Radiographic and Arthroscopic Findings in the Elbow Joints of 263 Dogs with Medial Coronoid Disease



    1. Fitzpatrick Referrals, Eashing, United Kingdom and Veterinary Clinical Sciences, University of Illinois at Urbana-Champaign, Urbana, IL
    Search for more papers by this author

    1. Fitzpatrick Referrals, Eashing, United Kingdom and Veterinary Clinical Sciences, University of Illinois at Urbana-Champaign, Urbana, IL
    Search for more papers by this author

    1. Fitzpatrick Referrals, Eashing, United Kingdom and Veterinary Clinical Sciences, University of Illinois at Urbana-Champaign, Urbana, IL
    Search for more papers by this author

    1. Fitzpatrick Referrals, Eashing, United Kingdom and Veterinary Clinical Sciences, University of Illinois at Urbana-Champaign, Urbana, IL
    Search for more papers by this author

  • Presented in part at American College of Veterinary Surgeons Annual Symposium, Washington DC, USA, 2006.

Address reprint requests to Noel Fitzpatrick, MVB, CertSAO, CertVR, Fitzpatrick Referrals, Halfway Lane, Eashing, Surrey GU7 2QQ, UK. E-mail:


Objective— To report clinical, radiographic, and arthroscopic findings in dogs with thoracic limb lameness attributed solely to disease of the medial aspect of the coronoid process (MCP).

Study Design— Case series.

Animals— Dogs (n=263) with MCP disease (MCD; 437 elbows).

Methods— Clinical records (January 2000–July 2006) and radiographs were reviewed and pertinent data recorded. Radiographic interpretation included measures of periarticular osteophytosis, gross assessment of MCP integrity, and measurement of ulnar subtrochlear sclerosis (STS). Statistical analysis was performed to evaluate associations between data; confidence interval was set at 95%.

Results— Labrador Retrievers were 50.2% of all dogs with MCD. Mean age at diagnosis was 32 months and duration of lameness was 14.5 weeks. Thirteen elbows (3%) were considered radiographically normal. Osteophytosis was identified on the anconeal process (70.2%), radial head (37.3%), and lateral epicondyle (56.5%), and STS was identified in 86.7% of elbows. Median osteophytosis score was 1; mean absolute osteophytosis score was 1.7. Arthroscopic findings included: fissuring (18.3%) and fragmentation (64.1%) of the MCP and kissing lesions (49.0%) of elbows. Median-modified Outerbridge score of the MCP was 2 and the humeral condyle, 0. Weak or moderate correlations were found between osteophytosis and modified Outerbridge scores and weak correlation between modified Outerbridge scores of the MCP and medial humeral condyle.

Conclusions— Wide ranges in clinical, radiographic, and arthroscopic findings are recognized in dogs with MCD but correlations between such factors are generally weak. Radiographic and arthroscopic findings do not correlate with owner-reported duration of lameness.

Clinical Relevance— Radiographic measures of osteophytosis are poor predictors of severity of arthroscopic pathology for MCD.


DISEASE OF the medial aspect of the coronoid process (MCD), a component of elbow dysplasia, includes fissuring and fragmentation of the medial aspect of the coronoid process (MCP), cartilage and/or subchondral bone pathology. MCD is frequently diagnosed as a cause of lameness in medium to large breed dogs and is the most common cause of thoracic limb lameness in juvenile dogs of certain breeds.1–3 The exact cause of MCD is debated but elbow incongruence is reportedly a primary contributor to the pathogenesis of MCD.1,4–6 Concomitance with other manifestations of elbow dysplasia like osteochondritis dissecans (OCD) of the medial humeral condyle and ununited anconeal process (UAP) is well recognized.7–9

Diagnostic imaging techniques commonly used to establish a diagnosis of MCD include radiography,2,10–12 computed tomography,13,14 scintigraphy,15 and magnetic resonance imaging (MRI)16; however, direct observation by arthroscopy or arthrotomy may be required for diagnostic confirmation.17

Radiography is widely available, inexpensive, and can be performed effectively with patient sedation. Use of oblique radiographic projections has been recommended to increase the likelihood of identifying free fragments.18 An inability to define the axial border of the MCP without superimposition of other bony structures on any radiographic projection2,19,20 remains a distinct disadvantage, because free fragments are infrequently detected even in retrospect, after identification by arthroscopic observation. Radiographic diagnosis of MCD is commonly based on recognition of nonspecific characteristics of degenerative joint disease (e.g., periarticular osteophytosis) combined with exclusion of other primary disease processes (e.g., OCD, UAP). Other radiographic changes such as increased peritrochlear ulnar radiopacity, ulna subtrochlear sclerosis (STS)14,21–23 and loss of definition of the MCP outline on mediolateral radiographs of the elbow have been proposed, but rely on subjective interpretation techniques. On the mediolateral projection, the cranial aspect of the MCP and the proximal ulna should have a shallow concave contour.21 Lack of outline definition of the proximal aspect of the MCP, a convex or flattened shape to the MCP, or cortical irregularity are strongly indicative of MCD.21,24 It is recognized that some elbows affected with clinical MCD are radiographically normal,25 and it has been reported that radiographic changes associate poorly with both clinical signs and gross pathologic disease,26 necessitating use of other imaging techniques.

Arthroscopy was reported in dogs in 1976,27,28 a technique for the elbow joint described in 1993,29 and arthroscopic findings in 100 dogs reported in 1997.30 Arthroscopic evaluation is considered safe and reliable, allows more thorough examination of the elbow joint compared with arthrotomy, and facilitates minimally invasive application of intra-articular surgical procedures,31 but requires general anesthesia.

Because radiographic interrogation of the elbow is noninvasive, we were interested in establishing its merits and limitations by comparing findings obtained from techniques commonly used in practice with arthroscopic findings. Specifically we were interested in the value of radiographic assessment of MCD as a means of guiding treatment approaches. We hypothesized that: (1) increasing severity of radiographic changes as assessed by clinician measurements would be associated with increasing severity of arthroscopic findings; and (2) increasing duration of lameness would be associated with increases in the same radiographic measures and/or arthroscopic findings.


Inclusion Criteria

Medical records (January 2000–July 2006) and radiographs of dogs that had thoracic limb lameness attributed solely to MCD were retrieved for review. Dogs with MCD that also had any other disease of the thoracic limb (including but not limited to osteochondrosis of the medial aspect of the humeral condyle or caudal humeral head, UAP) identified by any diagnostic technique modality were excluded.

Clinical Data

Retrieved data included age, limb (left, right), duration of lameness at time of diagnosis, and arthroscopic findings. Arthroscopic findings assessed by a single investigator (N.F.) included modified Outerbridge scores31 of MCP cartilage quality (Table 1), presence of gross fragmentation or fissuring (Table 2), modified Outerbridge scores for the articular cartilage of the medial aspect of the humeral condyle, and the subjective presence or absence of synovitis. Synovitis was defined as any visible morphologic deviation from the normal well-demarcated, pale pink, nonfimbriated, plicated appearance of canine elbow synovium. In our experience, normal elbow synovium has minimal visible vascularity with the exception of a small group of nondistended but slightly tortuous vessels in a synovial frond cluster immediately cranial to the tip of the MCP.

Table 1.  Modified Outerbridge Scoring System31 Used for Arthroscopic Evaluation of Cartilage Pathology
Outerbridge Score
Description of Gross Cartilage Quality
1Chondromalacia (assessed by use of an arthroscopic probe)
2Partial thickness fibrillation
3Deep fibrillation
4Full thickness cartilage loss
5Subchondral bone eburnation
Table 2.  Grading System for Integrity of the Medial Aspect of the Coronoid Process
MCP GradeDefinition/Description
Intact1. No signs of fissuring or fragmentation; no movement of the medial aspect of the coronoid process during probing
2. Cartilage damage, modified Outerbridge scores 1–5 may be present.
Fissure1. Localized full-thickness cartilage fissure but medial aspect of the coronoid process remains in situ during probing
Fragmentation1. Free osteochondral body with corresponding loss of MCP or
2. Freely manipulated upon probe palpation.

Elbows were considered to have MCD based on arthroscopic identification of either gross fissuring or fragmentation of the MCP, or where the MCP was grossly intact, but had a modified Outerbridge score of 1–5 for the cartilaginous surface of the MCP in conjunction with subjective presence of synovitis.

Radiographic Assessment

Radiographs (craniocaudal, flexed, and extended mediolateral projections) were assessed by 1 individual (T.S.). Elbows where radiographic quality was subjectively considered inadequate (including inappropriate exposure, developing faults, and inappropriate positioning [e.g., lack of concentricity of the medial and lateral humeral condylar silhouettes on the mediolateral projection]) were excluded.

Standard radiographic technique included direct contact between the elbow being imaged and radiographic cassette where possible (film-focal distance, 100 cm) to minimize magnification, or where direct contact between the limb and radiographic cassette was not possible, a radiopaque 10 cm marker was positioned immediately adjacent to the elbow to permit magnification correction. The investigator was unaware of patient data including arthroscopic and clinical findings, and availability of radiographs for the contralateral elbow. Radiographic evaluation included measures of periarticular osteophytosis, presence of a detectable osteochondral fragment of MCP on mediolateral or craniocaudal radiographic projections, evaluation of coronoid shape on the mediolateral projection, and quantification of STS as a percentage (%STS).

Presence of Osteophytosis. The caudal edge of the lateral epicondyle, the proximal surface of the anconeal process, and radial head were subjectively assessed for the presence of osteophytosis using the mediolateral projections and graded as: 1=normal or 2=osteophytosis.

Degree of Osteophytosis. Periarticular osteophytosis was quantified using 2 scoring schemes: (1) categorical grading (grades 0, 1, 2, 3) of arthrosis as a consensus across all radiographic views (Table 3); and (2) an absolute measurement of maximal osteophyte depth on the caudoproximal anconeal process, measured from a maximally flexed mediolateral radiographic projection (Fig 1). Measurements were made to the nearest 0.1 mm using a Vernier caliper (Tricle Brand, Shanghai, China). No adjustments or post-hoc calculations were performed to account for variation in dog size.

Table 3.  Grading Scale for Periarticular Osteophytosis
Radiographic Finding
0No osteophytosis, subtrochlear sclerosis may be present
1Osteophytosis in one or more locations <2 mm in height
2Osteophytosis in one or more locations 2–5 mm in height
3Osteophytosis in one or more locations >5 mm in height
Figure 1.

  Measurement of maximal absolute depth of osteophytosis (X) on the caudoproximal anconeal process, measured perpendicular to the anconeal process.

Assessment of MCP Morphology. MCP shape was subjectively assessed and graded as: 1=normal; 2=abnormally shaped (based on the degree of concavity of the cranial aspect of the MCP); and 3=evidence of fragmentation.

Assessment of STS. Each flexed mediolateral projection was assessed for presence of STS. When present, STS was quantified as a percentage of the length between 2 fixed points (Fig 2). To measure %STS a line was created perpendicular to the most caudal margin of the ulnar proximal metaphyseal cortex (point α) and to the most proximocaudal aspect of the radial head (point β). The STS caudal border (point δ) which constituted a subjective radiographic assessment of the junction between sclerotic and normal trabecular bone pattern, was created along line α–β, point δ. The distance α–δ (X) was expressed as a percentage of the total distance α–β (Y). The %STS was calculated as 100(X/Y). Elbows without STS were scored as 0%.

Figure 2.

  Flexed mediolateral radiograph of showing landmarks used to calculate %STS. Point α, caudal margin of the ulnar proximal metaphyseal cortex; point β, proximal caudal aspect of the radial head; point δ, STS caudal border. %STS was calculated as 100 (distanceX/distanceY).

Statistical Analysis

Statistical software (Minitab® 14.20 Statistical Software, Minitab Ltd., Coventry, UK) was used to calculate descriptive statistics and to investigate potential associations between clinical, radiographic, and arthroscopic variables. Association between major clinical, radiographic, and arthroscopic findings was assessed using Spearman's rank correlation coefficient (JMP version 6.0.0, SAS Institute, Cary, NC). Association between presence of medial humeral condyle pathology with gross arthroscopic fragmentation of the coronoid process was investigated by calculation of odds ratio. Arthroscopic assessment of gross MCP fragmentation and kissing lesions extending into the subchondral bone (modified Outerbridge score 5) were compared with radiographic assessments of MCP fragmentation and radiographically evident subchondral bone defects and sensitivities calculated. Confidence interval was set at 95%.



Two hundred and sixty-three dogs (n=437 elbows) diagnosed with MCD met the inclusion criteria. Labrador retrievers (132 dogs, 50.2%) were most commonly represented of 38 breeds (Table 4). Cross breeds (16 dogs, 6.1%) were predominantly “Labrador-crosses”, assessed subjectively from their phenotypic appearance. Dog age ranged from 3 to 135 months (mean ± SD, 32 ± 36.0 months; median 14 months) with 58.4% of dogs being <18 months old. Male:female ratio was 5:3. There were 49 (18.6%) intact and 48 (18.3%) spayed females, 100 (38.0%) intact males, and 66 (25.1%) castrated males.

Table 4.  Breed Distribution of Dogs with Medial Coronoid Disease
BreedNumber of
Dogs (%)
  • *

    One of each: Airedale Terrier, Belgium Shepherd Dog, Bullmastiff, Bouvier des Flanders, Beagle, Briard, Cavalier King Charles Spaniel, Curly Coated Retriever, Dog De Boudeaux, Flatcoated Retriever, German Shepherd Dog Great Dane, German Wire Haired Pointer, Huntaway, Jack Russell, Newfoundland, Pyrenean Mountain Dog, Rough Collie, Long Haired Collie, Rhodesian Ridgeback.

Labrador Retriever132 (50.2)
Rottweiler32 (12.2)
Cross-breed16 (6.1)
Golden Retriever13 (4.9)
English Springer Spaniel8 (3.0)
Boxer7 (2.7)
Border Collie6 (2.3)
Australian Cattle Dog4 (1.5)
Newfoundland4 (1.5)
Weimaraner4 (1.5)
Old English Sheepdog3 (1.1)
Bearded Collie2 (0.8)
Bernese Mountain dog2 (0.8)
Cocker Spaniel2 (0.8)
Italian Spinone2 (0.8)
Hungarian Viszla2 (0.8)
Samoyed2 (0.8)
Staffordshire Bull Terrier2 (0.8)
Single Breeds*20 (7.6)
Total263 (100)

Distribution of MCP disease was left (48.7%; 213 elbows) and right (51.3%; 224) including 174 dogs (66.2%) that were bilaterally affected (Table 5).

Table 5.  Distribution of Dogs by Categorical Age Group and Associated Gender Ratios and Lameness Duration
Age Range
Number of
Dogs (% Total)
Gender Ratio
Mean (± SD)
Duration (Weeks)*
  • *

    Data when available represents duration of lameness before admission/elbow.

0–5.996 (2.3%)5:14 ± 2.6 (6 elbows)
6–11.9992 (35.0%)12:112 ± 8.5 (99 elbows)
12–23.9973 (27.7%)2:118 ± 15.4 (78 elbows)
24–35.9915 (5.7%)1:114 ± 9.1 (14 elbows)
36–47.9919 (7.2%)1:112 ± 8.4 (19 elbows)
48–71.9922 (8.4%)1:112 ± 0.4 (22 elbows)
>7236 (13.7%)2:119 ± 36.1 (34 elbows)

Lameness Duration

Mean (± SD) duration of lameness before admission (272 elbows) was 14.5 ± 14.20 weeks (median, 12 weeks; range, 1–108 weeks; Fig 3, Table 5). Sixty-five percent of dogs were admitted for investigation of unilateral thoracic limb lameness but clinical examination revealed bilateral lameness in 66% of these dogs referred for unilateral lameness.

Figure 3.

  Histogram of lameness duration per elbow before diagnosis of MCD.

Arthroscopic Findings (Table 6)

Table 6.  Incidence of Modified Outerbridge Scores for Pathology of the Humeral Cartilage Surface and Medial Coronoid Process Identified Arthroscopically
Outerbridge Score
Medial Coronoid Process
  1. Data Represents Number of Elbows/Score Combination and (Relative % of Non-missing Values).

 07 (1.7%)84 (20.9%)74 (18.4%)20 (5.0%)13 (3.2%)2 (0.5%)200 (49.8%)
 112 (3.0%)15 (3.7%)6 (1.5%)1 (0.2%)3 (0.7%)1 (0.2%)38 (9.5%)
 21 (0.2%)9 (2.2%)18 (4.5%)7 (1.7%)6 (1.5%)0 (0.0%)41 (10.2%)
 30 (0.0%)16 (4.0%)10 (2.5%)6 (1.5%)14 (3.5%)0 (0.0%)46 (11.4%)
 41 (0.2%)8 (2%)6 (1.5%)1 (0.2%)49 (12.2%)2 (0.5%)67 (16.7%)
 50 (0.0%)2 (0.5%)1 (0.2%)0 (0.0%)5 (1.2%)2 (0.5%)10 (2.5%)
Total21 (5.2%)134 (33.3%)115 (28.6%)35 (8.7%)90 (22.4%)7 (1.7%)402 (100.0%)

Synovitis present in all examined elbows (437) was manifested by subjectively assessed erythematous discoloration of the synovium and increased frequency of visible blood vessels and increased distension or tortuosity of vessels. Other common findings included fimbriation of the synovial surface with large numbers of fine villous protrusions and increased plication of the synovium that did not resolve with capsular distension during arthroscopy. Subjectively, these changes equally affected the craniolateral (as far as could be assessed) and craniomedial joint compartments as evaluated through standard medial arthroscopy portals. Focal increases in severity of synovitis immediately adjacent to the MCP were also identified, but not consistently so.

The MCP was grossly intact in 77 (17.6%), fissured in 80 (18.3%) and fragmented in 280 (64.1%) of 437 elbows. Modified Outerbridge scores for MCP cartilage (excluding the existence of gross fragmentation of fissuring) were available for 405/437 (92.7%) elbows with a median score of 2 (range, 0–5). Modified Outerbridge cartilage scores for the medial aspect of the humeral condyle were available for 418/437 (95.6%) elbows with a median score of 0 (range, 0–5). A kissing lesion (modified Outerbridge >0) was identified in 205/418 (49.0%) of elbows.

Radiographic Findings

Appropriate radiographs for assessment were available for 423/437 (96.8%) elbows.

Presence of Osteophytosis. Osteophytosis was present on the anconeal process in 307 (70.2%), the radial head in 163 (37.3%), and the lateral epicondyle in 247 (56.5%) elbows.

Degree of Osteophytosis. Using categorical grading, the range of osteophytosis grades was 0–3, with a median score of 1. Distribution of scores was 0=28.8% (122 elbows); 1=32.6% (138 elbows); 2=28.6% (121 elbows); and 3=9.9% (42 elbows). Using absolute measurement of anconeal osteophytosis, mean ± SD maximum depth of osteophytosis on the anconeal process for 423/437 (96.8%) elbows was 1.7 ± 1.90 mm (median, 1.2 mm; range, 0–12.3 mm; Fig 4).

Figure 4.

  Histogram of magnitude of absolute osteophyte depth on the caudoproximal aspect of anconeal process on a flexed mediolateral radiographic projection of the elbow.

Assessment of MCP Morphology

Of 437 elbows, MCP was considered normal in 107 (24.5%), fragmented in 55 (12.6%), and abnormally shaped in 275 (62.9%).

Assessment of STS

Ulna STS was present in 379 (86.7%) of elbows. Mean (± SD) %STS value for all elbows was 38 ± 15.7% (median, 39%; range, 0–100%) and 42 ± 11.7% (median, 41%; range, 25–100%) in elbows where STS was considered present. Ulna STS was scored as 0% (absent) in 58 (13.3%) of elbows whereas five elbows (1.1%) had %STS scores of 100%, evidenced by an increase in bone opacity and loss of trabecular pattern extending to the caudal ulnar cortex at the level of the MCP.

Relationship between Clinical, Arthroscopic, and Radiographic Findings

Although statistically significant correlation was found between several clinical, arthroscopic, and radiographic variables, in general the associations were negligible (rs<.2) or weak (rs=0.2–0.4; Table 7).

Table 7.  Associations between Major Clinical, Arthroscopic and Radiographic Variables
Variable ComparisonSpearman
  1. MCP, medial coronoid process.

Duration of lameness
Osteophytosis measurement (anconeal process)
 Osteophytosis measurement (anconeal process)0.18.0003Negligible
Osteophytosis score
 Osteophytosis measurement (anconeal process)0.92<.0001Strong
Modified Outerbridge score: MCP
 Osteophytosis measurement (anconeal process)0.33<.0001Weak
 Osteophytosis score0.34<.0001Weak
Modified Outerbridge score: Humerus
 Osteophytosis measurement (anconeal process)0.41<.0001Moderate
 % STS0.17.0007Negligible
 Osteophytosis score0.45<.0001Moderate
 Modified Outerbridge MCP0.39<.0001Weak

Medial Humeral Condyle Pathology

Kissing lesions (humeral-modified Outerbridge score >0) were identified in 55.5% elbows with fragmentation, 27.5% elbows with fissuring and 34.4% elbows with intact coronoid. Kissing lesions were significantly more likely to occur in association with fragmentation than without free fragmentation with an odds ratio of 2.75 (P=0.000; 95% confidence interval range, 1.84–4.11). There was no significant variation within severity of kissing lesions between coronoid integrity groups. No other variable had any statistically significant association with gross integrity (graded as fragmented, fissured or intact).

Radiographic Sensitivity

Comparing radiographic and arthroscopic findings, radiography had a sensitivity of 16% for diagnosing gross fragmentation of MCP and 30% for diagnosing modified Outerbridge scores of 5 (subchondral bone defect) for the medial aspect of the humeral condyle (where loss of the subchondral bone was observed arthroscopically and would be expected to be detected radiographically).

STS-Free Elbows

In 58 elbows where STS was absent, 24/48 (41.4%) were from Labrador Retrievers and 9/58 (15.5%) from Rottweilers. Mean (± SD) age/elbow was 918 ± 1090 days (57 elbows) and mean duration of lameness 177 ± 17.9 weeks (35 elbows). Of these 58 elbows, radiographic abnormalities included osteophytosis of the radial head in 17 (29.3%), osteophytosis of the anconeal process in 29 (50.0%), osteophytosis of the lateral epicondyle in 26 (44.8%), kissing lesions in 2 (3.4%) and fragmentation of or an abnormally shaped MCP in 34 (58.6%) (Fig. 5).

Figure 5.

  Arthroscopic images (left) with matched extended and flexed mediolateral radiographs and craniocaudal radiograph (left to right) demonstrating the spectrum of joint pathology identified. (A) 12-month-old male neutered German Short Haired pointer with moderate synovitis and modified Outerbridge grade 2 cartilage pathology at the tip of the MCP without overt fissure or fragment formation (left elbow); (B) 11-month-old male entire Rottweiler with fissure formation in the region of the radial incisure of the MCP (right elbow); (C) 18-month-old female neutered Labrador retriever with gross fragment formation at the radial incisure of the MCP (left elbow); (D) 28-month-old male neutered Golden Retriever with modified Outerbridge grade 3 pathology of both MCP and medial humeral condylar surface demonstrating typical linear striations in absence of gross fragmentation of the MCP (left elbow); (E) 15-month-old male entire English Springer Spaniel with modified Outerbridge grade 3–4 pathology of both MCP and medial humeral condylar surface including gross fragmentation of radial incisure of MCP with apparently undiseased cartilage cap (right elbow); (F) 60-month-old female neutered German Shepherd Dog with modified Outerbridge grade 5 pathology of both MCP and medial humeral condylar surface in absence of gross fragmentation of the MCP (right elbow); (G) 20-month-old male entire Golden Retriever with modified Outerbridge grade 5 pathology of both MCP and medial humeral condylar surface including gross fragmentation of radial incisure of MCP with apparently undiseased cartilage cap (left elbow).

Elbows Considered Normal on Radiographic Examination

Of 437 elbows, 13 (3%) were considered free from any radiographic evidence of disease. Arthroscopic findings in these elbows included fragmentation (6 elbows), fissuring (2), and grossly intact MCP (5). Median-modified Outerbridge score for the MCP was 2 (range, 1–4) whereas 11/12 had a modified Outerbridge score for the medial humeral condyle of 0.


MCD occurred in dogs with a spectrum of signalments and breed diversity. Juvenile dogs (<18 months old) were over represented but there were also a substantial number of mature dogs (>18 months old). Likewise there was a diverse spectrum of radiographic and arthroscopic changes. Like others,2,18,20 we found that free fragmentation of the MCP was often not visible on standard radiographic projections, and further, that radiographic signs of disease generally correlated poorly with the severity of disease visible on arthroscopy. Indeed, 13 elbows that were considered radiographically normal had joint pathology (identified as fragmentation, fissuring, or grossly intact MCP with a modified Outerbridge score ≥1) on arthroscopic examination. We also found that radiographic and arthroscopic findings had no correlation with duration of lameness.

Study Limitations

Our inclusion of dogs with visibly intact MCP but mild cartilage disease by our definition of MCD may be considered controversial but has been reported before.17,20 Diagnosis was justified based on clinical signs, including lameness and pain of elbow manipulation, radiographic evidence of arthrosis in the absence of other conditions, and presence of arthroscopically visible synovitis. Whether these dogs should be classified more accurately as medial compartment disease,32 which may include cases where subchondral bone microfissuring17 may precede gross fissuring or fragmentation, is beyond the scope of this report. We believe such cases should be included within the diagnosis of MCD and in fact, resolution of lameness in most of these dogs after surgical intervention supports our contention. Clinically affected elbows had subtotal coronoid ostectomy and excised coronoid specimens were assessed histomorphometrically.17 Every specimen where cartilage was visibly diseased but intact, was permeated by subchondral microcracks and diffuse microdamage. However, we acknowledge that inclusion of these cases may have contributed to an overall reduction in scores for arthrosis and cartilage pathology compared with previous reports.

The lack of a disease-free control group or cohorts of dogs afflicted by other elbow joint pathology limits interpretation of our results, particularly related to our radiographic findings which are likely to be relatively nonspecific for MCD. Lack of a comparison population limits clinical application of our radiographic findings as potential predictors of MCD. However, to our knowledge, this is the first study where the radiographic findings of elbows affected by MCD in isolation (i.e., excluding of all other forms of elbow pathology) have been reported, and the data is therefore of epidemiological importance for this population.

Investigator factors represent a further study limitation. Arthroscopic and radiographic assessments were performed by a single observer. Because several of the assessments were at least partly subjective this could be a potential source of inaccuracy or bias. Furthermore, because the assessments we used have not been validated, inter- and intraobserver variations remain unknown. We attempted to limit the likelihood of errors of interpretation by use of multiple assessments to investigate similar radiographic features (application of an ordinal scoring scheme together with an absolute measurement for assessment of osteophytosis), and by the large number of cases; however, our conclusions and potential clinical applicability should still be interpreted with caution.

Whereas our study encompassed a large number of cases with a wide range of signalments and severities of arthroscopic pathology, it is of significance that each dog was evaluated only once at the time of admission to a referral clinic. This has implications for interpretations about the chronology of disease progression, because none of the cases within this study can be reliably defined as being “early” or “late” in disease progression. Duration of lameness might have provided further information, but could only be established from owner anamnesis. Given the high proportion of bilaterally affected dogs referred for unilateral disease, owner knowledge cannot be considered sufficiently accurate to draw firm conclusions about disease progression.

Breed Influence

In our study, Labrador Retrievers and Rottweilers (62.4%) were the mostly commonly reported breeds compared with Bernese Mountain dogs, Rottweilers, German Shepherd dogs, and Labrador Retrievers (63.9%) in another report.33 This difference may reflect the geographic location of our clinic and the relatively low numbers of Bernese Mountain dogs in the United Kingdom. Any influence of differing breed ratios on interpretation of our results or comparison with other studies is unknown.

Radiographic Findings

Mediolateral extended and craniocaudal radiographic projections are the most commonly performed orthogonal imaging techniques for the canine elbow joint. To image the more complex anatomy of the elbow joint, including the MCP, special radiographic projections have been proposed to improve either sensitivity or reliability.2,18,34–38 These views reportedly improve visibility of the MCP by radiographically isolating it from other structures. We did not use these projections because of difficulty in standardizing them between cases and this may also represent a study limitation. However, despite the purported benefits of such views for improved visibility of the MCP (e.g., detection of free fragments) their sensitivity and specificity18 seemingly do not provide further consequential information predicting MCD or associated prognostic variables. We included a maximally flexed mediolateral projection that isolated the anconeal process from superimposition by the medial epicondyle. This projection allowed clear evaluation of osteophytes2,39 and exclusion of UAP. Our selection of radiographic projections may have decreased the sensitivity of assessment of fragmented MCP. Incongruency or subluxation of the humero-antebrachial joints was not assessed because it has been reported to be an insensitive and nonspecific method of measuring incongruence.40,41

Radiographic measures of osteophytosis are commonly used as components of the assessment criteria in elbow dysplasia screening schemes. Absence of osteophytosis in 28.8% of clinically affected dogs suggests that such screening tests may fail to identify some dogs with MCD. Recognition of STS in 86.7% of clinically affected dogs suggests that this may be a more sensitive indicator of MCD. However, as discussed, the lack of a control group of unaffected elbows and the lack of multiple observers for radiographic evaluation warrants further investigation to establish specificity of these diagnostic variables.

Shape assessment of the MCP is in our opinion is observer subjective. We considered a change in shape indicative of either osteophytosis or distortion of the cranial aspect of the MCP, but it must be noted that the assessment of a three-dimensional structure using a two-dimensional radiographic projection is suboptimal and likely unreliable.

Arthroscopic Findings

The common arthroscopic finding of kissing lesions of the medial aspect of the humeral condyle concurs with observations in previous reports.3,30 Presence of gross cartilage pathology involving the MCP and medial aspect of the humeral condyle in the absence of visible fragmentation of the MCP has been reported22,26,30,42,43; however, the cause of these lesions has not been reported. It is plausible that abnormal forces acting on normal cartilage or normal forces acting on diseased cartilage of the medial aspect of the elbow joint could explain the moderate correlation found between measures of cartilage lesions and periarticular osteophytosis. Danielson's study17 supports the former postulate of humeroulnar conflict44 and increased subchondral bone pressure (i.e., supraphysiologic pressure). Other authors have suggested that primary disease is a more likely contributor to cartilage pathology of the medial compartment.45 It seems likely based on our results that MCD and progression of cartilage erosion of the medial compartment is likely attributable to a common pathway.

Our recognition of elbow joints without radiographic signs of disease but with confirmed MCD on arthroscopic examination concurs with a previous report25 and suggests the need for further diagnostic tests when there are clinical signs indicative of MCD but no radiographic abnormalities. Presence of STS in 87% of elbows in our study is similar to the frequency reported after CT assessment of dogs with thoracic limb lameness attributable to elbow disease (not restricted to MCD) where 94% were considered affected by STS.46 The sensitivity and specificity of STS to differentiate MCD from disease-free elbows and elbows affected by other forms of elbow disease has not been reported. Subjective assessment of MCD using %STS is reportedly a sensitive method of predicting MCD47 but further clinical studies are warranted. In our opinion, the value of assessing STS is of greater importance in elbows where no other radiographic abnormality is evident.

Our results agree with previous comments that diagnosis of MCD is challenging when based on plain radiographs alone.46 We consider the presence of osteophytes on the radial head and anconeal process to be objective assessments whereas assessment of “mild” osteophytosis on the epicondylar ridges may be more subjective. Radiographic findings in a large group of dogs (263 dogs, 332 elbows) found fragmented MCP in 58 (17.5%) elbows48 whereas we identified 55 of 437 (12.6%) elbows as affected. Our lower frequency may be explained in part by inclusion of elbows with grossly intact MCP. Arthroscopic findings in 100 dogs (148 elbows),30 and after visual assessment during arthrotomy in 64 elbows26 in dogs with elbow lameness, predominantly but not exclusively attributable to MCD, are similar to our results. The MCP was grossly intact in 17/417 elbows (17.6%) compared with “chondromalacia-like” lesions previously reported in 27/148 (18.2%).30 Fissuring was identified in 80/437 elbows (18.3%) compared with reported incidences of 13/148 (8.7%)30 or 17/64 (26.6%).26 Fragmentation was identified in 280 elbows (64.1%) compared with reported incidences of 91/148 (61.5% including nondisplaced and displaced fragments)30 or 45/64 (70.3%) loose fragment, fissure/fragment and peripheral fragment.26 Potential explanations for variation between studies include the presence of a limited number of cases diagnosed with OCD in other studies and potential variation between direct visual and arthroscopically enhanced appraisal of the joint. Kissing lesions reported in 224 of 437(51.3%) elbows in our study are similar to another reports (83/148; 56.1%).30

Nine of 10 elbows with a modified Outerbridge score of 5 for the medial aspect of the humeral condyle had a fragment present suggesting that lesions of humeral subchondral bone are more likely when fragmentation is present. This does not necessarily infer that the fragment is the source of abrasion. It may in fact just be a sentinel of profound humeroulnar conflict, wherein fragmentation is an early occurrence before erosion of the remaining and more resilient portion of the MCP. We found moderate correlation of modified Outerbridge scores for the surfaces of the medial humeral condyle and the MCP.

The lack of correlation between duration of preoperative lameness with radiographic and arthroscopic findings was an unexpected finding. We hypothesized that there would be an increase in gross joint pathology with increased duration of clinical signs. Potential explanations for this contrary finding include variations between owners' assessment of lameness and the presence of bilateral lameness, delays between onset of lameness and diagnosis, potential insensitivity of arthroscopic findings as a measure of disease,20 and patient factors. Patient factors include individual variation in the response to pain, exercise regimes, weight, and the progression of pathologic change that may be influenced by individual genetics.49 Further investigation of these variables, including objective measurements of lameness duration and intensity is warranted. Our findings indicate that arthroscopic findings cannot be accurately predicted based on history and radiographic examination.


We found only weak or moderate correlations between extent of radiographic findings and arthroscopic findings. Thus, we rejected our hypothesis that increasing severity of quantifiable radiographic changes would be associated with increasing severity of arthroscopic findings. We also rejected our hypothesis that severity of quantified radiographic and/or arthroscopic findings would be associated with increasing duration of lameness identified preoperatively. Despite some reports that advocate usefulness of radiography as an indicator of MCP, we consider its primary indication to be exclusion of other causes of elbow-associated lameness, including traumatic, neoplastic, and other developmental conditions including OCD and UAP. Elbows considered radiographically “normal”, whether subjectively or objectively, can still be affected by MCD, the severity of which is difficult to predict based on clinical findings. We also conclude that duration of preoperative lameness is poorly correlated with arthroscopic and radiographic findings and that further studies documenting the association of clinical signs with CT, MRI, scintigraphy, synovial fluid analysis, arthroscopic, and histologic examination of MCP are needed.


This project did not receive external financial support and there is no proprietary interest.