COMPUTED TOMOGRAPHIC IDENTIFICATION OF DYSPLASIA AND

PROGRESSION OF OSTEOARTHRITIS IN DOG ELBOWS PREVIOUSLY

ASSIGNED OFA GRADES 0 AND 1

CHELSEA M. KUNST, ANTHONY P. PEASE, NATHAN C. NELSON, GREG HABING , ELIZABETH A. BALLEGEER

Elbow dysplasia is a heritable disease that is a common cause of lameness and progressive elbow osteoarthritis
in young large breed dogs. The Orthopedic Foundation for Animals (OFA) screens elbow radiographs, and
assigns grades 03 based on presence and severity of bony proliferation on the anconeal process. Grade 1 is
assigned when less than 3 mm is present and considered positive for dysplasia. We investigated the incidence of
elbow dysplasia and progression of osteoarthritis in elbows with grades 0 and 1 in 46 elbows screened at least
1 year previously, using CT as a gold standard and with the addition of CT absorptiometry. The incidence of
dysplasia based on CT was 62% in grade 0, and 75% in grade 1 elbows, all of which had medial coronoid disease.
Progressive osteoarthritis at recheck was consistent with elbow dysplasia. The sensitivity and specificity of
the OFA grade for elbow dysplasia compared to CT findings was 75% and 38%, respectively. Increased bone
mineral density of the medial coronoid process as characterized by osteoabsorptiometry warrants further
investigation with respect to elbow dysplasia. Proliferation on the anconeal process without CT evidence of
dysplasia or osteoarthritis was present in 20% of the elbows, and is theorized to be an anatomic variant or
enthesopathy of the olecranon ligament/synovium. Results of our study suggest that the anconeal bump
used for elbow screening by the OFA is a relatively insensitive characteristic, and support the use of CT for
identifying additional characteristics of elbow dysplasia. 
C 2014 American College of Veterinary Radiology.

Key words: canine orthopedic foundation for animals, computed tomography, elbow dysplasia, medial coronoid
process, osteoabsorptiometry.

Introduction condyle, and elbow incongruity.6 Elbow dysplasia can cause

E LBOW DYSPLASIA IS A common heritable disease in
dogs. Elbow dysplasia involves one or more of four dis-
tinct pathologic processes: medial coronoid disease (which
is the most common type of dysplasia,15 and includes but
is not limited to fragmented medial coronoid process), un-
united anconeal process, osteochondrosis of the humeral
From the Department of Small Animal Clinical Sciences (Pease, Nel-
son, Ballegeer); College of Veterinary Medicine, Michigan State Univer-
sity, East Lansing, MI 48823 (Pease); California Veterinary Specialists,
Murrieta, CA 92563 (Kunst); and Department of Veterinary Preventa-
tive Medicine, Ohio State University, Columbus, OH 43210 (Habing).
Funding sources: Michigan State University, College of Veterinary
Medicine Endowed Research Funds Project Grant.
Previous abstract: Kunst CM, Habing G, Ballegeer EA. CT identifica-
tion of dysplasia and progression of osteoarthritis in dog elbows pre-
viously assigned Orthopedic Foundation for Animals (OFA) grade 1.
2012 ACVR Annual Scientific Conference, Las Vegas, Nevada: Vet Rad
Ultrasound, Vol. 53, No. 6, 681682.
Previous presentation: 2012 ACVR Annual Scientific Conference, Las
Vegas, Nevada.
Address correspondence and reprint requests to Elizabeth A.
Ballegeer Department of Small Animal Clinical Sciences, Michigan
State University, East Lansing, MI 48823 at the above address. E-mail:
ballegee@cvm.msu.edu.
Received September 6, 2013; accepted for publication January 15,
2014.
doi: 10.1111/vru.12171
debilitating lameness, and arthroscopic treatment does not
palliate pain in all dogs.7 Additionally, elbow dysplasia
is characterized by progressive osteoarthritis,2,3,811 which
surgical treatment does not resolve.7 Therefore, it is impor-
tant to identify and remove affected dogs from the breeding
pool to decrease prevalence of elbow dysplasia in the canine
population.
The Orthopedic Foundation for Animals (OFA) in the
United States maintains a screening process of pure bred
dogs for elbow dysplasia. Dogs must be at least 2 years
of age and be registered with a recognized organization,
such as the American Kennel Club. The Foundation as-
sesses a single mediolateral radiographic projection of the
elbow during extreme flexion for the presence of bony pro-
liferation on the anconeal process, as this is often the first
place in the elbow that evidence of osteoarthritis occurs
secondary to a dysplastic lesion, and diagnosis of dyspla-
sia based on the presence of secondary signs has been well
documented.1,2,8,1221 The OFA grades range from 0 to 3. A
grade 0 is assigned when there is no evidence of proliferation
on the anconeal process, and is considered negative for the
presence of elbow dysplasia. Grades 13 are assigned when
Vet Radiol Ultrasound, Vol. 55, No. 5, 2014, pp 511520.

511
512 KUNST ET AL. 2014

FIG. 1. Flexed lateral radiographs of a grade 0 elbow (A) and grade 1 elbow (B) taken at Time 1 from the same animal, a 4-year-old female Bernese
Mountain Dog. Note the bony prominence along the proximal anconeal process in (B), depicted by the white arrow. This is the anconeal bump that is
measured for elbow grading by the Orthopedic Foundation for Animals (OFA). Both elbows were presumed to have medial coronoid process disease, based on
hyperattenuation of the coronoid process that was detected on CT only.

anconeal proliferation is present, and measure a height of
up to 3 mm, 35 mm, and over 5 mm, respectively. Grades
13 elbows are considered positive for elbow dysplasia and
American Kennel Clubs encourage breeders to not breed
dogs assigned a positive status. Figure 1 represents two el-
bows, one of each grades 0 and 1, from the same dog. Some
studies have theorized that proliferation on the anconeal
process may be a normal anatomic variant,17,22,23 which
could lead to false positive diagnoses in mildly affected el-
bows. Additionally, dogs with elbows assigned a grade 1
are commonly never lame,12,13 leading some breeders to
question their disease status. Breeders experience financial
loss if potentially healthy animals are removed from their
breeding stock. Alternatively, some dogs affected by elbow
dysplasia never develop proliferation on the anconeal pro-
cess, nor other evidence of elbow osteoarthritis.10,24,25 These
cases would lead to false-negative diagnoses of elbow dys-
plasia. Studies have demonstrated that CT is more sensi-
tive in detection of medial coronoid disease compared to
radiographs.8,22,2628 This is intuitively linked to the lack of
superimposition provided by cross sectional imaging. How-
ever, there is additional roughly quantitative bone density
information that can be afforded by its more precise deriva-
tion of X-ray attenuation. As medial coronoid disease has
often been associated with sclerosis of the subtrochlear
ulna,29 quantifying this bone density may provide some
more insight into the presence or absence of disease, for the
purpose of identifying dysplasia. Computed tomography
absorptiometry is a process that uses a tissue density equiv-
alent phantom to calibrate the attenuation values obtained
from a scan to mathematically derived bone mineral den-
sity (BMD) values (in units of mg K2 HPO4 /ml) that can
be compared between different machines,30 and account for
variability in X-ray beam energies during the scan.31
The purpose of this study was to determine and com-
pare the incidence of elbow dysplasia, using CT as the
gold standard, in elbows assigned OFA grades 1 and 0.
We also assessed the relationship of osteoarthritis progres-
sion with the incidence of elbow dysplasia by comparing
the initial OFA images to radiographs taken at least 1 year
subsequent to initial diagnosis. As a possible additional
correlation with medial coronoid process disease, we also
investigated the attenuation of the medial coronoid process
with CT absorptiometry.
Materials and Methods
Client-owned dogs were recruited using the OFA
database. The study was approved by the Michigan State
University Animal Care and Use Committee. Inclusion cri-
teria for these patients included screening by the Founda-
tion at least 1 year previously (indicated as Time 0), with
elbows assigned a grade 1 in one elbow, and a 1 or 0 in
the opposite elbow. Exclusion criteria were animals that
were clinically ill, pregnant, or deemed a risk for sedation.
Animals were not excluded for lameness. A history was
obtained from the owners, including whether the patients
ever experienced lameness, and dogs were given a complete
physical exam.
An IV injection of 0.005 mg/kg of dexmedetomidine
hydrochloride (Orion Corporation, Espoo, Finland) and
0.2 mg/kg butorphanol tartrate (Fort Dodge, Iowa) was
VOL. 55, NO. 5 ELBOW DYSPLASIA IN OFA GRADE 0 AND 1 ELBOWS
administered in a cephalic vein. Once sedated, the patients
were placed in dorsal recumbency on the CT couch with a
calibration phantom (Mindways R
13002 Model 3 CT cali-
bration phantom, Mindways Software, Inc., Austin, Texas)
placed underneath the patient for attenuation correction.
The phantom is composed of five parallel rods of differ-
ent reference materials: water, K2 HPO4 in concentrations
of 50, 100, 200 mg/ml, and a high-density polyethylene.
The forelimbs were positioned such that there was approx-
imately 90 degree flexion of both the shoulders and elbows
and secured in place with foam padding and porous ban-
dage tape. Using a 16 row multidetector CT machine (GE
Brightspeed, General Electric Company, Milwaukee, WI),
transverse, 0.625 mm slice thickness images of the elbows
were acquired in a high-pass algorithm with a collimator
pitch of 1, table speed of 13.75 mm/s, tube rotation time of
1 revolution/s, kV of 120, and mA of 250. Field of view was
tailored to the patient, to include the phantom and both el-
bows on initial scan. Reformatted transverse narrow field
views of each elbow were constructed simultaneous with
large field of view images. Dorsal and sagittal multiplanar
reformatted images of the elbows were performed post ac-
quisition, at the time of interpretation. Once the CT was
completed, three digital radiographic projections (Canon
CXDI-70 C Wireless digital plate, Canon USA, Melville,
NY) of each elbow were obtained: one mediolateral pro-
jection with the elbow in approximately 90 degree flexion,
another mediolateral projection during extreme flexion for
optimal visualization of the anconeal process, and a cranio-
caudal projection of the extended elbow. Exposure ranged
from mAs 2.53.2, and kVp of 6872 based on thickness of
the elbow. After the radiographs were obtained, the origi-
nal sedation was reversed with an intramuscular injection
of atipamizole hydrochloride (Orion Corporation, Espoo,
Finland) dosed at the same volume as the dexmedetomidine
hydrochloride.
Digitized copies of the radiographs originally graded by
the OFA at Time 0 were obtained. Of these, 10 elbows were
originally digital in nature, with the remainder (36 elbows)
scanned on a Lumisys Lumiscan 75 radiograph digitizer
(Lumisys Inc, Sunnyvale, CA). The radiograph and CT
images taken at our facility will be referred to as being
acquired at Time 1.
All identifying information was removed from the im-
ages from Time 0 and Time 1, which were then individ-
ually randomized and independently assessed by two of
the authors (C.M.K. and E.M.B.) in a blinded manner.
On radiographs, osteophytes and the height of the prolif-
eration on the anconeal process were measured as previ-
ously described,23 which applied measurements in thick-
ness, defined as the dimension perpendicular to the cortex.
Reviewers were allowed to zoom and manipulate window
and level of images as desired. Images were subjectively
assessed for the presence of linear lucency at the base of
513
the anconeal process, indicating ununited anconeal process,
medial humeral condylar lucency indicating cartilage and
subchondral bone defect (osteochondrosis or erosion from
medial coronoid process disease), blunting or indistinct-
ness of the cranial margin of the medial coronoid process,
and incongruity, seen as a separation of humeroradial and
humeroulnar joints on neutral lateral view. On the radio-
graphs acquired at Time 0 and Time 1, osteophytes were
recorded and scored according to the process used by the
International Elbow Working Group.32 Per this method,
no evidence of osteophytosis is assigned a score of 0, and
scores 13 are assigned to the greatest osteophyte height
within an elbow regardless of location. Scores 1, 2, and
3 are assigned to osteophyte height of up to 2 mm, be-
tween 2 and 5 mm, and over 5 mm, respectively, located on
any surface of the elbow. These scores were used to monitor
progression of elbow osteoarthritis on radiographs taken at
Time 0 and Time 1. Proliferation on the anconeal process
on radiographs acquired at Time 1 was graded according to
height based on the OFA screening process. Thus, grade 0
is assigned to elbows with no proliferation on the anconeal
process, and grades 1, 2, and 3 are assigned for anconeal
proliferation height of up to 3 mm, 35 mm, and over 5
mm, respectively.
Computed tomography images were also assessed, with
reformatted images able to be adjusted by the viewer to
obtain dorsal and sagittal planes precisely parallel or per-
pendicular to the plane of the ulna on Voxar 3D soft-
ware (Toshiba Medical Visualization Systems Europe Ltd.
Edinburgh, UK), and variable zoom and viewer-preferred
window and level. Dorsally oriented multiplanar reformat-
ted CT images were assessed for the presence of prolifera-
tion on the anconeal process, as previously described.17,22
If proliferation was present, the height was measured using
either the aforementioned Voxar 3D software (if a plane
different from standard multiplanar reconstruction was
needed), or PACS-associated viewing software (Horizon
Rad Station, McKesson, San Francisco, CA) as the height
from the base of the medial aspect of the process (which was
consistently flat) to the highest point of the proliferation.
Images were subjectively assessed for the presence of linear
hypoattenuation at the base of the anconeal process, indi-
cating ununited anconeal process, medial humeral condylar
hypoattenuation indicating cartilage and subchondral bone
defect (osteochondrosis or erosion from medial coronoid
process disease), and significant (>1.3 mm) incongruity
of the radioulnar articulation at the humeral condyle, as
seen on sagittally reformatted images, at the base of the
coronoid.33
On the CT images, the medial coronoid process was con-
sidered diseased if the process was fragmented or fissured,
sclerotic with loss of trabecular medullary pattern, hypoat-
tenuating, and/ or irregularly margined or shaped, as de-
scribed previously.4,9,10,34 If a discrepancy in interpretation
514
FIG. 2. Demonstration of the ellipsoid region of interest drawn on the transverse plane CT images at the slice location with the largest medial coronoid
process height for the purpose of recording the regional Hounsfield units. (A) is an elbow with a presumed normal medial coronoid process, and (B) is an elbow
with a presumed diseased medial coronoid process.
was present between reviewers, a consensus was reached
after mutual examination of images; these were mostly lim-
ited to minimal differences in height measurements, and
subjective assessments of sclerosis or medial coronoid pro-
cess blunting, as seen on Time 0 radiographs. In addition
to subjective assessment, a similar-sized ellipsoid region of
interest was hand drawn on transverse plane images at the
slice of largest medial coronoid process height to include
the subchondral bone of the ulna in the medial coronoid
process and ulnar notch (Fig. 2) to more quantitatively
measure Hounsfield units. Care was taken to include only
the perceived medullary bone, and not cortical bone, by
following visible cortices adjacent to the nonhyperattenu-
ating medulla, and excluding the bone following the same
thickness along the entire cortex. Additionally, the region
of interest edge was placed at the corticomedullary junction
(when visible) and the caudal extent of the region of interest
only to the level of the lateral coronoid process. Additional
regions of interest (ROIs) were drawn on each bar of the
previously described phantom, taking care not to include
edges, or recognizable beam-hardening artifact.
Statistical Analysis
All statistics were performed by an epidemiologist (G.H.)
using commercially available statistics software (SAS v. 9.2,
Cary, NC). All p values for significance were set at 0.05.
Descriptive statistics were calculated, including the pro-
portion of elbows at each time point with each grade, os-
teoarthritis, and fragmented medial coronoid process. Ad-
ditionally, linear correlation between height of proliferation
KUNST ET AL. 2014
on the anconeal process on radiographs obtained at Time
1 compared to the CT images was performed using the
PROC CORR procedure. Attenuation correction for ul-
nar Hounsfield units on CTs was performed by translating
Hounsfield units to bone mineral densities with phantom
attenuation corrections, as a more accurate means of re-
porting bone density, as previously described.30,35,36 Four
elbows were excluded due to phantom absence from the
scan. The remaining 42 BMD values were calculated for
each elbow from attenuation measurements from phan-
tom bars, linear regression slope and intercept values, and
the equation: BMD = (region of interest value inter-
cept)/slope. To test differences in the mean BMD between
dogs with and without medial coronoid process disease, a
generalized linear model was constructed using the PROC
GENMOD procedure. To account for the expected non-
independence of elbows within dogs, a repeated statement
was included for dog. Bone mineral density was tested for
correlation with age, separated into separate elbows to re-
move the correlation of elbows within dogs, by calculat-
ing the Pearson correlation coefficient using PROC CORR
procedure. Exact confidence intervals for the proportions,
including sensitivity and specificity, were calculated using
the PROC FREQ procedure.
Results
Twenty-three dogs met the inclusion criteria. They
ranged in age from 3.2 to 8.7 years (mean 5.6 years old).
Ten breeds were represented: six Golden Retrievers, four
Bernese Mountain dogs, three each of English Springer
VOL. 55, NO. 5 ELBOW DYSPLASIA IN OFA GRADE 0 AND 1 ELBOWS 515

FIG. 3. Dorsal plane CT maximum intensity projection images of the anconeal processes of a (A) presumed healthy elbow, and (B) an elbow with a
fragmented medial coronoid process. Note the increased height of the lateral aspect of the anconeal process in (B), as represented by the vertical black line,
and the osteophyte along the medial aspect of the medial coronoid process (white arrow).

Spaniels and Rottweilers, two Belgian Shepherds, and one
of each Labrador Retriever, German Shepherd dog, Rhode-
sian Ridgeback, English Setter, and Newfoundland. There
were nine intact females, four spayed females, eight intact
males, and two neutered males. The time ranging from Time
0 to Time 1 was 15.6 years (mean 2.9 years). None of the
clients reported lameness at Time 1. Ten dogs were assigned
a grade 1 in both elbows by the OFA at Time 0, and 13
were assigned a grade 1 in one elbow and a 0 in the oppo-
site elbow, thus, 33/46 (72%) elbows included in the study
were assigned a grade 1, and 13/46 (28%) were assigned a
grade 0.
On CT, 33/46 (72%) of all assessed elbows had evidence
of elbow dysplasia on CT. Of these 33, 9/46 (20%) had
fragmented medial coronoid process. The remaining elbows
(24/46, or 52%) had other evidence of medial coronoid dis-
ease, to include hyperattenuation (n = 14), hypoattenuation
(n = 3), irregular margination or shape (n = 2), or atten-
uation and shape changes (n = 5). Of these elbows, 9/24
(38%) had periarticular osteophyte formation beyond the
anconeal bump. 13/46 (28%) had no evidence of elbow
dysplasia. Other types of presumed elbow dysplasia were
identified in five elbows (three with medial humeral condy-
lar lesions consistent with osteochondrosis, and two with
partial ununited anconeal processes) on the radiographs
and CT performed at Time 1. Only one of the three elbows
with medial humeral condylar lesions had a separate me-
dial coronoid fragment. However, all five elbows also had
evidence of medial coronoid process disease on CT, and as
such were not categorized differently. For this reason, dis-
tinction between medial humeral condylar osteochondro-
sis and a cartilage erosion from adjacent medial coronoid
disease was not attempted. One elbow was not dysplastic,
however, had mild osteoarthritis and presumed triceps en-
thesopathy, characterized by mild proliferation on the most
proximal aspect of the olecranon. On the CT images, when
proliferation on the anconeal process was present, it was
located consistently on the lateral aspect and best visual-
ized on the dorsal multiplanar reformatted images (Fig. 3).
Correlation of measurement of height of proliferation on
the anconeal process between radiographs and CT images
acquired at Time 1 was good (R2 = 0.85).
Bone mineral density values calculated from corrected
attenuation values were significantly (P = 0.0074) corre-
lated with the subjective presence of medial coronoid dis-
ease, meaning those classified having the disease typically
had higher mineral density within the medial coronoid pro-
cess and medullary ulna just caudal to it. Figure 4 demon-
strates the BMD values and classification. All animals with
fragments and BMD calculations (n = 8) had BMD val-
ues greater than 1005.6 mg K2 HPO4 /ml, while those with-
out medial coronoid process disease had values less than
722.5 mg K2 HPO4 /ml (mean of 596.3.) 19 elbows with
subjectively hyperattenuating processes without fragments
had bone mineral densities ranging from 739.2 to 1286.4
mg K2 HPO4 /ml. Conversely, there were three elbows clas-
sified with medial coronoid process disease with subjec-
tively focally hypoattenuating processes, with bone min-
eral densities ranging from 548.3 to 662 mg K2 HPO4 /ml.
Overlap existed between this category and normal elbows.
Elbows with osteoarthritis had a large range of bone
mineral densities, from 505.7 to 1404.4 mg K2 HPO4 /ml.
516
FIG. 4. Scatter plot demonstrating the ranges of calculated bone mineral densities compared to the presumed disease status. MCPD = Medial coronoid
process disease.
Borderline significant negative association of BMD with
age was found in the left elbow (P = 0.059) but not the right
(P = 0.089).
When broken down into grade 1 vs. grade 0 elbows, 25/33
(76%) elbows originally assigned a grade 1 were diagnosed
with elbow dysplasia on CT: 7/25 (28%) had a fragmented
medial coronoid process, 18/25 (72%) had other evidence
of disease of the medial coronoid process. Eight (24%) of
the 33 original grade 1 elbows had no evidence of elbow
dysplasia. Of the 13 elbows originally assigned a grade 0
on radiographs, 8/13 (62%) had signs of elbow dysplasia
with fragmented medial coronoid process (2/8, 25%) or ev-
idence of medial coronoid process disease (6/8, 75%), and
5/13 (38%) did not have evidence of medial coronoid dis-
ease, nor osteoarthritis (Fig. 5). Sensitivity and specificity
of OFA assignment of grades 0 and 1 of the study popula-
tion compared to diagnosis of elbow dysplasia based on CT
images were 75% (Confidence Interval 5486%) and 38%
(Confidence Interval 1945%), respectively. Of the twelve
elbows that were neither dysplastic nor arthritic on CT,
proliferation on the anconeal process was often present,
in 7/7 (100%) of the grade 1 and 2/5 (40%) of the grade
0 elbows. On one Time 0 grade 0 elbow, a bump on the
anconeal process was detected on radiographs but not on
CT. Therefore, 9/46 (20%) elbows included in the study
had proliferation on the anconeal process without evidence
of elbow dysplasia nor additional osteoarthritis. Based on
the International Elbow Working Group osteophyte mea-
surement and scoring method, nine elbows at Time 0 had
additional evidence of osteoarthritis beyond proliferation
on the anconeal process, only three of which progressed to
KUNST ET AL. 2014
a higher score at Time 1. All three were originally assigned
a grade 1, and on CT had evidence of dysplasia.
Discussion
We compared canine elbows assigned OFA grades 0 and
1 to the presence of dysplasia as determined by CT, and
found that the incidence of dysplasia is not significantly
different between the two groups. Progressive osteoarthri-
tis at recheck at least 1 year after initial evaluation was
positively correlated with elbow dysplasia. Multiple elbows
that had bony proliferation on the anconeal process did not
have evidence of dysplasia on CT or osteoarthritis in other
areas of the elbow joint. Increased attenuation of the medial
coronoid process as determined by CT osteoabsorptiome-
try is correlated with medial coronoid disease, though this
requires further investigation, given the diseased classifica-
tion was often made based on increased attenuation.
The lack of difference in prevalence of presumed dys-
plasia between differing grades is similar to two studies
specifically involving Belgian Shepherds and Labrador Re-
trievers, comparing elbows with Finnish grades 0 and 1
to CT findings regarding dysplasia.17,22 The Finnish grad-
ing system is similar to the OFA system in that grading
is dependent on the height of proliferation on the an-
coneal process. The two differences between the Finnish
and OFA screening processes are (i) the minimum age is
12 months compared to 24 months with the OFA, and
(ii) grade 1 is assigned to anconeal proliferation up to
2 mm with the Finnish system compared to 3 mm with
OFA. The sensitivity and specificity of the radiographic
VOL. 55, NO. 5 ELBOW DYSPLASIA IN OFA GRADE 0 AND 1 ELBOWS 517

FIG. 5. Distribution of presumed disease states of Grades 0 and 1 elbows based on CT findings at Time 1. Disease status of elbows originally graded Grade
0 and 1 elbows is represented as percentages, to demonstrate similarity of distribution. Note that while there were more grade 1 elbows overall, the relative
distribution of those that did not have dysplasia (on the bottom of bars), and fragments (middle), or otherwise diseased medial coronoid processes (top) are
similar, though not identical. Numbers within the bars represent the number of elbows for each category.

Finnish screening process and using CT as a gold standard
was 40% and 29% in Belgian Shepherds,22 and 92% sen-
sitivity and 79% specificity in Labrador Retrievers.17 The
sensitivity and specificity of the OFA screening process with
gold standard of CT was 75% and 38% in our study. There
are multiple possible reasons as to why there are large dif-
ferences between the calculations from each study. The first
is the 1 mm difference in absolute measurement of the an-
coneal bump. Second, proliferation on the anconeal process
is theorized to be an anatomic variant,17,22,23 which may
be breed-dependent and vary greatly between breeds. This
breed variation also relates to the prevalence of medial coro-
noid process disease, specifically that the previous studies
measured Belgian shepherds22 which have low prevalence of
medial coronoid disease, and Labrador Retrievers17 , which
have high prevalence. Diseased status could be incidentally
present with the bump in the more prevalent breed, rather
than true causation. Last, the OFA minimum age is 24
months, thus allowing progression of proliferation making
it more easily detectable, as it has been documented that
with age, osteophytosis progresses and is better detected
radiographically.22,3739
Differences exist in the CT criteria used in this study and
previous studies to diagnose medial coronoid disease. In
the previous study involving Belgian Shepherds, inclusions
for a normal elbow on CT were no signs of fragmentation
or fissure formation, lack of subtrochlear sclerosis, and nor-
mal medial coronoid process shape. Five of the 36 elbows
in their study had fragmented medial coronoid process, the
remainder was considered normal, and no mention was
made to any other changes of the medial coronoid pro-
cesses. In our study, 30 of the 46 elbows were diagnosed
with disease of the medial coronoid process, 10 (33%) were
fragmented and 20 (67%) were otherwise diseased. False-
negative diagnosis of diseased medial coronoid processes
on the previous study or false positive diagnosis on our
study, possibly from differences in acquiring the CT im-
ages or interpretation methods, could lead to the variation
of sensitivities calculated. Given the large variation in me-
dial coronoid disease as interpreted by orthopedic surgeons
upon surgical examination of the bone, it seems oversimpli-
fied to simply classify only the fragmented ones as diseased.
A larger range of inclusion for those conditions reported
as abnormal by surgeons and previously reported changes
was deemed more appropriate.
Subchrondral bone sclerosis, as represented by increased
attenuation and loss of trabecular pattern, has previously
been correlated with medial coronoid process disease.29
It has also been correlated, as represented by BMD val-
ues from CT osteoabsorptiometry, with osteoarthritis in
horses,30,36 humans,40 and cats.41 This increased medullary
cavity BMD, to the authors knowledge, has yet to be
reported in correlation with medial coronoid disease, al-
though values have been reported in clinically normal
dogs.32,42 Our technique was slightly different from those
previously reported, as we attempted to exclude subchon-
dral cortical bone from measurements. True exclusion of
cortical bone, which may possibly be thickened with me-
dial coronoid disease, on attenuation ROIs placed in the
medullary cavity of the medial coronoid process, cannot
be proven. Even so, this was done as the most disparate
area visually between sclerotic and normal was the medulla
of the ulna, and cortical inclusion was suspected to artifi-
cially raise attenuation of those with a visible medullary
518 KUNST ET AL.
cavity, and precludes good comparison with previous
reports. Nonetheless, all patients with fragments had higher
values than those without medial coronoid disease. A gray
area arises, however, in that some of those elbows classified
with medial coronoid disease were categorized as such be-
cause of their sclerosis, rather than the presence of a true
fragment, and some with more focal hypoattenuating areas
have densities that overlap considerably with elbows clas-
sified as normal. In fact, increased density of the medial
coronoid process/subtrochlear notch may be associated
with an aging change only,42 as suggested by weak negative
correlation with age in our study. Though the presence of
additional osteoarthritis in some of the elbows may argue
against this, all these compounding factors emphasize the
need for further study comparing attenuation in those with
surgically or histopathologically proven medial coronoid
disease or lack thereof, rather than CT only.
We found that proliferation on the anconeal process was
best detected on CT using the dorsally reformatted mul-
tiplanar images, similar to previous studies.17,22 The pro-
liferation was consistently found on the lateral aspect of
the process (see Fig. 3). Twenty percent of the elbows in
this study had proliferation on the anconeal process, seen
on radiographs and CT, without evidence of elbow dyspla-
sia, nor osteoarthritis elsewhere in the elbow. Such findings
may lead to false-positive diagnosis of elbow dysplasia in
radiographic screening programs that rely solely on detec-
tion of secondary lesions. This finding is also similar to
previous studies that theorized an anatomic variant may be
responsible.17,22 The International Elbow Working Group
has also discussed this possibility.21,23,25
Several theories may elucidate the asymmetry of the an-
coneal process seen in otherwise normal elbows. One the-
ory is that the proliferation seen radiographically may be
secondary to an anatomic variant combined with a small
amount of obliquity.23 Obliquity is unlikely to be the case
in our study, as the proliferation was confirmed with CT
imaging. The ability of CT to create multiplanar reformat-
ted images in different planes removes the possibility of
artifactually creating a bump where none exists. Digital ra-
diographs are thus likely a close representation of true an-
coneal shape, provided that positioning standards for the
mediolateral radiograph are adhered to.
One anatomic study described the topography and
course of the olecranon ligament in dogs and postulated
that it is associated with proliferation on the anconeal
process.43 A second study demonstrated that the prolifera-
tion on the anconeal process in elbows with osteoarthritis
secondary to elbow dysplasia is along the site of attach-
ment of the olecranon ligament and joint capsule16 , though
a third states it is not at the olecranon ligament, but at syn-
ovial attachment only.44 None of these studies described
laterality of the proliferation on the anconeal process. Con-
sidering that the proliferation occurs on the lateral aspect
2014
of the process in presumed dysplastic and nondysplastic
elbows in our study, the theory that proliferation in all el-
bows may be secondary to enthesopathy of the olecranon
ligament and/ or joint capsule is supported.
A different study described variation of ossification of
the canine anconeal process, including that in some elbows,
a separate center of ossification occurs in the lateral part of
the process.45 Additionally, the size and shape of the lateral
part of the process can be larger and more convex compared
to the medial side. They also note that the craniolateral
aspect of the anconeal process is exposed to higher loading
forces, which may explain the earlier onset of ossification
and advanced ossification process laterally. Unfortunately,
this study was limited to neonatal to 44-week-old dogs.
The laterally of the proliferation identified in our study is
supportive of these theories, and may explain the presence
of the proliferation in dogs with otherwise normal elbows
as determined by CT.
To the authors knowledge, an anatomic study of shape
of mature elbows specifically looking at asymmetry of the
anconeal process does not exist. Ideally, a study comparing
CT images to anatomic and histologic findings in normal
and dysplastic elbows should be performed, and until then,
the implication of proliferation on the anconeal process in
elbows that are otherwise normal is unknown.
We compared osteoarthritis score in Time 0 and Time 1
acquired radiographs in order to potentially identify occult
elbow dysplasia not seen on CT, as has been reported.26,29
Only three elbows demonstrated increased score from Time
0 to Time 1, all of which were OFA grade 1 and dysplas-
tic on CT. Thus monitoring of osteoarthritis progression
was not helpful in identifying occult elbow dysplasia be-
yond more obvious CT findings, although if documented
clinically would strongly suggest the presence of elbow dys-
plasia. Arthroscopy would have been the best method to
confirm normal CT findings on our dogs, however would
have introduced undesired morbidity in our non-lame
population.
Radiographic findings in OFA grade 1 elbows are sub-
tle, presenting a challenge to clinicians in discussion with
the owners about its implications, particularly when the
animal in question may have never exhibited lameness.
Consequently, the International Elbow Working Group has
added additional findings consistent with elbow dysplasia
to grading in 2010. However, they have also experienced
an increasing number of requests from owners to include
CT in a re-evaluation process of dogs previously assigned
grade 1 elbow scores.21 At this point, a CT protocol has not
yet been established and neither group allows re-evaluation
with CT. In this study, presence (or absence) of dyspla-
sia as determined by CT differed from the OFA findings
for 16/46 (39%) elbows, which suggests that the use of
CT in a re-evaluation process would be valuable. More re-
search needs to be performed to establish a standard CT
VOL. 55, NO. 5 ELBOW DYSPLASIA IN OFA GRADE 0 AND 1 ELBOWS
protocol and validate it with arthroscopic and histologic
findings.
Several limitations are present in this study. The main
limitation is lack of arthroscopic and/ or histologic confir-
mation of diagnosis of diseased versus nondiseased elbows.
However, given the lack of clinical lameness in the patients,
this was not feasible. Original Foundation radiographs were
often digitally scanned, and positioning done by other vet-
erinarians, thus their quality and the resulting interpreta-
tion, as seen with discrepancy between reviewers, often af-
fected. Positioning of the elbows without additional body
parts in the CT gantry, particularly the head, to avoid beam
hardening artifact,46,47 was not considered possible in our
study as the dogs were all large breed and only sedated, not
anesthetized, preventing the extreme neck flexion necessary
to remove the head and neck from the same cross section
as the elbow. This contributed significant beam harden-
ing to acquired CT images, and though care was taken to
exclude visible artifact, could significantly affect the atten-
uation values for osteoabsorptiometry. Another limitation
concerns the population of the dogs. First, it is a small
population of small numbers of varied breeds, thus breed
REFERENCES
1. Grondalen J, Grondalen T. Arthrosis in the elbow joint of young
rapidly growing dogs. V. A pathoanatomical investigation. Nord Vet Med
1981;33:116.
2. Olsson SE. The early diagnosis of fragmented coronoid process and
osteochondritis dissecans of the canine elbow joint. J Am Anim Hosp Assoc
1983;19:616626.
3. Wind AP. Incidence and radiographic appearance of fragmented coro-
noid process. California Veterinarian 1982;36:1925.
4. Van Ryssen B, Van Bree H. Arthroscopic findings in 100 dogs with
elbow lameness. Vet Rec 1997;126:360362.
5. Lavrijsen IC, Heuven HC, Voorhout G, et al. Phenotypic and ge-
netic evalutation of elbow dysplasia in Dutch Labrador Retrievers, Golden
Retrievers, and Bernese Mountain Dogs. Vet J 2010;193:486492.
6. Palmer RH. Clinical examination of the dog with elbow lameness. Pro-
ceedings 25th annual meeting of the International Elbow Working Group 2010,
Bologna, Italy.
7. Tobias TA, Miyabayashi T, Olmstead ML, Hendrick LA. Surgical
removal of fragmented medial coronoid process in the dog: comparative
effects of surgical approach and age at time of surgery. J Am Anim Hosp
Assoc 1994;30:360368.
8. Carpenter LG, Schwartz PD, Lowry JE, Park RD, Steyn PF. Com-
parison of radiologic imaging techniques for the diagnosis of fragmented
medial coronoid process of the cubital joint in dogs. J Am Vet Med Assoc
1993;203:7883.
9. Walde I, Tellhelm B. Fragmented medial coronoid process of the ulna
(FCP) and osteochondritis dissecans (OCD) of the canine elbow joint and
hock joint- literature review, diagnosis, and therapy. Wien Tierarztl Monatss-
chr 1991;78:414424.
10. Meyer-Lindenberg A, Langhann A, Fehr M, Nolte I. Prevalence of
fragmented medial coronoid process of the ulna in lame adult dogs. Vet Res
2002;151:230234.
11. Macpherson GC, Lewis DD, Johnson KA, Allen GS, Yovich JC.
Fragmented coronoid process associated with premature distal radial physeal
closure in four dogs. Vet Comp Orthop Traumatol 1992;5:9399.
12. Grondalen J. Arthrosis in the elbow joint of young rapidly growing
dogs. VI. Interrelation between clinical, radiographical, and pathoanatomical
findings. Nord Vet Med 1982;34:6575.
519
comparisons could not be made. Second, the study likely
attracted those owners whose dogs received a grade 1,
however, were never clinically lame, and wanted confir-
mation (or contradiction) of the grade. Thus, a bias to-
ward a nonlame population likely selected for more ani-
mals free of disease and those that were falsely diagnosed at
Time 0.
In conclusion, this is the first study investigating inci-
dence of elbow dysplasia in elbows graded 1 by the OFA.
Our results indicated that incidence of elbow dysplasia, par-
ticularly disease of the medial coronoid process, in elbows
assigned grade 0 and grade 1 by the Foundation based
on radiographs, was not significantly different between the
two grades, as seen on CT. Additionally, and similar to
previous reports, proliferation on the anconeal process was
found to exist in elbows without CT evidence of dysplasia
or osteoarthritis elsewhere in the elbows. We theorize that
this may be enthesopathy of the olecranon ligament and/
or joint capsule, or an anatomic variant. In future studies,
CT absorptiometry may be useful for improving diagnostic
sensitivity in elbows with suspected disease of the medial
coronoid process.
13. Read RA, Armstrong SJ, OKeefe JD, Eger CE. Fragmentation of
the medial coronoid process of the ulna in dogs: a study of 109 cases. J Small
Anim Pract 1990;31:330334.
14. Pederson NC, Wind A, Morgan JP, Pool RR. Joint diseases of dogs
and cats. In: Ettinger SJ, (eds): Textbook of Veterinary Medicine. Philadel-
phia: W.B. Saunders, 1989;23292377.
15. Berry CR. Evaluation of canine elbow for fragmented medial coro-
noid process. Vet Radiol Ultrasound 1992;33:273276.
16. Keller GG, Kreeger JM, Mann FA, Lattimer JC. Correlation of
radiographic, necropsy and histologic findings in 8 dogs with elbow dysplasia.
Vet Radiol Ultrasound 1997;38:272276.
17. Lappalainen AK, Molsa S, Liman A, Snellman M, Laitinen-
Vapaavuori O. Evaluation of accuracy of the Finnish elbow dysplasia screen-
ing protocol in Labrador retrievers. J Small Anim Pract 2013;54:195200.
18. Hornof WJ, Wind AP, Wallack ST, Schulz KS. Canine elbow dys-
plasia. The early radiographic detection of fragmentation of the coronoid
process. Vet Clin North Am Small Anim Pract 2000;30:257267.
19. Olsson SE. Lameness in the dog. Proceedings of the Am Anim Hosp
Assoc 1975;42:363370.
20. Olsson SE. Pathophysiology, morphology and clinical signs of osteo-
chondrosis (chondrosis) in the dog. In: Bojrab MJ, (ed.): Pathophysiology in
Small Animal Surgery. Philadelphia: Lea & Febiger, 1983;604617.
21. Tellhelm B. The IEWG screening protocol for elbow dysplasia. Pro-
ceedings of the 22nd Annual Meeting of the International Working Group;
2007; Munich, Germany.
22. Lappalainen AK, Molsa S, Liman A, Laitinen-Vapaavuori O, Snell-
man M. Radiographic and computed tomography findings in Belgian Shep-
herd dogs with mild elbow dysplasia. Vet Radiol Ultrasound 2009;50:364
369.
23. Audell L. Scoring for secondary lesions. Proceedings of 20th Annual
Meeting of the International Elbow Working Group; 2005; Munich, Germany.
24. Punke JP, Hulse DA, Kerwin SC, et al. Arthroscopic documentation
of elbow cartilage pathology in dogs with clinical lameness without changes
on standard radiographic projections. Vet Surg 2009;39:209212.
25. Dennis R. The BVA/KC Elbow dysplasia grading scheme. Proceedings
of the27th Annual Meeting of the International Elbow Working Group; 2012;
Birmingham, UK.
520 KUNST ET AL.
26. Moores AP, Benigni L, Lamb CR. Computed tomography versus
arthroscopy for detection of canine elbow dysplasia lesions. Vet Surg 2008;
37: 390398.
27. Van Bree H, Van Ryssen B. Imaging the canine elbow: radiog-
raphy, computed tomography and arthroscopy. Vet Annu 1995;35:118
129.
28. Reichle JK, Snaps F. The elbow. Clin Tech Small Anim Pract
1999;14:177186.
29. Groth AM, Benigni L, Moores AP, Lamb CR. Spectrum of com-
puted tomographic findings in 58 canine elbows with fragmentation of the
medial coronoid process. J Small Anim Pract 2009;50:1522.
30. Young BD, Samii VF, Mattoon JS, Weisbrode SE, Bertone AL. Sub-
chondral bone density and cartilage degeneration patterns in osteoarthritic
metacarpal condyles of horses. Am J Vet Res 2007;68:841849.
31. Birnbaum BA, Hindman N, Lee J, Babb JS. Multi-detector row CT
attenuation measurements: assessment of intra- and interscanner variability
with an anthropomorphic body CT phantom. Radiology 2007;242:109119.
32. Draffan D, Carrera I, Carmichael S, Heller J, Hammond G. Radio-
graphic analysis of trochlear notch sclerosis in the diagnosis of osteoarthri-
tis secondary to medial coronoid disease. Vet Comp Orthop Traumatol
2009;22:715.
33. Kramer A, Holsworth IG, Wisner ER, Kass PH, Shulz KS. Com-
puted tomographic evaluation of canine radioulnar incongruence in vivo. Vet
Surg 2006;35:2429.
34. Reichle JK, Park RD, Bahr AM. Computed tomographic findings of
dogs with cubital joint lameness. Vet Radiol Ultrasound 2000;41:125130.
35. Samii VF, Les CM, Schulz KS, Keyak JH, Stover SM. Computed
tomographic osteoabsorptiometry of the elbow joint in clinically normal
dogs. Am J Vet Res 2002;63:11591166.
36. Olive J, dAnjou MA, Alexander K, Beauchamp G, Theoret CL. Cor-
relation of signal attenuation-based quantitative magnetic resonance imag-
ing with quantitative computed tomographic measurements of subchondral
bone mineral density in metacarpophalangeal joints of horses. Am J Vet Res
2010;71:412420.
2014
37. Swenson L, Audell L, Hedhammer A. Prevalence and inheritance of
and selection for elbow arthrosis in Bernese Mountain Dogs and Rottweilers
in Sweden and benefit: cost analysis of a screening and control program. J
Am Vet Med Assoc 1997;210:215221.
38. Maki K, Liinamo AE, Ojala M. Estimates of genetic parameters for
hip and elbow dysplasia in Finnish Rottweilers. J Anim Sci 2000;78:1141
1148.
39. Malm S, Fikse WF, Danell B, Strandberg E. Genetic variation and
genetic trends in hip and elbow dysplasia in Swedish Rottweiler and Bernese
Mountain Dog. J Anim Breed Genet 2008;125:403412.
40. Kawasaki T, Sashi R, Moriya S, et al. Computed tomography os-
teoabsorptiometry for assessing the density distribution of subchondral bone
as a measure of long-term mechanical stress in the rugby shoulder. J Shoul-
der Elbow Surg 2013;22:800806.
41. Ryan JM, Lascelles BD, Benito J, et al. Histological and molecular
characterization of feline humeral condylar osteoarthritis. BMC Vet Res
2013;9:110.
42. Dickomeit MJ, Bottcher P, Hecht S, Liebich HG, Maierl J. Topo-
graphic and age-dependent distribution of subchondral bone density in the
elbow joints of clinically normal dogs. Am J Vet Res 2011;72:491499.
43. Engelke E, Koch R, Brunnberg L, Waibl H. The olecranon ligament
of the elbow joint in dogs and cats. Kleinteirpraxis 2005;50:313323.
44. Seelig U. Klinisch-pathologische Untersuchungen zum Ligamentum
olecrani bei Hunden mit Ellbogengelenksdysplasie. Dissertation, Freie Uni-
versitat Berlin, Germany. 2009.
45. Breit S, Kunzel W, Seiler S. Variation in the ossification process of
the anconeal and medial coronoid processes of the canine ulna. Res Vet Sci
2004;77:916.
46. De Rycke LM, Gielen IM, van Bree H, Simoens PJ. Computed
tomography of the elbow joint in clinically normal dogs. Am J Vet Res
2002;63:14001407.
47. Seeram E. Image quality. In: Seeram E, (ed.): Computed tomography:
physical principles, clinical applications, and quality control. Philadelphia:
WB Saunders Co., 1994:173197.