Abstract
Objective—To define scintigraphic, physical examination, and scapular ultrasonographic findings consistent with bone fragility syndrome (BFS) in horses; develop indices of BFS severity; and assess accuracy of physical examination, scapular ultrasonography, and serum biomarkers for BFS diagnosis.
Design—Prospective case-control study.
Animals—48 horses (20 horses with BFS and 28 control horses).
Procedures—Horses underwent forelimb scintigraphic evaluation, physical examination, scapular ultrasonography, and serum collection. Scintigraphy was used as a reference standard to which physical examination, scapular ultrasonography, and concentrations of serum biomarkers (carboxy-terminal telopeptide of collagen crosslinks and bone-specific alkaline phosphatase activity) were compared for assessing accuracy in BFS diagnosis.
Results—A diagnosis of BFS was strongly supported on scintigraphy by ≥ 2 regions of increased radiopharmaceutical uptake, including 1 region in the scapular spine and 1 region in the scapular body or ribs; on physical examination by lateral bowing of the scapulae; and on ultrasonography by widening of the scapular spine. None of the tests evaluated were accurate enough to replace scintigraphy for mild disease; however, physical examination and scapular ultrasonography were accurate in horses with moderate to severe BFS. Serum biomarkers were not accurate for BFS diagnosis.
Conclusions and Clinical Relevance—Scintigraphy remained the most informative diagnostic modality for BFS, providing insight into disease severity and distribution; however, physical examination and scapular ultrasonographic abnormalities were diagnostic in horses with moderate to severe disease. Proposed severity indices classified the spectrum of disease manifestations. Clearly defined criteria for interpretation of diagnostic tests aid in the detection of BFS. Severity indices may be useful for assessing disease progression and response to treatment.
Equine BFS (also known as bone fragility disorder or silicate-associated osteoporosis) is a progressive, debilitating skeletal disease in which affected horses develop systemic osteoporosis.1,2,a Clinical signs are variable and often include body stiffness, acute or chronic and intermittent or continuous lameness, cervical stiffness or reduced range of motion, nontraumatic fractures of the axial and proximal appendicular skeleton, and skeletal abnormalities that include lateral bowing of 1 or both scapulae and severe lordosis.1,3,a Severe disease often culminates in catastrophic fracture or euthanasia. The time course from onset of subclinical and clinical disease to catastrophic fracture or euthanasia is unknown. The full pathophysiologic mechanism of BFS is unknown; however, recent characterization of BFS indicated high bone turnover with high numbers of abnormal osteoclasts.2,a Additionally, in 1 study,2 85% of horses with BFS had concurrent pulmonary silicosis, and the 2 diseases are hypothesized to be associated with each other.3,4,a Treatment for BFS includes pain management with analgesic medication or experimental injection of bisphosphonates. Zoledronic acid treatment has been associated with improvement in clinical signs.5 Early detection of disease might improve response to treatment.
Horses with BFS have historically been detected in geographic clusters in northern California1,3,a; however, affected horses have also been identified in Oregon, Texas, Virginia, Illinois, and Kentucky.a Because of the strong association between BFS and pulmonary silicosis,2,a a history of living on a BFS-endemic farm, in a BFS-endemic region, or in a region with a high incidence of pulmonary silicosis should increase the index of suspicion of BFS in horses with clinical signs compatible with BFS. Additionally, given that the cytotoxic crystals associated with pulmonary silicosis are found worldwide, this disease may be more widespread than reported.
Diagnosis of BFS can be problematic, especially at early stages of disease. Currently, scintigraphic examination is the most accurate antemortem test available and is considered the reference standard. Positive findings include foci of IRU in multiple bones of the axial and proximal portions of the appendicular skeleton where the degree of uptake is associated with the rate of new bone formation.6 Typical sites include the scapula, ribs, cervical portion of the vertebral column, and pelvis.1 Lateral (outward) bowing of the scapula on physical examination, widening of the scapular spine with an irregular surface contour on scapular palpation and ultrasonography, and mixed lytic and productive lesions on cervical radiographs have been described in horses with BFS.1–4,a
The polyostotic nature of scintigraphic findings and the apparent high degree of radionuclide uptake indicate that BFS is a systemic, high–bone-turnover disease. Bone turnover can be quantified by measurement of the concentration of serum biomarkers of bone formation and resorption.7 Serum biomarkers provide a dynamic assessment of current and total metabolic bone activity, are minimally invasive, and can be performed sequentially without bias.8 In horses, bALP and CTX-1 have been suggested as useful markers of bone formation and resorption, respectively.8
Although scintigraphic, physical examination, and scapular ultrasonographic abnormalities have been described for BFS in horses,1,3,4 appropriate diagnostic criteria need to be clearly defined. There is also a need to assess the diagnostic accuracy of less expensive and more readily available tests such as physical examination, scapular ultrasonography, and measurement of the concentration of serum biomarkers of bone turnover. Finally, establishment of a scale of clinical findings associated with disease severity would aid in monitoring disease progression and response to treatment as well as in the conduct of future clinical studies. The objectives of the study reported here were to define scintigraphic, physical examination, and scapular ultrasonographic findings consistent with BFS in horses; develop indices of disease severity for scintigraphic and physical examination findings; and assess accuracy of physical examination, scapular ultrasonography, and serum biomarkers of bone turnover for a diagnosis of BFS.
Materials and Methods
Study design—A prospective case-control study was used to evaluate the diagnostic accuracy of physical examination, scapular ultrasonography, and measurement of the concentration of serum biomarkers of bone turnover to detect BFS in horses, with scintigraphic findings as the reference standard. The study was designed to include 20 BFS-affected and 40 unaffected (control) horses. Results of standardized examinations (scintigraphy, physical examination, and scapular ultrasonography) and serum biomarker concentrations were compared between horses with BFS and control horses. Each examination (physical, ultrasonographic, and scintigraphic) was conducted by different investigators (AMA and RB, MBW, and SMP, respectively) blinded to the results of other diagnostic examinations. Serum biomarker analysis (performed by AMA) and scintigraphy analysis (performed by SMA and AMA) were conducted with a separate, randomized identification number that blinded the investigators to the true disease status of the horse.
The study protocol was reviewed and approved by the Institutional Animal Use and Care Committee of the University of California-Davis. Owner consent was obtained for each horse that participated in the study.
Horses—Beginning May 15, 2008, all horses that were referred to the William R. Pritchard Veterinary Medical Teaching Hospital of the University of California-Davis for forelimb scintigraphy were initially included in the study. Because the veterinary medical teaching hospital is not located in a BFS-endemic region, horses suspected to have BFS were also recruited through referral veterinarians from known endemic regions including Monterey, Sonoma, and Napa counties in California. Recruitment criteria included clinical signs suggestive of a BFS diagnosis (eg, history of shifting leg lameness, appreciable lateral [outward] bowing of the scapula, or rapid onset of lordosis) or origination from a farm in which BFS was considered endemic. Enrollment concluded January 15, 2010, when the target numbers of case and control horses were obtained. Thirty-nine horses underwent scintigraphy as part of diagnostic testing of forelimb lameness, 14 underwent scintigraphy for diagnostic testing for BFS because of suggestive clinical signs, and 7 underwent scintigraphy because of a history of living on a BFS-endemic farm. A total of 60 horses, therefore, underwent scintigraphy.
Scintigraphy—Horses were administered 99mTc-MDP (5,550 to 6,100 MBq, IV) 4 to 5 hours before scintigraphy to maximize skeletal uptake in older horses, furosemide (200 mg, IV) 3 hours after 99mTc-MDP administration to aid in soft tissue clearance of 99mTc-MDP, and detomidine hydrochloride (0.01 to 0.03 mg/kg [0.005 to 0.014 mg/lb], IV) and butorphanol tartrate (0.01 to 0.02 mg/kg [0.005 to 0.009 mg/lb], IV) for sedation. Images in a 256 × 256-pixel matrix were obtained for 1 minute or up to a minimum of 100,000 counts with a rectangular, large field-of-view gamma camerab with a low-energy, all-purpose, parallel-hole collimator mounted on a custom-made, central-column movable gantry.c Horses were positioned squarely to attempt to distribute weight on all 4 limbs during image acquisition. For purposes of the present study, bilateral views of interest included the forelimbs, cervical portion of the vertebral column, and ribs.
Scintigraphic lesions were categorized, and intensity was quantitated.d The number, character, and location (bone) of lesions were recorded. Uptake in the scapulae was qualitatively assessed as mild, moderate, or severe. Intensity (CPP) was recorded for the brightest lesion, a rectangular region of interest that isolated the width of the central third of the right scapular spine, and an oval region of interest of approximately 300 pixels (mean ± SD, 282 ± 43 pixels) centered at the point of greatest curvature of the diaphysis in the middle of the right radius. Intensity was calibrated by total counts per scan with the following formula:
Intensity was also adjusted (adjusted CPP) for general skeletal uptake with the following formula:
Scintigraphic diagnostic criteria and SSI—Scintigraphic images from all 60 horses were evaluated, and a diagnosis of BFS was made subjectively on the basis of the presence of multiple regions of IRU in the axial skeleton and proximal part of the forelimb. After categorization of horses as case or control horses, scintigraphic images from cases were further assessed for specific locations of IRU, shape of scapulae and ribs, and presence or absence of photopenia (ie, decreased radiopharmaceutical uptake giving the appearance of lack of bone) of the caudal cervical vertebrae (Figure 1).
Scintigraphic images of the scapula, ribs, and cervical vertebrae in various horses without (A, B, and G) and with BFS (C, D, E, F, and H). Normal scintigraphic appearance of the scapula (A) and rib cage (B). Areas of IRU are evident along the scapular spine (C) and in several ribs (D). Altered curvature of the caudal margin of the scapula and narrowing of the infraspinatus fossa are indicative of lateral bowing of the scapula (E). Lateral bowing of the rib cage is evidenced by altered orientation of the ribs (F). Comparison of normal uptake of radiopharmaceutical (G) with photopenia in the caudal cervical vertebrae (H).
Citation: Journal of the American Veterinary Medical Association 242, 1; 10.2460/javma.242.1.76
The prevalence of predominant scintigraphic features among the cases was tabulated. This tabulation resulted in groups of horses with visible scintigraphic lesions upon which the SSI (0 to 5) was developed. Each SSI score was defined by the presence of a particular feature and included all features of the lower scores. Prevalence of features was presumed to be related to disease severity such that the most prevalent feature (IRU in the spine of the scapula) was the first to develop and the least prevalent feature (photopenia of the caudal cervical vertebrae) was the last to develop. Horses were assigned SSI scores on the basis of the presence of predominant scintigraphic features as follows: 0 = horse with none of the specific scintigraphic findings; 1 = horse with IRU in the scapulae; 2 = horse with IRU in the scapulae and ribs; 3 = horse with IRU in the scapulae and ribs and the skeletal deformity of lateral (outward) bowing of the scapulae; 4 = horse with IRU in the scapulae and ribs and skeletal deformities of lateral (outward) bowing of the scapulae and rib cage; and 5 = horse with IRU in the scapulae and ribs, skeletal deformities of lateral (outward) bowing of the scapulae and rib cage, and photopenia evident in the caudal cervical vertebrae.
Physical examination and severity index—The standardized physical examination included palpation of scapular spines, assessment of range of motion in the neck, subjective assessment of lordosis, and lameness assessment (scaled 0 to 5)9 with horses traveling in a straight line and on a 20-m-diameter circle. The standardized physical examination was performed the day before the scintigraphic examination on 32 horses (20 horses with BFS and 12 control horses). Some control horses did not receive the standardized physical examination because of scheduling errors.
Similar to the development of the SSI, the prevalence of predominant physical features among case horses, which included lameness at the time of examination, presence or absence of signs of pain on scapular spinal palpation, lateral bowing of the scapula, and severe lordosis, was tabulated. This tabulation also resulted in horses with visible lesions upon which the PSI (0 to 4) was developed. Each PSI score was defined by the presence of a particular feature and included all features of the lower scores. Prevalence of features was again presumed to be related to disease severity such that the most prevalent features (lameness and signs of scapular spinal pain on palpation) developed earlier in the disease process than did the least prevalent features (lateral bowing of the scapula or lordosis). Horses were assigned PSI scores on the basis of the presence of specific physical examination findings as follows: 0 = horse with none of the specific physical examination findings; 1 = horse with either lameness or signs of pain on palpation of scapular spine, but not both; 2 = horse with lameness and signs of pain on palpation of the scapular spine; 3 = horses with lameness, signs of pain on palpation of the scapular spine, and lateral (outward) bowing of the scapulae; and 4 = horses with lameness, signs of pain on palpation of the scapular spine, lateral (outward) bowing of the scapulae, and severe lordosis.
Ultrasonography—Ultrasonographic evaluation was conducted the day after scintigraphic examination and included assessment of both scapulae via transverse and longitudinal imaging techniques with a 10- to 14-MHz linear transducer or a 4- to 8-MHz curvilinear transducere and a scanning depth of 4 to 7 cm. Initial survey scans of each scapular spine were performed to detect areas of widening, abnormal bone surface contours, and any associated overlying soft tissue abnormalities. Representative transverse images were then obtained from the proximal, middle, and distal thirds of each scapular spine for 6 images/horse (Figure 2). Care was taken to avoid images of the tuber spinae scapulae in the proximal region of the scapular spine and to acquire images at the location deemed maximally affected in all abnormal regions. The exact location of each image was recorded as the distance (cm) dorsal to the cranial eminence of the greater tuberosity of the humerus (point of the shoulder). Qualitative assessment of each scapular spinal region was recorded as wide (abnormal) or narrow (normal). Quantitative measurements of scapular spinal widths were taken at the time of image acquisition with digital calipers and the calculation package within the ultrasound machine.e Bone surface abnormalities were characterized as a focal (< 1-cm length of irregular cortical surface) or regional (surface irregularities extended throughout most of a region). Bone contours of the scapular spine were further described as having a tufted, smooth, or spiculated appearance at each level of the scapulae (Figure 3). The tufted lesion pattern was further described as a hypoechoic expansion of the normally smooth and hyperechoic cortical surface of the scapular spine. The presence or absence of a hypoechoic rim within the soft tissues overlying the scapular spine was also recorded. When needed, horses were sedated with detomidine hydrochloride (0.01 to 0.03 mg/kg, IV) and butorphanol tartrate (0.01 to 0.02 mg/kg, IV).
Depiction of the 3 locations (proximal [A], middle [B], and distal [C]) on the equine scapula used to obtain ultrasonographic images of the scapular spine. Images A, B, and C are the corresponding transverse ultrasonographic images of the scapular spine. Notice the spinal widening in a horse with BFS (A and C) that is readily detectable, compared with the normal thin appearance (B). Images A and B were obtained from the same scapula of the same horse, demonstrating regional scapular spinal differences within a horse. Double arrows illustrate the position of calipers to obtain spinal width measurements (A = 17 mm, B = 9 mm, and C = 31 mm).
Citation: Journal of the American Veterinary Medical Association 242, 1; 10.2460/javma.242.1.76
Transverse ultrasonographic images of equine scapular spines showing typical findings in horses without (A) and with BFS (B, C, and D). A—Scapular spine has a normal width with a normal contour. B—Scapular spine is wide with a tufted appearance. Notice the hypoechoic rim between the scapular spine and overlying soft tissues. C—Scapular spine is wide and smooth. D—Scapular spine is wide and spiculated.
Citation: Journal of the American Veterinary Medical Association 242, 1; 10.2460/javma.242.1.76
The standardized ultrasonographic examination was performed on 31 horses (18 horses with BFS and 13 control horses). Ultrasonographic examinations were not performed on all horses because of lack of owner consent or scheduling conflicts. One ultrasonographic examination was performed 1 day prior to the scintigraphic examination. This horse had been referred to the veterinary medical teaching hospital for diagnostic testing of forelimb lameness and not specifically for scintigraphy.
Serum collection and processing—Blood (15 to 20 mL) was collected from all horses between 8:30 am and 10:30 am via either a jugular vein indwelling catheter or venipuncture via an aseptic technique immediately prior to 99mTc-MDP injection. If collected via catheter, blood (3 mL) was withdrawn and discarded to remove heparin before the blood sample was withdrawn into a separate syringe. Two milliliters of blood was placed in 2 tubes containing EDTA to obtain plasma. The remainder of the blood was evenly divided and placed in 2 serum collection tubes and allowed to clot at room temperature (approx 22°C) for 40 minutes. Serum and tubes that contained EDTA were spun in a centrifuge at 25°C (1,200 × g) for 40 and 3 minutes, respectively. Serum and plasma were removed. One milliliter of serum and 0.5 mL of plasma were used immediately; the remainder was placed into cryogenic storage tubes and then frozen at −80°C until analysis (1 to 21 months). All processing of blood samples was completed within 2 hours after collection.
Complete blood counts and serum biochemical analysis (determination of calcium, phosphorus, total protein, glucose, sodium, potassium, chloride, creatinine, urea nitrogen, albumin, globulin, and bilirubin concentrations and aspartate transaminase, creatinine kinase, γ-glutamyl transpeptidase, sorbitol dehydrogenase-37, and alkaline phosphatase activities) were performed at the veterinary medical teaching hospital hematologyf and chemistryg laboratories on the day of blood collection. Intact parathyroid hormone,10 parathyroid hormone–related protein,8 ionized calcium,h and 25-hydroxy vitamin Di analyses were performed at a commercial laboratoryj within 1 to 21 months.
ELISA—An ELISA was performed with kits (bALPk and CTX-1l) that used murine monoclonal antibodies that cross-react with equine bone formation and resorption by-products and that have been validated for healthy horses.8 Samples were run in triplicate with the mean value for each horse used in data analysis.
Statistical analysis—All data were determined to be nonnormally distributed via univariable analysis and the Shapiro-Wilk test.m Two-way ANOVAl and the Wilcoxon rank sum testm were used to calculate and compare means and medians, respectively, among continuous variables between case and control horses. Spearman correlation coefficientsl were calculated to assess correlations among continuous and ordinal variables, including scintigraphic intensity of all lesions measured, number of scintigraphic areas of IRU, serum biomarkers of bone turnover, serum markers of calcium-phosphorus balance, and severity indices of physical and scintigraphic findings, with an r > 0.6 considered clinically relevant and values of P ≤ 0.05 considered significant. The accuracy of physical examination and scapular spinal ultrasonographic findings for the diagnosis of BFS was assessed by calculation of sensitivity, specificity, and 95% exact binomial CI relative to scintigraphic diagnosis. Accuracy of bALP and CTX-1 to differentiate horses with BFS from control horses was assessed via the ROC curve analysis.n The AUC was used as a global summary statistic of test accuracy that was cutoff independent.
Results
Study sample—On initial scintigraphic evaluation, 20 horses had BFS and 40 horses had non-BFS diseases or conditions (control horses). Diagnoses other than BFS included osteoarthritis (n = 19), stress remodeling of bone (8), no skeletal abnormalities noted (6), suspensory ligament desmitis or superficial digital flexor tendonitis (3), cervical facet disease (2), a dental lesion (1), and an abscess caused by Corynebacterium pseudotuberculosis within the triceps musculature (1). Of the 39 horses that underwent scintigraphy as part of diagnostic evaluation of forelimb lameness, 4 were found to have BFS and 35 had other conditions. Of the 14 that underwent scintigraphy because of clinical signs suggestive of BFS, 10 had BFS and 4 had other conditions. Of the 7 horses that underwent scintigraphy because of a history of living on an endemic farm, 6 had BFS and 1 was clinically normal.
Age (< 5 years) was used as an exclusion criterion for further evaluation. The median age of horses with BFS was 14 years (range, 5 to 25 years), and the median age of control horses was 7 years (range, 2 to 16 years). Because the youngest horses with BFS were 5 years old (n = 2) and normal growth can affect scintigraphic findings11 as well as serum biomarkers of bone turnover,8 the control group was limited to only those horses that were ≥ 5 years old. Hence, the remainder of the present study included 20 case horses and 28 control horses, of which 29 were geldings and 19 were mares of various breeds including Appaloosa, Arabian, Friesian, Lipizzaner, Morgan cross, American Paint Horse, pony (Connemara, Icelandic, or Pony), Quarter Horse, Rocky Mountain Horse, Thoroughbred (American or Irish), and warmblood (Dutch, Hanoverian, Holsteiner, Oldenburg, or Trakehner).
Scintigraphic findings—Horses with BFS had IRU in the scapula in consistent locations, and had more lesions with IRU in the scapulae and ribs and typically had greater intensity of lesions than did control horses. Lesions identified by IRU in 40 scapulae of 20 horses with BFS were concentrated in 5 locations: 40 of 40 (100%) scapulae had lesions identified by IRU in scapular spines, 18 (45%) in the caudodorsal scapular angle, 12 (30%) in the caudal border of the scapula, 7 (18%) in the dorsal aspect of the scapula, and 3 (8%) in the supraglenoid tubercle. The number of IRU lesions in scapulae and ribs was significantly higher in horses with BFS (median, 19.5; range, 2 to 47) than in control horses (median, 0; range, 0 to 1). Median intensity of the brightest visible lesion on forelimb scintigraphy was 2-fold and significantly higher in horses with BFS (median, 16.2 CPP; range, 6.8 to 39.6 CPP) than in control horses (median, 7.4 CPP; range, 4.2 to 51.2 CPP). The median intensity of the brightest visible lesion adjusted to the radius was significantly higher in horses with BFS (median, 2.8 adjusted CPP; range, 1.1 to 6.6 CPP) than in control horses (median, 1.5 adjusted CPP; range, 1.0 to 8.2 CPP). Intensity of the standardized region of interest in the right scapular spine was significantly higher in horses with BFS (median, 7.2 CPP; range, 2.8 to 21.9 CPP) than in control horses (median, 6.5 CPP; range, 3.8 to 10.8 CPP). The intensity of the right scapular spine adjusted to the radius was significantly higher in horses with BFS (median, 1.5 adjusted CPP; range, 0.7 to 4.1 CPP) than in control horses (median, 1.2 adjusted CPP; range, 0.4 to 2.1 CPP).
Physical examination findings—The PSI ranged from 0 to 4 in horses with BFS and 0 to 2 in control horses. Palpation of the scapular spine elicited signs of pain in 14 of 20 (70%) horses with BFS and 2 of 28 (7%) control horses. Visible lateral bowing of the scapula was present in 9 of 20 (45%) horses with BFS and 0 of 28 (0%) control horses. Decreased range of motion in the cervical portion of the vertebral column (neck) was present in 8 of 20 (40%) horses with BFS and 0 of 28 (0%) control horses. Severe lordosis was observed in 4 of 20 (20%) horses with BFS and 0 of 28 (0%) control horses. Lameness was not observed in 3 of 20 (15%) horses with BFS and 3 of 28 (11%) control horses. Lameness was mild (grade 1 to 3/5) in 5 of 20 (25%) horses with BFS and 23 of 28 (82%) control horses and was moderate (grade 3 to 4) in 9 of 20 (45%) horses with BFS and 2 of 28 (7%) control horses. A lameness evaluation was not performed on 3 of 20 (15%) horses with BFS because of contraindications (risk of fracture in 2 horses with advanced BFS and exacerbation of ocular disease in 1 mildly affected horse with BFS).
Compared with actual BFS status (present or absent on scintigraphic examination), physical examination findings with the greatest ability to differentiate BFS from other conditions were visible lateral bowing of the scapula (sensitivity, 53%; specificity, 100%) and signs of pain on palpation of the scapular spines (sensitivity, 70%; specificity, 83%). Sensitivity and specificity of all physical examination findings to differentiate horses with BFS from other horses ranged from 53% to 80% and 4% to 100%, respectively (Table 1).
Ultrasonography—Qualitative widening of the scapular spine was observed in 12 of 18 horses with BFS on ultrasonographic evaluation as follows: the left scapula of 4 horses, the right scapula of 1 horse, and both scapulae of 7 horses; none of the control horses had this lesion found on ultrasonographic examination. In horses with BFS, the lesion was specifically distributed to the proximal third (12 scapulae), middle third (12), and distal third (7) regions of the scapular spine. Only 3 of 12 affected horses had scapulae with widening of the scapular spine detected in all 3 regions (proximal, middle, and distal). Measurements of widened regions of the scapular spine in horses with BFS (median, 24 mm; range, 15.6 to 40.0 mm) were greater than regions deemed qualitatively normal in both horses with BFS (median, 12.1 mm; range, 7 to 21.5 mm) and control horses (median, 12.1 mm; range, 6.9 to 22.1 mm).
Variations in surface contour by region of affected scapulae were as follows: proximal third (2 had a tufted appearance, 3 spiculated, and 7 smooth), middle third (4 tufted, 3 spiculated, and 6 smooth), and distal third (3 tufted, 4 spiculated, and 1 smooth). Surfaces that appeared tufted or spiculated were only found on wide scapular spine regions, whereas a smooth surface was found on both wide and normal scapular spine regions. Focal irregularities were found on 3 scapular spines of normal thickness in 2 horses with BFS and 1 control horse. A thin hypoechoic rim overlying the scapular spine was observed in 9 regions of 6 of 18 horses with BFS and 0 of 28 (0%) control horses and was only found in association with the tufted lesion pattern. Abnormal surface features in 4 affected scapulae spanned more than a single region.
Compared with BFS status (present or absent on scintigraphic examination), the ultrasonographic finding with the greatest ability to detect BFS was the presence of a widened scapular spine with a sensitivity of 67% and a specificity of 100% (Table 1). Specific ultrasonographic scapular spinal surface contours (tufted appearance, smooth, and spiculated) had low sensitivity (5% to 15%) for BFS detection. Increased scapular spinal width, with or without surface contour abnormalities, was evident in moderately (SSI, 2 to 3) to severely (SSI, 4 to 5) affected horses with BFS.
Sensitivity and specificity of findings on physical examination (20 horses with BFS and 28 control horses) and ultrasonographic examination of the scapular spine (18 horses with BFS and 28 control horses) to differentiate horses with BFS from horses with other conditions.
| Index test | Sensitivity | Specificity |
|---|---|---|
| Physical examination | ||
| Signs of pain on palpation of scapula | 70 (46–88) | 83 (52–98) |
| Lateral (outward) bowing of the scapula | 53 (28–77) | 100 (69–100) |
| Lameness | 80 (56–94) | 4 (0–19) |
| Ultrasonography | ||
| Wide scapular spine | 67 (41–87) | 100 (88–100) |
| Wide, tufted appearance | 15 (3–38) | 100 (88–100) |
| Wide, smooth appearance | 5 (0–25) | 100 (88–100) |
| Wide, spiculated appearance | 13 (2–38) | 100 (88–100) |
| Hypoechoic rim | 18 (4–43) | 100 (88–100) |
Data are reported as percentages (95% CI).
Scintigraphy was the reference diagnostic method.
Scintigraphy and ultrasonography of the scapular spine—Fifty-nine scapular spines had both scintigraphic and ultrasonographic findings for 3 regions (proximal, middle, and distal); therefore, 177 regions were used in the assessment of colocalization of findings between scintigraphy and ultrasonography. Overall, the ultrasonographic finding of a wide scapular spine had a sensitivity of 43% (95% CI, 31% to 55%) and specificity of 95% (95% CI, 90% to 99%) to detect IRU of any intensity. The specific shape (normal or wide), extent (focal irregularity or regional), and surface contours (tufted appearance, spiculated, or smooth) varied with the associated IRU (no IRU, mild IRU, moderate IRU, and marked IRU).
Scintigraphy and physical examination—The presence of detectable lateral bowing of the scapula on physical examination was strongly correlated with the presence of visible lateral bowing of the scapula on scintigraphy (r = 0.76) and had a sensitivity and specificity of 79% (95% CI, 49% to 95%) and 94% (95% CI, 70% to 100%), respectively. However, lateral bowing of the scapula was detectable on scintigraphy at an earlier stage than on physical examination.
Serum biomarkers—Determination of serum bALP concentration was a more accurate test than determination of serum CTX-1 concentration in differentiating horses with BFS from control horses. The median serum bALP concentration was significantly higher for horses with BFS (median, 38.3 ng/mL; range, 17.9 to 106.0 ng/mL) than for control horses (median, 22.7 ng/mL; range, 14.9 to 51.7 ng/mL), whereas the serum CTX-1 concentration was not significantly (P = 0.83) different between horses with BFS (median, 0.31 ng/mL; range, 0.06 to 0.78 ng/mL) and control horses (median, 0.32 ng/mL; range, 0.06 to 0.92 ng/mL). On ROC curve analysis, serum bALP concentration was moderately accurate for BFS diagnosis and had an AUC of 0.81 (95% CI, 0.69 to 0.94). Serum CTX-1 concentration was not accurate for BFS diagnosis (AUC, 0.48; 95% CI, 0.31 to 0.65). Serum biochemical profile and CBC values and intact parathyroid hormone, parathyroid hormone–related protein, 25-hydroxy vitamin D, and ionized calcium concentrations were not significantly different between horses with BFS and control horses.
Correlation of severity indices—The PSI was moderately correlated (r = 0.67) with the SSI (Figure 4). The presence of wide scapular spines on ultrasonography was strongly correlated with SSI (r = 0.81) and PSI (r = 0.82).
Scatterplot of PSI versus SSI for the detection of disease severity in 20 horses with BFS and 28 control horses. Notice the inability of a physical examination to accurately reflect true, overall disease severity at lower SSI (0 to 2) values.
Citation: Journal of the American Veterinary Medical Association 242, 1; 10.2460/javma.242.1.76
Serum bALP concentration was moderately correlated with SSI (r = 0.59; P < 0.01) and weakly correlated with PSI (r = 0.42; P < 0.01). Serum CTX-1 concentration was not correlated with either SSI (r = 0.10; P = 0.51) or PSI (r = 0.24; P = 0.10; Figure 5).
Scatterplots of serum bALP concentration versus SSI and PSI for the detection of disease severity in 20 horses with BFS and 28 control horses (upper panels); scatterplots of serum CTX-1 concentrations versus SSI and PSI for the detection of disease severity in 20 horses with BFS and 28 control horses (lower panels). Notice the weak to moderately positive relationships between serum bALP concentration and severity indices. Notice the lack of correlation between serum CTX-1 concentration and severity indices.
Citation: Journal of the American Veterinary Medical Association 242, 1; 10.2460/javma.242.1.76
Discussion
This prospective case-control study used forelimb scintigraphic findings as the reference standard to assess the accuracy of physical examination, ultrasonographic examination of the scapular spine, and concentrations of serum biomarkers of bone turnover for the detection of horses with BFS. In addition, indices of disease severity were developed for scintigraphic and physical examination findings. There were no adverse events associated with the reference or index tests used in this study.
Forelimb scintigraphy was the most informative diagnostic modality allowing for assessment of skeletal distribution of lesions, detection of early stages of BFS, and assessment of disease severity. Forelimb scintigraphy was used because it is more economical than whole-body scintigraphy, and scintigraphic findings of the forelimb, neck, and rib cage reflect whole-body findings.2,a Scintigraphy of the whole body, however, would be necessary to assess the full distribution of disease. Furthermore, full-body scintigraphy is medically warranted in some horses because hind limb lameness can be a main clinical sign and catastrophic pelvic fracture has been reported in horses with BFS.3,a
Development of a measure of IRU intensity (CCP and adjusted CPP) provided a potentially useful quantitative technique for monitoring disease activity over time. These measurements need to be validated and used prospectively to assess their repeatability and usefulness. However, future use of CCP or adjusted CPP may provide insight into disease progression or response to treatment.
A diagnosis of BFS could be made by the presence of lateral (outward) bowing of the scapula (sensitivity, 53%; specificity, 100%) with a PSI ≥ 3, such that physical examination was useful for the detection of 7 of 11 moderately (SSI, 3) and all 7 of 7 severely (SSI, 4 to 5) affected horses with BFS. Although not routinely recorded as a clinical sign in the present study, widening of the scapular spine could often be detected by palpation.
The presence of a widened scapular spine on ultrasonography was determined to be diagnostic (sensitivity, 67%; specificity, 100%) for BFS and was useful for the detection of 5 of 10 moderately (SSI, 2 to 3) and 7 of 7 severely (SSI, 4 to 5) affected horses with BFS. Given that this examination can be readily performed and interpreted in the field with standard ultrasound equipment, it was considered a good potential screening test.
Ultrasonographic examination focused on the scapular spine because abnormal findings in affected horses have been localized to this region.1 The entire scapula was examined in most horses of the study reported here but revealed abnormal contours of the supraspinous and infraspinous fossae in only a few severely affected horses. Such findings were more subtle and considered less detectable, compared with reported scapular spine abnormalities. Although scapular spinal width measurements were obtained, ultrasonographically abnormal scapular spines were readily apparent on survey scans prior to measurement acquisition. Examiners should be cautioned against the use of a numeric cutoff point for scapular widening because variations in spinal thickness are present among horse and pony breeds of various sizes. Abnormal scapular spines were similar in appearance and prevalence to results of another study,1 in which 7 of 11 horses had widened scapular spines, some with bony surface changes. In the study reported here, 11 scapulae in 8 horses had abnormal contours (tufted or spiculated appearance). The overlying hypoechoic rim observed with the tufted appearance was thought to reflect soft tissue inflammation and active bone modeling. Differential diagnoses for these bony surface changes include scapular trauma with periostitis, fracture with callus formation, osteomyelitis, and, less likely, neoplasia. Thus, the finding of increased scapular width and abnormal surface contours on ultrasonography should be considered in conjunction with medical history or other findings prior to making a diagnosis of BFS.
Determination of serum bALP concentration was overall a moderately accurate test but not clinically useful in this study population. Serum bALP concentrations of mildly and moderately affected horses with BFS were within the same range as for control horses, and bALP concentrations in severely affected horses were highly variable. When bALP concentrations in horses with BFS exceeded the range of that evident in control horses, BFS most often could be diagnosed by a thorough physical examination with a PSI ≥ 3. Neither CTX-1 nor any of the markers of calcium-phosphorus balance were accurate for detecting BFS in horses.
The SSI and PSI provided a measurable estimate of skeletal compromise and broadly captured the skeletal features associated with BFS. Most horses can be scored by the indices with discrete index categories. However, for those horses that did not fit the indices, the most severe finding (from scintigraphic or physical examination) was used to determine the SSI or PSI, respectively. Continued use of the indices will provide scales with which disease state can be categorized and disease progression can be monitored.
The detection of disease in mild to moderately affected horses remains a clinical challenge. These horses often have lameness without abnormal skeletal conformation. The present study found that scintigraphic lesions preceded both physical and ultrasonographic abnormalities. In contrast to another study,1 the present study demonstrated a moderate correlation between physical examination and scintigraphic findings. This correlation was likely influenced by the concordance of indices at the extreme (ie, the mild and severe) ends of the spectrum. Despite the apparent discordance of the indices in the middle of the spectrums, both indices can be used to follow disease progression. In particular, physical examination and the PSI are particularly useful for those without access to scintigraphy.
Significance, particularly with regard to serum biomarkers, may have been affected by the use of horses with lameness attributable to a non-BFS etiology in the control population. This population was determined to be the most clinically relevant control population because they are most likely to develop BFS. Both bALP and CTX-1 are nonspecific biomarkers of bone turnover and likely become elevated with many orthopedic problems. As such, if nonlame (sound) horses had been used as the control population, we may have found other statistical relationships. Thus, serum biomarkers of bone turnover may be useful in other situations such as screening for BFS and monitoring disease progression or responding to treatment.
Although this is the largest sample of horses with BFS studied, the sample size was small, which limited the ability to detect statistical differences among disease severity categories as well as to detect associations between ultrasonographic bone contour changes of the scapular spine and disease severity. However, the proposed disease severity indices (PSI and SSI) established discrete categories with which horses can be assessed and monitored. The continued prospective use of these indices in conjunction with early diagnosis may provide insight into the time needed for disease progression.
Because of the strong association between BFS and pulmonary silicosis,2,a a history of living on a BFS-endemic farm or in an endemic region should increase the index of suspicion of BFS. Six of 7 horses in the present study that had a history of living on a BFS-endemic farm were positive for BFS. The horse that was negative was 2 years old. The youngest horse to date with BFS was 4 years old.2 Thus, it is unknown whether the 2-year-old horse in the study reported here did not have BFS or whether the disease had not manifested to the degree necessary for detection by scintigraphy.
Cases of BFS in horses have been documented in Carmel, Monterey, Sonoma, and Napa counties in central and northern California,1,3 but as a result of increasing awareness of this disease, horses with BFS in the study reported here came from 2 additional California counties (Shasta and Ventura) that were not previously associated with either BFS or pulmonary silicosis. Considering that the cytotoxic silica dioxide crystals that can cause pulmonary silicosis are found worldwide, BFS may be more prevalent within the horse population than currently appreciated. Additionally, exposure of cytotoxic silica dioxide crystals to air, such as that which occurs during excavation, increases the cytotoxicity of the crystals.12 Thus, BFS may emerge as a new disease in areas where the disease was previously not seen.
Reported sensitivity and specificity of index tests likely differ among attending veterinarians. Detection of BFS, particularly in the early stage of disease, will depend on the experience of the individual performing the diagnostic assessment. Thus, the sensitivity to detecting changes associated with BFS, especially palpation of scapular spinal width or signs of pain, is likely higher for individuals with more experience than for those who have less experience detecting these changes.
In the present study, scinitigraphy was performed 4 to 5 hours after the administration of 99mTc-MDP to allow for maximum skeletal uptake in older horses. The consequence of this time offset was that the scan time approached the half-life of technetium and thus some technetium decay would have occurred at the time of the imaging. However, considering this was the protocol used in all horses of the study, any effect of the decay would have affected all horses equally. The main effect of decay would have been in the calculation of scintigraphic intensity. Thus, the calculated intensities should be interpreted as relative, and not absolute, differences between horses with BFS and control horses.
None of the diagnostic findings assessed in this study were pathognomonic for BFS. Horses with advanced nutritional secondary hyperparathyroidism develop similar conformational changes.13 The clinical signs could be attributed to any disorder that induces systemic osteoporosis. Additionally, the scintigraphic findings could also occur with disseminated neoplasia such as lymphoma or multiple myeloma.6 However, when taken together or in combination with the animal's medical history, these diagnostic findings are highly supportive of BFS.
In summary, scintigraphy remained the most informative diagnostic modality for BFS, providing insight into disease severity and distribution; however, physical examination and scapular spinal ultrasonographic abnormalities were diagnostic in horses with moderate to severe disease and could be used as a less expensive diagnostic tool. Negative findings on physical examination or scapular ultrasonography did not rule out a BFS diagnosis. Disseminated bone disease preceded skeletal deformations. Thus, horses with BFS, but without overt skeletal deformations, may also be at risk for pathological fractures. The proposed severity indices classified the spectrum of disease manifestations and may be useful for prospectively assessing disease progression and response to treatment.
ABBREVIATIONS
| 99mTc-MDP | Technetium Tc 99m medronate |
| AUC | Area under the curve |
| bALP | Bone-specific alkaline phosphatase |
| BFS | Bone fragility syndrome |
| CPP | Counts per pixel |
| CTX-1 | Carboxy-terminal telopeptide of collagen crosslinks |
| IRU | Increased radiopharmaceutical uptake |
| PSI | Physical examination severity index |
| ROC | Receiver operating characteristic |
| SSI | Scintigraphy severity index |
Arens MA. Bone fragility syndrome: disease characterization, assessment of cristobalite as an etiologic agent, and evaluation of diagnostic tests. PhD dissertation, Department of Comparative Pathology, University of California-Davis, Davis, Calif, 2012.
IS2 Gamma Camera, IS2 Medical Systems Inc, Ottawa, ON, Canada.
Ultra-Scan Lift System, Enhanced Technologies Inc, Bedford, Tex.
Mirage software, Digirad Corp, Paway, Calif.
Biosound Technos, Esaote North America, Indianapolis, Ind.
ADVIA 120, Siemens Healthcare Diagnostics, Tarrytown, NY.
Roche Hitachi 917, Roche-Diagnostics, Indianapolis, Ind.
DiaSorin, Stillwater, Minn.
Nova 8+ Electrolyte Analyzer, Nova Biomedical, Waltham, Mass.
Diagnostic Center for Population and Animal Health, Michigan State University, Lansing, Mich.
Metra BAP, Quidel Corp, San Diego, Calif.
Serum crosslaps, Immunodiagnostic Systems Inc, Fountain Hills, Az.
SAS, version 9.2, SAS Institute Inc, Cary, NC.
MedCalc, version 11.3.5, MedCalc Software BVBA, Mariakerke, Belgium.
References
1. Anderson J, Galuppo L, Barr B et al. Clinical and scintigraphic findings in horses with a bone fragility syndrome: 16 cases (1980–2006). J Am Vet Med Assoc 2008; 232:1694–1699.
2. Arens AM, Barr B, Puchalski SM et al. Osteoporosis associated with pulmonary silicosis in an equine bone fragility syndrome. Vet Pathol 2011; 48:593–615.
3. Durham M, Armstrong C. Fractures and bone deformities in 18 horses with silicosis, in Proceedings. 52nd Annu Conv Am Assoc Equine Pract 2006;311–317.
4. Durham M. Silicosis and osteoporosis syndrome. In: Current therapy in equine medicine. 6th ed. St Louis: Elsevier, 2007;303–306.
5. Katzman SA, Nieto JE, Arens AM et al. Use of zoledronate for treatment of a bone fragility syndrome in horses. J Am Vet Med Assoc 2012; 240:1323–1328.
6. Love C, Din AS, Tomas MB et al. Radionuclide bone imaging: an illustrative review. RadioGraphics 2003; 341–358.
7. Lepage OM, Carstanjen B, Uebelhart D. Non-invasive assessment of equine bone: an update. Vet J 2001; 161:10–22.
8. Allen MJ. Biochemical markers of bone metabolism in animals: uses and limitations. Vet Clin Pathol 2003; 32:101–113.
9. American Association of Equine Practitioners. Definition and classification of lameness. In: Guide for veterinary service and judging of equestrian events. Lexington, Ky: American Association of Equine Practitioners, 1991.
10. Miles L, Kipschitz D, Bieber C et al. Measurement of serum ferritin by a 2-site immunoradiometric assay. Anal Biochem 1974; 61:209–224.
11. Celen Z, Zincirkeser S, Ozkiliç S et al. Evaluation of growth in children using quantitative bone scintigraphy. J Int Med Res 1999; 27:286–291.
12. Rimal B, Greenberg AK, Rom WN. Basic pathogenetic mechanisms in silicosis: current understanding. Curr Opin Pulm Med 2005; 11:169–173.
13. Krook L, Lowe J. Nutritional secondary hyperparathyroidism in the horse with a description of the normal equine parathyroid gland. Pathol Vet 1964; 1:1–98.
