Materials and Methods

Description of the data—As part of a mandatory death registry program in Ontario, Canada, all Thoroughbred racehorses that died or were euthanized within 60 days of racing were submitted to the Animal Health Laboratory at the Ontario Veterinary College, University of Guelph, for a postmortem examination. Horses included in this study were necropsied during the 2007 racing season from April 22 to November 25. The following admission data for historical purposes only were collected: age calculated from the upper lip tattoo number, sex, weight, nature of the injury, and cause, time, and location of death.

Collection of samples and analysis of surface features—Forelimbs were removed from the horse at the radius approximately 10 cm proximal to the radiocarpal joint. Soft tissue was removed, and the metacarpophalangeal joint was carefully disarticulated to ensure that cartilage was not damaged during the process. Once disarticulated, the cartilage was kept moist by covering the joint surface with gauze sponges soaked in lactated Ringer's solution.

Macroscopic scoring—The metacarpophalangeal joint surface was divided into 15 subregions (Figure 1) on the basis of regional anatomy and weight bearing. The joint surface was placed over a nonreflective surface and digitally photographed with standardized lighting conditions including six 100-W white light bulbs in a 360° arrangement and a 60-cm cameraspecimen distance in a photographic stand.^a Additional examination of the joint surface was performed with low-angle incident light and a stereomicroscope to detect rugosities, cartilage thinning, cartilage thickening, and subchondral bruising. Each subregion was scored by use of reported lesion descriptors^11–17 in a modification of the OARSI¹⁸ scoring method for osteoarthritis. The macroscopic score for each subregion was a product of 3 components: lesion grade (1 to 3), lesion area (1% to 100%), and a correction factor (1 to 3) for weight bearing.

The lesion grade was assigned on the basis of 17 described pathological descriptors that took into account lesion depth and its probability of association with clinical signs. The mildest lesions were assigned a score of 1, and these included loss of reflectance, cartilage hypertrophy at joint margins, partial-thickness wear lines, partial thinning of cartilage, thickening of cartilage, and enlarged synovial fossae. For instance, wear lines were ubiquitous in the joints and because they are incidental findings in horses without clinical signs of joint disease or lameness,¹⁶ this lesion was given a numeric grade of 1. Full-thickness erosions and deeper lesions were assigned a grade of 2, reflecting their ability to contribute to lameness and affect performance.⁵ This included osteophytes, pitting, erosion, and fibrillation of cartilage. Chronic full-thickness lesions, including fissures, necrosis, full-thickness scoring lines, flattening, fibrous tissue replacement, and severe subchondral bruising, were given a grade of 3.

The following criteria were used to identify the lesions. The mildest degenerative changes were observed as loss of surface reflectance (sheen) and focal hypertrophy of the perichondrial margins, most commonly seen in dorsal and palmar aspects of the first phalanx. Early changes in the continuity of the cartilage were partial-thickness linear wear lines and thinning, thickening, and enlargement of synovial fossae. The palmar third metacarpal bone synovial fossae on each side of the sagittal ridge were considered enlarged when these depressed circular or oval areas of thin connective tissue exceeded 3 mm in diameter.¹² Wear lines were fine, parallel partial-thickness erosions in the sagittal plane.^5,15 Cartilage thinning was observed as a dark area caused by the presence of bone perceived through the thin overlying cartilage.¹ Thickened cartilage appeared opaque, white, and soft. Osteophytes were seen as a series of smooth periosteal spurs at the joint margins. Pitting was identified as small, randomly distributed 1- to 2-mm focal partial-thickness depressions in the articular surface.^14,15 Fibrillation appeared as fine, shaggy 1- to 3-mm-long vertical floating fibers attached to the joint surface when the cartilage was immersed in Ringer's solution.⁴ Erosions were identified when there was partial-to full-thickness attrition of cartilage, which exposed the glassy-appearing calcified cartilage.¹⁴ If a full-thickness erosion had a polished, smooth appearance and exposed the porous subchondral bone, this was considered eburnation.^12,13 Subchondral bruising was identified when the surface reflectivity and cartilage thickness appeared normal, but there was a discoloration caused by blood pigments beneath the cartilage. Fibrous connective tissue was noted when white continuous strands of tissue were found at the articular surface.¹⁵ Flattening of cartilage was observed when there was excessive loss of curvature in the palmar aspect of the third metacarpal bone condyle.¹⁵ Cartilage necrosis was identified as a clearly demarcated area of yellow or dark cartilage distinct from the white surrounding cartilage. Necrotic cartilage was dark, opaque, and partially delaminated from the subchondral bone.¹⁶ In contrast to wear lines, scoring lines were defined as full-thickness linear erosions extending into the subchondral bone that could be identified by detection of a gritty surface texture with a blunt probe.

The lesion area was derived from measurement of abnormal cartilage surface by applying 0.5 mL of India ink^b to the joint surface for 1 minute followed by repeated rinsing with lactated Ringer's solution and blotting (but not wiping) with a sponge to remove incidental ink not intercalated with surface features.^19,20 The joint was photographed again, a piece of clear flexible plastic film was placed on the joint surface, and a water-resistant black pen was used to outline the joint perimeter and the area of ink uptake. The resulting plastic film outlines were digitized, and morphometry software^c was used to measure the ink-stained area in each subregion, which was expressed as a percentage of total joint surface area.

A correction factor for weight bearing, the third component of the macroscopic score, was rated at 1, 2, or 3 on the basis of whether the lesion was in a non—weight-bearing zone (score, 1), a partially weightbearing zone (score, 2), or a fully weight-bearing zone (score, 3); zones were determined on the basis of previous studies^6,21 assessing pressure distribution at different gaits in the equine metacarpophalangeal joint (Figure 1). When more than one lesion was present in a subregion, the first 2 components (lesion grade X the correction factor for weight bearing) were used to compute a mean value, and the total area affected was the sum of the lesion areas. The global joint surface score for each metacarpophalangeal joint was the sum of the 15 subregion scores, and these were used to calculate age-related distributions of joint lesions.

View Full Size

Illustration of the articular surfaces of an equine metacarpophalangeal joint divided into 15 subregions of the distal aspect of the third metacarpal bone, proximal aspect of the first phalanx, and sesamoid bones in a study of the prevalence and severity of metacarpophalangeal joint osteoarthritis in Thoroughbred racehorses. In each subregion are listed the correction factor for weight bearing (1, 2, or 3) and the mean area of the subregion expressed as a percentage of the area of the entire metacarpophalangeal joint. LS = Lateral sesamoid. MLD = Metacarpal lateral dorsal. MLM = Metacarpal lateral middle. MLP = Metacarpal lateral palmar. MMD = Metacarpal medial dorsal. MMM = Metacarpal medial middle. MMP = Metacarpal medial palmar. MS = Medial sesamoid. PLD = Phalanx lateral dorsal. PLM = Phalanx lateral middle. PLP = Phalanx lateral palmar. PMD = Phalanx medial dorsal. PMP = Phalanx medial palmar. PMM = Phalanx medial middle. PSG = Phalanx sagittal groove.

Citation: American Journal of Veterinary Research 71, 11; 10.2460/ajvr.71.11.1284

• Download Figure
• Download figure as PowerPoint slide

In joints in which there was no surface disruption revealed by ink uptake or loss of surface sheen but color change was present from subchondral bruising, thinning, or thickening, these were manually drawn on plastic film as color-coded outlines. These outlines were scanned and digitized to create a second set of scaled image files to determine the area occupied by subchondral changes. The area of each was determined by setting a threshold value for each color and using the measurement function in the morphometry software to produce an output in square milliliters. This area in square millimeters was converted to a percentage of total joint surface for each region.

Histologic scoring—Full-thickness, 5-mm-diameter cartilage biopsy specimens were taken from the joint surface for histologic and biochemical evaluation by use of a No. 12 scalpel blade. Samples were taken from 2 subregions representative of weight bearing and non–weight-bearing areas for each joint.²¹ Weight-bearing samples were taken from the medial transverse ridge of the third metacarpal bone, and non-weight-bearing samples were taken from the medial palmar aspect of the first phalanx. One half of each biopsy specimen was used for histologic examination, and the other half was retained for biochemical analysis. All tissue underwent routine processing in paraffin at 61°C under vacuum followed by embedding with the cartilage oriented to create sections in the coronal plane. All sections were cut at 5-μm thickness and alternately stained with H&E,^d safranin-O,^d or picrosirius red.^e

Histologic scores were generated by use of the OARSI osteoarthritis scoring method¹⁵ performed by 2 independent observers. Interobserver agreement was calculated by use of a weighted Cohen κ statistic for nonexpert observers (graduate students) and experts (pathologists and clinician-scientists). This scoring method was simpler than the aforementioned macroscopic grading scheme because the lesion stage (area as a decimal fraction) was determined from ink uptake confirmed by evaluation of serial histologic sections, and the lesion grade was proportional to lesion depth. In this scoring system, lesion grades 1 to 5 can be briefly summarized as superficial zone cell injury or loss (grade 1), superficial zone matrix depletion and discontinuity (grade 2), vertical clefts (grade 3), erosion (grade 4), and denudation (grade 5). One notable departure from the OARSI system was the use of areas as a percentage of the joint area rather than use of discontinuous categories described in the original method.¹⁷

Polarized light intensity in the cartilage superficial zone—Collagen distribution and orientation were determined by use of 2 picrosirius red-stained sections viewed via polarized light microscopy. Images of cartilage sections at 100X magnification were obtained by use of a videography camera mounted to a compound microscope^f equipped with a rotating stage and 2 polarized filters, 1 below and 1 above the specimen. Images were stored in a personal computer and analyzed by use of morphometry software.^c Initially, glass slides were oriented horizontally in the microscopic field, and the 2 polarized filters were placed at 0° and 45°, respectively, to the horizontal plane. The specimen was rotated with the superficial zone of the cartilage parallel to the analyzer, resulting in a bright, multicolored birefringent superficial zone (Figure 2). In normal cartilage, the densely packed collagen fibers in this area converges and becomes parallel at the articular surface, resulting in a refractile, birefringent zone. The resulting image was converted to a 16-bit gray-scale image by use of a preset threshold to eliminate noise. The light intensity of each specimen was analyzed by use of a line scan function measuring from the deep zone to the superficial zone on each cartilage specimen. A background subtraction routine²² was used. Variable thickness of the samples was accounted for by measuring sample thickness and the proportion of superficial zone thickness. A series of 20 normal cartilage samples was used to determine the normal light-intensity profile of the cartilage and the thickness of the birefringent area. Three depth-dependent, light-intensity profiles through the superficial birefringent area were measured in 2 locations for each slide. If cartilage was absent within a portion of the denoted superficial zone, the light-intensity value was set to zero. A single number representing the difference in light intensities between experimental samples and the mean of the 20 normal specimens at equal depths within the superficial zone was calculated by subtracting line scan profiles, and the mean of the differences was determined to provide 1 number for statistical analysis. Larger numbers in this assessment are an indication of more loss of the superficial zone cartilage. Samples containing new loose fibrous connective tissue on top of the cartilage were omitted from polarized light microscopic analysis because this tissue was birefringent and created an abnormally thick zone of high light intensity.

sGAG in cartilage—The remaining half of each sample was used to measure sGAG concentration by use of the dimethylmethylene blue method.²³ Briefly, 5-to 10-mg samples of cartilage were weighed and placed into a 1.5-mL Eppendorf tube with a mixture of papain granules and digest buffer (1:1 ratio), placed in a 60°C water bath for 18 hours, and stored in a –80°C freezer until the assay was performed. The dimethylmethylene blue-dye molecules contain a positive charge, which allows them to bind to the negatively charged sGAGs. The absorbance was read from the resultant dye-sGAG complex and quantified by use of a spectrophotometer plate reader^g at 525 nm. A standard curve was created by use of shark cartilage chondroitin sulfate as a positive control for comparison. The r² value was ≥ 0.95 and had a slope from 0.13 to 0.15.

Statistical analysis—A statistical package was used for all analyses.^h All data were tested for normality by use of a Shapiro-Wilk test. Spearman correlations were calculated, and 95% confidence intervals were used to determine whether any correlation was significantly different from zero. This included possible correlations between macroscopic and histologic (OARSI) grades, results of polarized light microscopy, sGAG content, and ink uptake. A Pearson correlation analysis was performed between total ink-stained areas versus the subchondral lesion areas within subregions of each joint, and a 95% confidence interval was calculated to determine whether any given correlation was significantly (P < 0.05) different from zero. Macroscopic and histologic scores were not compared directly because both used ink staining as a component of the score; therefore, the macroscopic and histologic grades were compared. Spearman correlation analysis (P ≤ 0.05) was performed on the following variables: age, sex, cause of death, cartilage sGAG concentration, macroscopic score, OARSI histologic score, and polarized light intensity in the superficial zone.

Results

Population summary—Fifty horses were included in this study, and 100 metacarpophalangeal joints were evaluated. Twenty-six horses were female, 11 were sexually intact males, and 13 were geldings. Horses were from 2 to 9 years of age with a mean ± SD age of 3.5 ± 1.4 years for mares and 4.3 ± 2.0 years for all males (stallions and geldings). Two horses were of unknown age and were excluded from age-related calculations.

View Full Size

Polarized light microscopy image and bright-field microscopy image (inset) of the superficial cartilage zone of a metacarpophalangeal joint cartilage specimen from a Thoroughbred racehorse. Region of interest is indicated by the rectangle drawn on the articular surface. Line scan brightness values are illustrated in the graph. Polarized light microscopy image with picrosirius red stain; bar = 100 μm.

Citation: American Journal of Veterinary Research 71, 11; 10.2460/ajvr.71.11.1284

• Download Figure
• Download figure as PowerPoint slide

For all horses, weights ranged from 418 to 560 kg, with a mean ± SD of 460.8 ± 21.7 kg for mares and 419.9 ± 161.9 kg for males, respectively. Of the 50 horses, 35 originated from Woodbine Race Track, 13 were from Fort Erie Race Track, and 2 were at the home farm at the time of death.

Macroscopic scoring—Mean ± SD of the metacarpophalangeal joint surface area was 6,306.4 ± 81.2 mm²; mean subregion areas were calculated (Table 1). The spectrum of findings in the metacarpophalangeal joints ranged from normal to end-stage disease. The transverse ridge contained most of the lesions in the third metacarpal bone. For this study, the authors considered normal joints to be those with reflective, intact cartilage, but mild marginal perichondrial hypertrophy, thickening of peripheral cartilage, and subtle wear lines on the joint surface were common and probably normal in the studied population. Severely affected joints contained widespread wear lines, full-thickness scoring lines, pitting, full-thickness eburnation, fibrillation, and subchondral bone bruising. There was a broad spectrum of pathological lesions of the first phalanx, ranging from mild cartilage thickening to full-thickness eburnation at the dorsal articular margin (Table 2). The data were not normally distributed, so subsequent analyses were performed by use of the Spearman correlation. Mean area of ink uptake was 538.3 mm², constituting 8% of the total joint surface area, but some subregions, notably the metacarpal medial middle, phalanx lateral dorsal, metacarpal lateral middle, and phalanx medial dorsal, had values between 12% and 21% (Table 3). Comparison of microphotographs, ink-stained areas, and tracings of subchondral lesions (bone bruising and cartilage thinning without ink uptake) revealed that ink uptake did not colocalize with subchondral bone bruising and thinning, except in 1 subregion at the lateral dorsal articular margin of the first phalanx. Because subchondral lesions made little overall contribution to the affected cartilage area, these abnormalities were not included in calculation of abnormal cartilage areas. Macroscopic and histologic scores were calculated by use of ink-stained cartilage surface area alone.^17,20

View Full Size

Mean macroscopic cartilage lesion scores by subregion of the metacarpophalangeal joints and age in Thoroughbred racehorses. See Figure 1 for key.

Citation: American Journal of Veterinary Research 71, 11; 10.2460/ajvr.71.11.1284

• Download Figure
• Download figure as PowerPoint slide

Global macroscopic joint scores in the third metacarpal bone, first phalanx, and sesamoid bones in horses of various ages were determined (Table 4). No significant difference in macroscopic scores was detected between left and right metacarpophalangeal joints. Six-year-old horses had the highest macroscopic scores in the third metacarpal and sesamoid bones, and 5-year-old horses had the highest scores in the first phalanx (Figure 3). Patterns in development of pathological change were evident when macroscopic scores and age were considered. The dorsal phalangeal subregions, phalanx medial dorsal and phalanx lateral dorsal, had high scores in 2- and 3-year-old horses and reached a peak at 5 years of age, whereas in all other subregions, scores peaked at 6 years of age. Macroscopic scores were stable until year 5, but in 2- and 3-year-old horses, the mean area of abnormal cartilage in the sesamoid bones (8%), metacarpus (3%), and first phalanx (15%) indicated that joint disease was already well established in these young horses.

Histologic grading—Only cartilage from the medial middle transverse ridge of each third metacarpal bone condyle and the medial palmar aspect of the first phalanx was obtained for histologic analysis (Table 5). Normal joints (OARSI grade 1) had a distinctive superficial zone with many flattened chondrocytes and a superficial zone composed of layers of parallel retractile fibers. The middle and deep zones had randomly arranged or vertical columns of cells, respectively, with increasing safranin-O staining as depth increased. Histologic grades of the severely arthritic joints were 3 to 4 and had complete loss of the superficial zone, full-thickness fissures extending to the calcified cartilage, and severe glycosaminoglycan depletion based on faint or no safranin-O staining. Connective tissue replacement (Figure 4) of the superficial zone and tidemark duplication (Figure 5) were frequently associated with cartilage injury. Mean ± SD OARSI histologic scores for the third metacarpal bone condyle and first phalanx were 31.8 ± 64.4 and 23.9 ± 28.9, respectively. Weighted Cohen κ tests for interobserver agreement had values of 0.77 and 0.83 for nonexpert (graduate students) and expert (surgeons and a pathologist) panels, respectively.

Polarized light intensity in the cartilage superficial zone—Normal joints had much higher light intensity in the superficial cartilage zone than did joints with advanced osteoarthritis. Polarized light intensity for the metacarpal medial middle and phalanx medial palmar subregions was (mean ± SD) 7,226.8 ± 110.8 and 9,919.2 ± 125.5, respectively. Six percent of these samples had new connective tissue invasion from perichondrial borders that interfered with analysis and were not included in this analysis.

sGAG concentration in cartilage—The sGAG concentrations were (mean ± SD) 38.7 ± 14.3 μg and 33.2 ± 11.7 μg of CSC/mg of cartilage for the medial transverse ridge of the third metacarpal bone and mediopalmar aspect of the first phalanx, respectively. Glycosaminoglycan concentration in cartilage declined as macroscopic and histologic scores increased. For example, a horse with a global macroscopic joint degeneration score of 2,588, indicating established joint disease, had a corresponding mean ± SD sGAG concentration of 27.71 ± 2.8 μg of CSC/mg of cartilage. In contrast, a joint with a comparatively low macroscopic score of 150 had a corresponding mean ± SD sGAG concentration of 41.1 ± 2.4 μg of CSC/mg of cartilage.

Correlations among measurements—Significant relationships were found between cartilage sGAG concentration and macroscopic scores (Table 6), sGAG and histologic score, sGAG and ink staining, sGAG and age, and polarized light intensity and macroscopic score. Orthopedic injury as a cause of death was associated with macroscopic grade. There were strong positive correlations between macroscopic score and ink uptake, age and histologic grade, and sex (male) and macroscopic score.

View Full Size

Photomicrograph of a section of articular cartilage from the metacarpophalangeal joint of a Thoroughbred racehorse. Notice replacement of the superficial zone of articular cartilage with new fibrous connective tissue emanating from the periochondrial borders. Loss of proteoglycan staining and chondrocyte density is evident under the connective tissue. Safranin-O stain; bar = 250 μm.

Citation: American Journal of Veterinary Research 71, 11; 10.2460/ajvr.71.11.1284

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Photomicrograph of a section of articular cartilage from the metacarpophalangeal joint of a Thoroughbred racehorse. Notice multiple tidemarks (arrows) between the calcified region and the articular cartilage above it. Safranin-O stain; bar = 250 μm.

Citation: American Journal of Veterinary Research 71, 11; 10.2460/ajvr.71.11.1284

• Download Figure
• Download figure as PowerPoint slide

Discussion

This cross-sectional observational study correlated macroscopic, microstructural, and some biochemical data as indicators of the prevalence and severity of posttraumatic osteoarthritis in racing Thoroughbreds. In addition to establishing the prevalence of osteoarthritis, this study developed a macroscopic method of joint assessment that could be quickly used to quantify cartilage degeneration and provide insight into the time course of osteoarthritis development. The study population contained an even distribution of sexes, and age ranged from 2 to 9 years. Most of the sample population (30/50 horses) were 2 or 3 years old, and 19 died as a result of musculoskeletal injuries, confirming the high prevalence of fatal musculoskeletal injuries in young horses reported by Cruz et al²⁴ in the same geographic area. Interestingly, cause of death (musculoskeletal vs other causes) was correlated with severity of macroscopic joint disease (Table 5), confirming the relationship between catastrophic long-bone injury and arthritis.²⁴ Lameness arising from joint injury could cause a change in gait patterns, resulting in additional loading of a contralateral limb.

Thirty-three percent of the 2- and 3-year-old horses had global joint macroscopic scores consistent with at least 1 full-thickness (grade 3) weight-bearing (correction factor, 3) lesion in approximately half of 1 subregion area. Alternatively, larger partial-thickness or partial weight-bearing lesions may have been present. Because mean subregion area was approximately 420 mm², weight-bearing areas with lesions of 200 mm² were common in this subpopulation, which was consistent with previous reports^5,25 that identified high-risk areas such as the dorsal aspect of the phalanges, the transverse ridge, or the middle weight-bearing portion of the third metacarpal bone. The metacarpal bone locations are predisposed to injury because of the transition between the dorsal and palmar curvatures of the third metacarpal bone condyle. The different curvatures of the dorsal and palmar regions are accompanied by different biochemical compositions; therefore, they have inherently different mechanical and structural properties that predispose this area to osteochondral injury in race horses.¹⁵ The high dorsal phalangeal subregion scores were probably caused by hyperextension injury because this area is not weight bearing at other times.⁶

Marginal cartilage hypertrophy, wear lines, enlargement of synovial fossae, cartilage thinning, and fibrillation at the transverse ridge extending to adjacent condylar cartilage were common findings in the young horses in the present study. By contrast, older horses had more erosive lesions and full-thickness osteochondral defects at the dorsal articular margin of the first phalanx and the transverse ridge of the third metacarpal bone condyle. Subchondral pathological change, though not the subject of this study, was most evident in older horses, and this was consistent with previous findings in a similar population of horses.¹

In the population of horses reported here, biological responses to injury were observed, some of which may be species specific. A high percentage of joints contained marginal hypertrophy or cartilage thickening in the first phalanx and proximal sesamoid bones. At the palmar aspect of the first phalanx, there was extensive hypertrophy, which was determined to be perichondrial proliferation via histologic examination, resulting in new fibrous connective tissue overlying the articular cartilage. This proteoglycan-poor tissue was composed of fibroblasts producing loosely arranged, vascular fibrous connective tissue. On the basis of this observation, we concluded that this tissue was unlikely to mature into well-attached fibrous connective tissue or fibrocartilage, so this should be considered a failed repair attempt. None of the typical cartilage healing responses known as intrinsic repair, extrinsic repair, or cartilage flow were present. In prepubertal animals, chondrocytes can proliferate to create a scar-free intrinsic repair.²⁶ In adults, extrinsic repair is the result of new fibrovascular tissue originating from the bone marrow or perichondrial tissues. Where cartilage is thick, plastic deformation of the lesion edge can help fill the defect in a process known as cartilage flow. Because these horses were adults, no intrinsic repair was possible. There was no histologic evidence of repair tissue even when fullthickness lesions had created continuity between the subchondral bone and the synovial fluid. This may be attributable to the dense nature of osteoarthritic bone in the metacarpophalangeal joint. No cartilage flow was seen, possibly because of the thin nature of the articular cartilage in the metacarpophalangeal joint.

Linear grooves or wear lines orientated in the sagittal plane were found in all joints examined, and this was consistent with another study.¹⁸ Wear lines leading to full-thickness scoring lines are common in the metacarpophalangeal joint and may be a feature of a condylar joint with high loads. It has been speculated that wear lines in the metacarpophalangeal joint may be caused by debris in the joint, which, with normal movement, causes the wear lines or scoring lines in the direction of motion.²⁷ The debris likely originates from osteochondral fractures or erosions at the dorsal articular margins and palmar aspect of the third metacarpal bone as well as the first phalanx.¹⁵

Chondrocyte clones, areas of acellular matrix, fissures, and delamination of cartilage were observed frequently, leading to loss of superficial zone chondrocytes. Loss of these flattened superficial cells and the associated birefringent area was widespread in horses with high histologic and macroscopic grades and ubiquitous in 5- to 9-year-old horses. Because the lamina splendens is important for production of superficial zone protein (also called lubricin), loss of this layer implies deterioration in cartilage lubrication and wear properties.²⁸

In the process of macroscopic scoring, it was noted that areas with ink uptake did not always occur in areas of cartilage in which subchondral bone bruising and cartilage thinning were observed. India ink is a carbon colloid and has a high affinity for articular cartilage in which there is proteoglycan depletion and surface disruption; however, it cannot penetrate or bind to a fully intact proteoglycan-rich surface; therefore, ink uptake is a function of cartilage-surface disruption.²⁰ This explains why ink uptake did not correlate with subchondral bone bruising and cartilage thinning.

Data revealed that joint surface erosion and subchondral bone lesions contributed to the progression of osteoarthritis. When cartilage erosion occurred at the surface of the cartilage, this was easily seen as cartilage thinning, but when subchondral bone remodeling led to flattening of the palmar aspect of the third metacarpal bone condyle, collapse of the articular surface,²⁹ and advancement of the tidemark, changing the biomechanical gradient (as proposed by Radin et al³⁰), this also resulted in articular cartilage thinning but was harder to recognize.²⁰

A variety of macroscopic osteoarthritis scoring systems are used for preclinical studies, drug development, and the study of naturally occurring disease. The scoring system used in the present study was designed to provide qualitative as well as quantitative measurements of cartilage changes associated with trauma-induced osteoarthritis. The cartilage degeneration index used by Brommer et al²⁰ was limited to the first phalanx, did not account for area calculations on curved surfaces, and had a 10% rate of false-positive ink uptake. This method did not allow different lesions to be categorized according to their ability to cause clinical signs and relied heavily on ink staining. The present study put more emphasis on lesions in weight-bearing areas and avoided incidental ink uptake in normal cartilage by the use of blotting and rinsing. The scoring system also provided spatial resolution of lesions contributing to joint disease through the introduction of subregions in the third metacarpal bone, first phalanx, and sesamoid bones. We acknowledge that the division of the joint surface into subregions is time-consuming, but in our opinion, this allows greater differentiation of disease processes in a group of animals. For instance, lesions associated with the transverse ridge in the palmar compartment may have long-standing subchondral bone involvement, whereas diffuse lesions or lesions in the dorsal compartment of the metacarpophalangeal joint are usually caused by osteochondral chip fractures. Drum et al² introduced the use of a qualitative macroscopic grading system (grades 1 to 4) based on a general description of increasing cartilage lesion severity and a corresponding histologic grade (grades 1 to 4).² In contrast, our study used 17 categorical lesions related to etiopathogenesis of the disease to macroscopically characterize the disease. The histologic scoring system used by Drum et al² was similar to our modification of the OARSI scoring system, which reflects the volume of abnormal cartilage through the use of an assigned grade (depth) and stage (area). Young et al³¹ incorporated the area of affected cartilage as a categorical measurement, but a more robust statistical analysis is possible by making area measurements of the abnormal cartilage and total surface area in square millimeters, so the output is a percentage representing the total area affected. Measurement of abnormal cartilage area in millimeters avoided categorical steps that can substantially affect the final score. Better correlation between macroscopic and histologic scores would have been possible if more subregions had been examined histologically; nevertheless, there was a significant association between macroscopic and histologic grades, which substantiates the use of macroscopic evaluations.

Biochemical and histologic correlations were done in only 2 of the 15 subregions, so limited correlations among histologic findings, biochemical findings, and other outcome measures became a weakness of the study. A strength of the study was the high level of agreement between expert and nonexpert observers who used our modification of the OARSI histologic scoring sytem. This indicated that these assessments can be learned quickly by veterinarians, graduate students, and technicians. Calculation of abnormal cartilage area requires measurement of ink-stained area by use of morphometry equipment that is widely available. Strong correlations were found between histologic grade and sGAG loss; lack of other correlations may have been caused by limited sampling. Interestingly, macroscopic grading had significant correlations with sGAG depletion, lack of a bifrefringent superficial zone in polarized light microscopy, and ink staining, indicating that a trained observer can identify the many subtleties of lesions of the articular cartilage.

Although macroscopic grading can be done quickly and reliably, an objective measure of collagen orientation in the superficial zone of cartilage by use of polarized light microscopy was used to assess the integrity of the superficial zone. This was a continuous variable, unlike a previous study³² that used polarized light microscopy in which the outcome was limited to a categorical value such as fibers crossing frequently, occasionally, or never. Collagen orientation and distribution in the superficial zone of articular cartilage are critical to this tissue's low coefficient of friction and viscoelastic properties.^33,34 Loss of the superficial cartilage and disruption of the collagen network are a harbinger of osteoarthritis.^33,34 Given the importance of the lamina splendens in joint lubrication and rheological factors, a fully intact superficial cartilage zone is essential for functionality and protection against progression of the disease. The propensity for horses to produce exuberant connective tissue on the articular surface, similar to horses’ ability to produce granulation tissue in wounds, is worrisome because fibrous tissue overlying cartilage could interfere with cartilage nutrition and lubrication.

The factors that were most strongly correlated with progression of OA in this population were age and male sex. Fatal musculoskeletal injuries were strongly associated with advanced joint disease, indicating that joint injury and bone injury occur simultaneously, although the 0.30 correlation coefficient indicates that one cannot be used as a surrogate indicator for the other. Low correlation coefficients found in this study are typical of biological studies in relatively small populations in which many factors contribute to disease progression, so larger studies would probably identify more factors. In the present small study, correlation coefficients ≥ 0.30 were considered to be strong evidence of an association.

Results of the study reported here indicate that 2- and 3-year-old Thoroughbreds can have well-established joint disease and that the most severe lesions occurred in horses 5 years of age and older. This is consistent with a previous report¹⁰ of age-related OA progression, although our findings did not indicate a continuous increase in lesion severity with age. On the basis of guidelines for the human knee in which 1-cm² lesions are frequently symptomatic and 2-cm² lesions often require repair procedures to prevent progression,³⁵ we speculate that many of the 2- and 3-year-old horses had clinical signs. Although the threshold for pain in horses may be quite different than that in humans, it seems likely that intermittent lameness would be present and treatment of some kind would be needed to permit racing, including intra-articular injections. Difficult decisions need to be made by the racing industry regarding the use of intra-articular medications, particularly corticosteroids and other pain-relieving medications. The Thoroughbred industry may also need to consider whether races among 2- and 3-year-old horses should continue to be allocated the largest purses because results of the present study suggest that metacarpophalangeal joint osteoarthritis in young Thoroughbred racehorses may be compromising their longevity in racing. Perhaps breeding for long-term soundness and resistance to injury will become a priority.

The hypothesis regarding age and severity of joint disease was partly supported by the data, although 7-, 8-, and 9-year-old horses had fewer lesions than did the 5-year-old horses in this study. This may be because only those horses with relatively sound metacarpophalangeal joints are able to continue racing year after year. Older horses were a small portion of the sample population, but perhaps more attention should be given to these horses to identify genetic or other factors, such as training methods, that allow longer competitive lifetimes. Future work should focus on factors that might make 2-year-old horses susceptible to early joint disease, including genetic, nutritional, and management factors. High-resolution diagnostic imaging and cartilage biomarker panels that could identify horses at risk for joint disease would be valuable.

Rapid quantitative macroscopic methods of assessing joint disease in the metacarpophalangeal joint correlated well with quantitative measures of cartilage quality and health. Results of this study support the use of assessment systems that create data as a continuous output rather than scored categorical data. Although some additional work is needed beyond routine macroscopic observation, the strength of this system arises from assessment of the volume of tissue lost or damaged, with an emphasis on injury to weight-bearing areas.