 days. In comparison, fewer horses with ELLs returned to racing (28/43 [65.1%] horses) and the period from the time of diagnosis to return to racing was slightly shorter (184 days) in the present study. Seven Thoroughbreds had ELLs that caused lameness while still in training, prior to commencement of racing. Although those 7 Thoroughbreds all proceeded to race, they had significantly fewer race starts than did horses that had raced before ELL diagnosis. On the basis of these results, which are in contrast to findings of a previous case series of 21 horses with ELLs reported by O'Neill and Bladon,2 we concluded that ELLs likely had a negative influence on racing longevity.
The likelihood of return to racing decreased as the number of ELLs diagnosed at a single admission increased. This apparent association may have been due to horses being retired early, having prolonged rehabilitation periods, or having recurrent undiagnosed lameness. Without follow-up clinical and scintigraphic examinations, it is uncertain whether the lower rate of return to racing was a result of development of new ELLs, persistence of the primary lesion, or unrelated causes. Regardless, if horses develop an ELL while still in training and before commencement of racing or have numerous ELLs diagnosed simultaneously, the prognosis for future racing is less favorable.
Equids with ELLs were older than the general population of equids undergoing bone scintigraphy in the present study. This finding differs substantially from what has been reported for young dogs with similar intramedullary lesions associated with panosteitis.9 This suggests that the underlying pathogenesis of ELLs in equids likely differs from that of panosteitis in dogs.
In the present study, the recommended rehabilitation period before return to training was more likely to be longer for racehorses with ELLs that had higher lameness scores than for racehorses with an ELL that was un-associated with lameness or was an incidental finding. It is not known whether owners and trainers elected to give the more severely lame horses more prolonged rest because of the magnitude of the initial degree of pain, concern about recurrence of lameness if horses were to return to training prematurely, or persistence of clinical signs.
The data obtained from a large group of horses with ELLs in the present study indicated that ELLs were more common in older Thoroughbreds and lameness was directly attributable to a scintigraphically evident ELL in approximately half of the admissions. The most common anatomic locations for ELLs were the tibia and the radius, yet humeral ELLs caused more lameness than did ELLs in other locations. The intensity of radiopharmaceutical uptake in the ELLs during scintigraphic examinations mirrored the severity of lameness. Racehorses with > 1 ELL found during a single scintigraphic examination were less likely to race after diagnosis than those with a solitary ELL. The cause of ELLs in horses is still unknown; however, information from this retrospective study has provided greater understanding of factors associated with its epidemiology and may enable clinicians to more accurately determine prognosis for affected equids.
ABBREVIATION
| ELL | Enostosis-like lesion |
Stata, version 13, StataCorp, College Station, Tex.
Boswell RP, Equine Sports Medicine and Diagnostic Imaging, West Palm Beach, Fla: Personal communication, 2001.
References
1. Bassage LH, Ross MW. Enostosis-like lesions in the long bones of 10 horses: scintigraphic and radiographic features. Equine Vet J 1998; 30: 35–42.
2. O'Neill HD, Bladon BM. Retrospective study of scintigraphic and radiological findings in 21 cases of enostosis-like lesions in horses. Vet Rec 2011; 168: 326–329.
3. Stieger-Vanegas SM, Kippenes-Skogmo H, Nilsson E. Imaging diagnosis—enostosis-like lesion in the femur of a horse. Vet Radiol Ultrasound 2009; 50: 509–512.
4. Fenger CK, Bertone JJ, Biller D, et al. Generalized medullary infarction of the long bones in a horse. J Am Vet Med Assoc 1993; 202: 621–623.
5. Ramzan PH, Newton JR, Shepherd MC, et al. The application of a scintigraphic grading system to equine tibial stress fractures: 42 cases. Equine Vet J 2003; 35: 382–388.
6. Ross M. Movement. In: Ross M, Dyson S, eds. Lameness in the horse. St Louis: Elsevier, 2003;60.
7. Ramzan P. Equine enostosis-like lesions: 12 cases. Equine Vet Educ 2002; 14: 143–148.
8. O'Sullivan CB, Lumsden JM. Stress fractures of the tibia and humerus in Thoroughbred racehorses: 99 cases (1992–2000). J Am Vet Med Assoc 2003; 222: 491–498.
9. Demko J, McLaughlin R. Developmental orthopedic disease. Vet Clin North Am Small Anim Pract 2005; 35: 1111–1135.
Appendix 1
System used to grade the severity of ELLs in horses with a scintigraphic diagnosis of an ELL.
| Grade | Radiopharmaceutical intake |
|---|---|
| 1 | Mild |
| a | Focal |
| b | Diffuse |
| 2 | Moderate |
| a | Focal |
| b | Diffuse |
| 3 | Intense |
| a | Focal |
| b | Diffuse |
(Adapted from Ramzan P, Newton J, Shepherd M, et al. The application of a scintigraphic grading system to equine tibial stress fractures: 42 cases. Equine Vet J 2003;35:382–388.)
Appendix 2
System used to grade the radiographic severity of ELLs in horses with a scintigraphic diagnosis of an ELL.
| Grade | Description |
|---|---|
| 1 | Subtle radiographic abnormalities |
| 2 | Moderate radiographic abnormalities typical of an ELL |
| 3 | Marked radiographic abnormalities typical of an ELL |
(Adapted from Ramzan P, Newton J, Shepherd M, et al. The application of a scintigraphic grading system to equine tibial stress fractures: 42 cases. Equine Vet J 2003;35:382–388.)
