Arthrodesis is a well-established surgical procedure for the treatment of
arthritis of the ankle joint and has been accepted as yielding good
intermediate-term
results1-8.
Specific indications for arthrodesis have expanded over the years, but the
general principles remain the same: correction of deformity, alleviation of
pain, and maintenance of stability.
Current published reports on ankle fusion have consisted of retrospective
clinical reviews that have focused on indications, patient satisfaction,
surgical technique, the proper position of fusion, and the effects of fusion
on
gait1-40.
The majority of these reports have had a short to intermediate-term follow-up
period. Several reports with long-term outcomes have demonstrated that
patients have foot pain caused by the development of ipsilateral hindfoot
arthritis, particularly of the subtalar
joint4,5,10,11,13,28.
The deficiencies of previously published reports include the lack of current
validated outcome measures, poorly defined inclusion criteria, and the lack of
a control group for comparison.
The purpose of the current study was to combine three-dimensional kinematic
gait analysis with validated clinical outcome measures to evaluate patients
who had been managed with an ankle arthrodesis for the treatment of isolated
posttraumatic or primary ankle joint arthritis. These results were compared
with the findings for an age and gender-matched group of healthy controls to
provide the treating physician with a functional outcome study of ankle
arthrodesis in a homogeneous group of patients who had been managed with
modern surgical techniques and methods of outcome evaluation.
The present study was performed with the approval of the Research Ethics
Board of our institutions, and all participants signed an informed-consent
form prior to inclusion.
A retrospective review of patients who had undergone isolated ankle
arthrodesis between December 1993 and December 1999 was undertaken. Patients
were identified from the operative records of the senior author (T.R.D.). A
review of the medical records of these patients was then performed to identify
those who met the inclusion criteria for the study.
Only patients who had undergone arthrodesis for the treatment of conditions
isolated to the ankle joint (i.e., posttraumatic arthritis, primary
osteoarthritis, or osteochondral lesions) were included in the study. In
addition, only patients who had undergone arthrodesis with use of modern
internal fixation techniques and who had at least twelve months of follow-up
were included in the study.
Any patient who had abnormal function of, or trauma to, other areas of the
lower limbs or the contralateral ankle was excluded from the study. Other
exclusion criteria included another midfoot or hindfoot arthrodesis,
inflammatory arthritis, Charcot arthropathy, neuromuscular disease, and
congenital deformity. In addition, patients with cardiopulmonary symptoms that
were severe enough to limit their exercise tolerance during gait analysis were
excluded from the study.
Thirty-six suitable candidates were identified on the basis of a chart
review. Of these, thirty-three (92%) could be contacted and three had been
lost to follow-up. An additional two patients were excluded because of knee
osteoarthritis, one patient was undergoing talonavicular fusion, and four
patients declined to participate. The remaining twenty-six patients included
fourteen men and twelve women. The mean age at the time of surgery was
fifty-four years (range, eighteen to eighty-three years), and the mean
interval between surgery and assessment was forty-four months (range, twelve
to eighty-six months).
Posttraumatic arthritis was the preoperative diagnosis in all cases but
one. One patient presented with primary osteoarthritis.
At the time of surgery, access to the ankle was gained through a medial and
lateral approach. A transfibular osteotomy was performed in all cases.
Initially, the fibula was resected and used as local bone graft (nineteen
cases), but more recently the distal part of the fibula was retained. At the
end of the operation, a distal syndesmosis fusion was obtained by denuding the
medial third of the fibula of cartilage and then securing it to the tibia and
talus with compression screws (Figs.
1-A and
1-B). The medial malleolus was
partially resected to help to obtain compression and to prevent valgus tilt of
the talus. All fusion sites were stabilized with use of three cannulated
compression screws. Two 4.5-mm screws were inserted into the talar body, with
a 7.0-mm screw being placed from the posterolateral aspect of the tibia into
the talar neck. The optimum fusion position was plantigrade, with the heel in
as much as 5° of valgus and with external rotation equivalent to that on
the contralateral side. The talus was also displaced slightly posteriorly on
the tibia. Twelve patients underwent Achilles tendon lengthening as part of
the procedure. In one patient, a revision procedure was performed two weeks
after the initial procedure to reposition the ankle from equinus into a
plantigrade position.
Clinical Assessment
At the time of gait analysis, all patients underwent a clinical
examination. The position of fusion was recorded. Sagittal plane movement of
the foot was measured with use of a goniometer referencing the long axis of
the shank and the sole of the foot. The motion of the subtalar complex was
also assessed. Ranges of motion were then compared with those on the
contralateral side. Circumferential measurements were made from the widest
diameter of both calves.
Patients were questioned specifically about their symptoms, satisfaction
with surgery, gait pattern, and use of orthotics. On the basis of the answers
to the clinical questions and the findings of the clinical examination, the
American Orthopaedic Foot and Ankle Society (AOFAS) Ankle-Hindfoot
scale41 was
completed. This instrument is a nonvalidated questionnaire that is divided
into pain and function subsections. The maximum score is 100 points, and a
value of >80 points is considered to be a very good to excellent
result.
Validated Outcome Questionnaires
All twenty-six patients completed an Ankle Osteoarthritis Scale (AOS) and a
Musculoskeletal Outcomes Data Evaluation and Management Systems (MODEMS)
baseline foot and ankle questionnaire.
The AOS is a reliable and validated self-assessment scoring system that
specifically measures symptoms and disabilities related to arthritis of the
ankle42. The scale
was adapted from the Foot Function
Index43 and uses a
visual analog scoring system comprising eighteen questions divided into pain
and disability subgroups.
The MODEMS baseline foot and ankle questionnaire forms part of the American
Academy of Orthopaedic Surgeons/Council of Musculoskeletal Specialty Societies
outcomes data collection questionnaires. MODEMS is an assessment instrument
that aims to standardize functional outcome information on musculoskeletal
conditions44. The
questionnaire comprises both a general health survey (Short Form-36 [SF-36])
as well as specific questions related to the foot and ankle. The SF-36 is an
accepted patient-oriented questionnaire that has been used previously for the
evaluation of musculoskeletal
disease45. The part
of the questionnaire related to the foot and ankle has questions on pain,
stiffness, stability, activity, shoewear, and physical appearance. Questions
are answered on the basis of the average severity of symptoms experienced
during the past week. The results are presented as four disease-specific
subscales: physical health and pain, satisfaction with symptoms, shoe comfort,
and a global foot and ankle scale. As the patients had already had surgery,
the section on treatment expectations was omitted. Scores are rated on a scale
of 100 (with 100 corresponding to the best possible score), apart from
satisfaction, which is rated on a scale of 5.
Radiographic Examination
At the time of assessment of the study group, standard radiographs were
made as part of the normal follow-up in the outpatient clinic. These consisted
of weight-bearing anteroposterior and lateral radiographs of the ipsilateral
foot and ankle. Radiographs could not be made for one patient because of time
constraints. Radiographs made at the time of ankle arthrodesis were available
for twenty-three of the twenty-six patients.
Two authors (T.R.D., R.T.) reviewed the radiographs and assessed
osteoarthritic changes in the subtalar, talonavicular, and calcaneocuboid
joints. Scores were given according to the grading system of Kellgren and
Moore46, a system
that was recently used in a review of long-term results following ankle
arthrodesis13. With
this system, scores can range from 1 (normal) to 5 (severe arthritis). Both
authors scored the radiographs independently after having first gained
consensus on grading by simultaneously scoring radiographs of ten ankle
fusions that were not included in this study.
The tibiotalar angle (the angle subtended by the long axis of the tibia and
the axis of the talus) and the anterior translation of the head of the talus
in front of the anterior border of the tibia were measured on the
weight-bearing lateral radiographs of the ankle. Anterior translation was
compared with the findings on preoperative lateral radiographs. The tibiotalar
angle was compared with the preoperative measurement only if a weight-bearing
radiograph was available (seven patients).
Gait Analysis
Bilateral barefoot gait data were collected for the patient population.
Patients wore T-shirts and shorts during the examination to allow for the
accurate placement of markers. Three-dimensional kinematic data were obtained
with use of an adapted two-segment foot model to describe the motion of the
forefoot relative to the hindfoot and the hindfoot relative to the
shank40.
Thirty-five spherical reflective markers, 2.5 cm in diameter, were used to
define the rigid-body model (Fig.
2). Markers were placed on the heads of the first and fifth
metatarsals; on the measured midpoint between the bases of the first and fifth
metatarsals; and on the posterior, medial, and lateral aspects of the
calcaneus. Markers also were placed on the lateral ankle malleolus, the distal
part of the anterior aspect of the shank, the head of the fibula, the distal
part of the anterior aspect of the thigh, the lateral aspect of the thigh, the
greater trochanter, the anterior superior iliac spine, and the iliac crest to
define the shank, thigh, and pelvis segments. Four additional markers were
placed to define the body-embedded coordinate systems and joint centers for a
neutral standing static frame. These markers were located on the medial
malleolus, the lateral and medial condyles of the femur, and the sacrum. Local
dynamic and bone-embedded coordinate systems were defined for the forefoot
(metatarsals, cuneiforms, and cuboid), the hindfoot (calcaneus, talus, and
navicular), the distal part of the shank, the proximal part of the shank, the
thigh, and the pelvis. The motion of the distal segment relative to the
proximal segment was obtained with use of Euler angles. A rotation sequence
from the sagittal plane to the coronal plane and then to the transverse plane
was used to define the forefoot and hindfoot relative angles to allow for
comparisons with previously published
data40. The
remainder of segment rotations were defined with use of an existing
mathematical Euler sequence of coronal followed by transverse and then
sagittal plane
rotations47. A
six-camera Vicon system (Vicon Peak, Lake Forest, California) was used to
monitor the three-dimensional body-segment kinematics at a sampling rate of 60
Hz with cameras that were optimized to capture a volume of approximately 2 m
(length) × 1.5 m (height) × 1 m (width). Coordinate data were
processed with use of custom software (MATLAB; The MathWorks, Natick,
Massachusetts). Data were passed through a low-pass 6-Hz digital Butterworth
filter prior to Euler-angle calculations. Temporal measures of velocity and
percentage time spent in the stance and swing phases of gait were calculated
on the basis of the posterior heel marker trajectories.
Patients were instructed to walk at a comfortable self-selected speed down
the walkway. Five sets of data were obtained, time-normalized, and averaged to
obtain a single representative trial.
Control Group
The control group included thirteen healthy women and fourteen healthy men
with a mean age of fifty-two years (range, twenty-six to seventy-five years).
Controls had no systemic disease and no symptoms or surgery related to the
lower limbs. The control group was matched to the study group with respect to
age and gender; the patients weighed slightly more than the controls did
(Table I).
The control group underwent questioning as well as clinical examination of
the foot and ankle, and an AOFAS ankle-hindfoot score was completed. The
self-assessment Ankle Osteoarthritis Scale and the MODEMS Foot and Ankle
questionnaire were also completed.
Gait analysis was performed in a manner identical to that used for the
study group. Data were collected for each side of the body when time permitted
(eighteen controls) or for one side chosen randomly (nine controls). The side
used for final analysis in patients with bilateral gait data was chosen
randomly.
Statistical Analysis
Statistical comparisons were made between the control group (for one side
chosen randomly) and the study group (for the affected side) with use of a
series of multiple analysis of variance tests (SPSS, version 11.5; SPSS,
Chicago, Illinois). Three multiple analysis of variance tests were conducted
to separately analyze kinematic gait parameters, temporal gait parameters, and
clinical and questionnaire outcome parameters. The dependent kinematic gait
parameters included range of motion, during a stride of gait, in the sagittal
plane for the pelvis, hip, knee, hindfoot, and forefoot; in the coronal plane
for the hindfoot and forefoot; and in the transverse plane for the hindfoot
and forefoot. The dependent temporal gait parameters included cadence
(steps/minute), stride length (meters), and percentage of gait cycle spent in
the stance phase. The dependent clinical and outcome questionnaire parameters
included the AOS subgroup scores for pain and disability; the AOFAS total
score; and the MODEMS mental composite, physical composite, physical health
and pain, satisfaction with symptoms, shoe comfort, and global foot and ankle
scores. For all analyses, the level of significance was set at p <
0.05.
The level of agreement between the reviewers' radiographic scores was
statistically analyzed, and a Kappa value was calculated.
Clinical Assessment
The position of fusion was determined to be clinically satisfactory in all
patients. The hindfoot was in slight valgus alignment in all cases, and no
hindfoot was positioned in detectable dorsiflexion or plantar flexion.
All patients but one demonstrated muscle atrophy in the ipsilateral calf.
The mean difference between the involved limb and the contralateral limb was 3
cm (range, 1 to 7 cm).
Twenty patients (77%) were completely satisfied or were satisfied with only
minor reservations with the result of the arthrodesis. No patient was
completely dissatisfied with the result. All patients but one stated that they
would undergo the procedure again if faced with an identical scenario.
Twenty-five of the twenty-six patients stated that they would recommend an
ankle fusion to a friend given the same degree of preoperative symptoms that
they had experienced.
Five (19%) of the twenty-six patients stated that they did not notice a
gait abnormality. Twelve (46%) of the twenty-six patients wore custom-made
orthotics of various types, and all believed that use of the orthotics had
improved their gait.
Outcome Questionnaires
Multiple analysis of variance indicated that there was a significant
difference between the study group and the control group with regard to the
clinical and outcome questionnaire scores (multivariate F = 6.76, p <
0.0005). The results of univariate analysis are summarized in
Table II.
The mean postoperative AOFAS ankle-hindfoot score in the study group (74
points; range, 43 to 93 points) was significantly lower than that in the
control group (98 points; range, 88 to 100 points) (p < 0.0005).
The mean AOS scores for pain and disability were 20 (range, 0 to 72) and 20
(range, 0 to 73), respectively, in the study group, compared with 5 and 4,
respectively, in the control group (p < 0.134, p < 0.05).
The SF-36 physical composite and mental composite scores were not
significantly different between the study group and the control group.
The MODEMS is used to assess four subscales specific to foot and ankle
symptoms (Table III). No
significant difference was seen between the study group and the control group
with regard to satisfaction with foot and ankle symptoms, with the scores for
both groups (3.8 and 4.0, respectively) indicating that the respondents were
"somewhat satisfied." However, there was a significant difference
between the study group and the control group with regard to the global foot
and ankle scale (p < 0.01).
Radiographic Examination
Statistical analysis of osteoarthritis assessment by both surgeons revealed
a significant level of agreement between preoperative and postoperative
scores. Four patients (15%) progressed from no, doubtful, or minimal subtalar
joint arthritis preoperatively to moderate or severe arthritis at the time of
assessment. The calcaneocuboid joint was also affected in one of these
subjects, progressing from stage 2 to stage 5.
The mean tibiotalar angle as measured on a weight-bearing lateral
radiograph was 103° (range, 93° to 115°) postoperatively, compared
with 109° (range, 94° to 117°) preoperatively.
Gait Analysis
The results of gait analysis in the study group were statistically compared
with data obtained from the age and gender-matched control group. Multiple
analysis of variance indicated that there was a significant difference in
temporal measures between the study group and the control group (multivariate
F = 7.47, p < 0.0005). Multivariate and univariate F-test results for
temporal parameters are summarized in Table
IV. The range of motion discussed relates to the total dynamic
motion identified during steady-state walking and does not reflect the full
arc of passive range of motion of the joints stipulated.
With the numbers available, there was no significant difference between the
study group and the control group with regard to the percentage of time spent
in stance. However, the study group had significant decreases in both cadence
and stride length as compared with the control group (p < 0.01). Gait
velocity was 1.06 ± 0.22 m/sec in the study group, compared with 1.29
± 0.17 m/sec in the control group. Multiple analysis of variance of the
kinematic parameters indicated that there was a significant difference between
the groups with regard to range of motion measures during a stride of gait
(multivariate F = 12.97, p < 0.0005).
The kinematic parameters are summarized in
Table V. There were no
significant differences between the study group and the control group with
regard to the range of motion of the pelvis or the knee joint. In the sagittal
plane, the study group had significant decreases in range of motion at the hip
(p < 0.01), hindfoot (p < 0.0005), and forefoot (p < 0.0005) as
compared with the control group. Similarly, in the coronal and transverse
frontal planes, the study group had significantly decreased range of motion in
the hindfoot and forefoot.
The patients in the present study were a homogeneous group in that they had
isolated ankle arthritis that was treated surgically with ankle fusion.
Previous studies reviewing the results of ankle arthrodesis have included
patients with inflammatory arthropathy, ipsilateral foot arthritis, congenital
foot disease, and associated lower limb
pathology3-5,7,9-11,13,16,19,22,23,26-28,32,34,35,40,48,49.
None of those studies compared the clinical results of ankle arthrodesis with
the findings in a control group, and functional outcome questionnaires have
only been recently
developed13,50.
The primary purpose of the present study was to compare the intermediate-term
functional results for a group of patients who had been managed with ankle
arthrodesis with the findings for a group of normal age and gender-matched
controls with use of modern evaluation techniques.
Clinical and Radiographic Evaluation
Previous clinical reviews of ankle arthrodesis with an average duration of
follow-up of less than seven years can be divided into five broad categories
with an emphasis on clinical
outcomes9,16,23,51-53,
complications17,54,55,
surgical
techniques6,23,24,30,31,56-61,
gait
analysis4,22,27,40,62,
and patients with rheumatoid
arthritis15,63,64.
Previous work has established that union rates are improved with internal
fixation with use of compression screws and that gait parameters are more
normalized when the arthrodesis is performed with the foot in neutral
dorsiflexion and slight hindfoot valgus. The current study adhered to these
recommendations, with a 100% union rate and with all ankles being fused in an
acceptable position. Those same previous studies also indicated a high rate of
patient satisfaction, with acceptable pain relief. The results of the clinical
evaluation were similar in that patient satisfaction was high, pain relief was
good, and most patients stated that they would undergo the surgery again,
given a choice. The evaluation of patient outcomes in the studies mentioned
above did not go beyond these parameters, particularly with respect to further
assessment with use of pain analog scales or validated functional outcome
scores.
Outcome Questionnaires
Our outcome questionnaires included nonvalidated scores such as the
American Orthopaedic Foot and Ankle Society (AOFAS) hindfoot scale. A recent
meta-analysis demonstrated the AOFAS questionnaire to be the most commonly
used scoring system for the
foot65; it is
considered to be a regional scale that is designed to assess function and
disability of only the hindfoot. This scale has been rated as having poor
reliability, validity, and responsiveness. In the present study, the AOFAS
hindfoot score demonstrated a significant difference between the study and
control groups.
The validated scores that were used in the present study included the Ankle
Osteoarthritis Scale (AOS) and the Musculoskeletal Outcomes Data Evaluation
and Management Systems (MODEMS) baseline foot and ankle questionnaire. The AOS
questionnaire has been adapted from the Foot Function Index and modified to be
more disease-specific for ankle
arthritis42. The
MODEMS is both a global and a regional questionnaire that includes the Short
Form-36 (SF-36) and disease-specific sections for the foot and ankle. It has
been slightly modified and now exists as part of the American Academy of
Orthopaedic Surgeons lower limb outcomes assessment instruments that have been
validated66. Both
of these questionnaires are considered to be reliable, with good criterion and
construct validity, but weak with regard to
responsiveness65.
In the present study, the scales that were specific for foot and ankle
function demonstrated a significant difference between the study group and the
control group whereas the scales that were specific for foot and ankle pain
did not. The AOS disability section, the AOFAS hindfoot scale, and the MODEMS
global foot and ankle scale demonstrated significant differences between the
study group and the control group (Table
II and
III). The AOS pain scale and
the MODEMS physical health and pain sections did not demonstrate a significant
difference between the study group and the control group. These results
suggest that arthrodesis of the ankle substantially decreases the pain caused
by ankle arthritis but creates disability with regard to foot and ankle
function.
The more global scales (such as the SF-36 physical composite and mental
composite scores) and satisfaction with symptoms did not demonstrate
significant differences between the two groups. These results suggest that the
limitations identified in the disease-specific scores are not substantial
enough to significantly affect the general health and physical function of the
individual.
Gait Analysis
Gait analysis revealed that the patients in the present study had
significantly decreased sagittal, coronal, and transverse ranges of motion of
the hindfoot and forefoot during the stance and swing phases of gait. A
simplified ankle-foot model was utilized to explore relative foot segment
motions. The joint motions of the foot from a clinical perspective are more
complex and may not be fully reflected in the results obtained.
Clinical series that have focused on the sagittal (dorsiflexion-plantar
flexion) arc of motion of the midtarsal (Chopart) joints have shown variable
results: some have demonstrated increased
motion3-5,11,40,67,
and some have demonstrated decreased motion or no
change13,25,28,32,35,50.
In the current study, a decreased cadence and stride length were observed
along with decreased flexion at the hip. Rotation at the pelvis and motion at
the knee were not significantly altered.
The current data were also compared, where applicable, to kinematic and
temporal results as reported in previous studies of gait following ankle
arthrodesis in ten
subjects40 and in
nine subjects68.
The study by Beyaert et
al.68 and the
current study identified similar values and trends for temporal measures of
cadence and the percentage of time spent in stance phase. Both studies
indicated little difference in the time spent in stance phase, with patients
having a reduced cadence as compared with controls. Wu et
al.40 reported
slightly different results, indicating increased cadence in patients as
compared with controls and significantly increased time spent in swing phase
as compared with stance phase for patients as compared with controls. The
current study could be compared only with the study by Wu et
al.40 in terms of
the hindfoot and forefoot. Both studies demonstrated a significant decrease in
hindfoot dorsiflexion range of motion in patients as compared with controls.
This is where the similarities ended, with reverse trends obtained for
forefoot range of motion measures and hindfoot range of motion measures in the
coronal and frontal planes. The trends identified by Wu et
al.40 indicated
increased forefoot range of motion in patients as compared with controls,
whereas we found the opposite trend.
The inclusion criteria in these three studies ensured that all patients had
the arthrodesis with the ankle in the recommended position of neutral
dorsiflexion, slight external rotation, and valgus. As with the outcome
questionnaires, gait analysis suggested that the maximum disruption of the
gait parameters were observed at the hindfoot and midfoot, with the
abnormalities in gait kinematics becoming less noticeable as the analysis
moved further away from the zone of pathology. The influence of ankle fusion
on midfoot and forefoot range of motion among studies involving gait analysis
remains variable.
Radiographic Analysis
As early as 1978, it was recognized that subtalar stiffness and arthritis
can develop following ankle
fusion67. Since
that time, a number of studies with both intermediate and long-term follow-up
have indicated that arthrofibrosis and arthritis of the subtalar joint can be
a source of disability following ankle
fusion10,11,13,16,23,28,48,50,52.
Only one study, with intermediate-term follow-up, assessed the progression of
subtalar arthritis, with progression being identified in 31% (eight) of
twenty-six
patients9. The
current study identified progression of subtalar arthritis in 15% (four) of
twenty-six patients. Beyaert et
al.68 performed
gait analysis on nine patients with minimal comorbidities who had had an
arthrodesis with the ankle in neutral dorsiflexion. Their findings indicated
that the restricted hindfoot motion resulted in decreased forward progression
of the tibia, causing early heel-rise and increased ground-reaction forces to
the midtarsal joints. That study suggests that these abnormal mechanics can
contribute to the development of radiographic evidence of subtalar and
midtarsal arthritis in the intermediate follow-up period, even when modern
surgical techniques and accepted positions of fusion are employed.
Overview
Since trauma is the primary etiology of ankle arthritis, it is more likely
to occur in young, active
individuals69.
Patients undergoing ankle arthrodesis for the treatment of end-stage arthritis
are, on the average, forty-six years of age (range, twenty-nine to sixty-three
years)3,5,6,11,13,16,22-24,28,30,31,40,48-51,56-62.
Consequently, the goal of treatment is not only pain relief and restoration of
function but good long-term results. The current study suggests that, in the
intermediate phase, ankle fusion provides acceptable results in terms of pain
relief, general health, and physical function. Closer analysis focusing on
gait and the hindfoot indicated that there were substantial differences
between the group of patients managed with ankle fusion and the cohort of
healthy controls. These data suggest that the detrimental effects of ankle
fusion first occur in the joints immediately affected by the restricted ankle
motion; this idea is supported by the fact that subtalar arthritis had evolved
in 15% of the patients in the current study and in >60% of patients
reviewed in long-term
studies13,50.
Given the results of the present study and recent long-term reviews, patients
should be advised that, although ankle fusion will help to relieve pain and to
improve overall function, it is considered to be a salvage procedure that
causes persistent alterations in gait, with the potential for deterioration
due to the development of ipsilateral hindfoot arthritis.