Abstract
Background: Patients who have had a hip
arthrodesis have been reported to have pain in the back and the
knee due to an altered gait. There is little information about the
specific compensatory mechanisms that are adopted when walking.
The purpose of this study was to objectively define gait adaptations
after an arthrodesis of the hip and to correlate the kinematic findings
with pain and other patient outcomes.
Methods: Nine patients who had had an arthrodesis
of the hip at an average age of thirteen years and five months (range,
ten years and nine months to sixteen years and eleven months) were
evaluated with gait analysis and muscle strength-testing and completed
a questionnaire related to pain and function. The average duration
of follow-up was eight years and ten months (range, two years and
one month to thirteen years and ten months). The frequency of the
postoperative visits varied. Seven patients were adults at the time
of the study and were called back specifically for the study.
Results: All patients had decreased cadence
and step lengths. The kinematic findings included decreased dorsiflexion
of the ipsilateral ankle, hyperextension of the ipsilateral knee
during the stance phase, and a tendency toward increased genu valgum
during gait. In normal gait, there is no genu varum or valgum during
the stance phase.
The patients had an average (and standard deviation) of 7 ± 4
degrees of genu valgum. Pelvic and lumbar motion in the sagittal
plane was excessive in all patients. Strength-testing revealed clinically relevant
weakness in the ipsilateral quadriceps in all patients, with a difference
of more than 20 percent between the two extremities in six patients. The
gastrocnemius-soleus muscle was stronger on the side with the fused
hip in six patients.
The questionnaire, designed by Harris in 1969 and completed by
the patients at the time of the gait analysis, revealed back pain
in seven patients. The questionnaire was administered only once.
The functional outcome as measured with use of the same questionnaire
worsened as the duration of follow-up increased.
Conclusions: The gait analysis showed excessive
motion in the lumbar spine and the ipsilateral knee in all nine
patients. This abnormal motion led to pain as the duration of follow-up
increased, and all patients who had been followed for four or more
years after the arthrodesis complained of back pain. We hypothesized
that excessive motion for an extended duration can lead to back
pain. The preferred position of the hip for the arthrodesis was
20 to 25 degrees of flexion, neutral abduction-adduction, and neutral
rotation.
Hip arthrodesis remains the most viable operative treatment
for the active adolescent who has severe degenerative disease of
the hip. Although total hip arthroplasty is performed in patients
who have juvenile rheumatoid arthritis, it is not currently advocated
for children who have unilateral joint destruction and are otherwise
normal. The goal of the arthrodesis is to alleviate pain in the
hip by eliminating motion, but this has been reported to produce
abnormal pressures on the lumbar spine and the ipsilateral knee2,3,12. Benaroch et al.2 performed a preliminary study assessing
cadence parameters and function in thirteen patients who had undergone
an arthrodesis of the hip and found that single-limb support was
of shorter duration on the affected side.
The purpose of the current study was to further analyze the gait
of patients who had had an arthrodesis of the hip, with particular
attention paid to motion at the knee and the lumbar spine, and to correlate
the gait-analysis findings with the radiographic findings and the
clinical symptoms.
We reviewed the medical records of nineteen patients who had
had an arthrodesis of the hip between January 1980 and December
1996 and had had a minimum of two years of follow-up. Nine patients
agreed to participate in the study. (Six patients could not be located,
two had had a hip replacement, one had osteogenesis imperfecta,
and one was seriously ill.) The study was approved by the institutional
review board, and all patients or their legal guardians signed informed-consent forms.
The operative technique was identical in all patients; it consisted
of an intra-articular arthrodesis with the hip in the position of
maximal contact between the femoral head and the acetabulum, a subtrochanteric
osteotomy, and immobilization in a spica cast as described by Mowery
et al.8 (Fig. 1-A and Fig. 1-B). The position
of the hip was verified with direct visualization fluoroscopy.
There were six male patients and three female patients in the
study group (Table I).
The average age at the time of the operation was thirteen years
and five months (range, ten years and nine months to sixteen years
and eleven months), and the average age at the time of follow-up
was twenty-two years and four months (range, fourteen years and
six months to twenty-seven years and eleven months). The average
duration of follow-up was eight years and ten months (range, two
years and one month to thirteen years and ten months).
The preoperative diagnoses included slipped capital femoral epiphysis
with avascular necrosis in four patients; idiopathic chondrolysis
in two patients; and avascular necrosis following a femoral neck
fracture, Legg-Calvç¬erthes disease, and a septic hip dislocation
in one patient each.
Three patients sustained a total of four fractures (a fracture
through the fusion site, an ipsilateral intertrochanteric femoral
fracture, and ipsilateral distal femoral and ankle fractures) after
the arthrodesis. The distal femoral fracture was treated with a
cast for four months; the fracture through the fusion site was treated
with Ilizarov external fixation for six months; and the ipsilateral
intertrochanteric femoral fracture, sustained after a motor-vehicle accident
fifteen months postoperatively, was treated with open reduction
and internal fixation followed by Ilizarov external fixation because
of limb-length discrepancy. In addition, one patient was treated
with Ilizarov fixation because of limb-length discrepancy.
A physical examination was performed to assess the position of
the fused hip and to measure the range of motion of the knee. Since
this examination was done postoperatively, we could assess only
the position in which the hip had been fused and could not determine
the best position for the arthrodesis. Limb-length discrepancy was
quantified with use of blocks to level the posterior iliac spine
with the patient in the standing position. Scanograms were not obtained;
thus, any contribution of pelvic asymmetry to the limb-length discrepancy
was not excluded. The circumference of the thigh was measured bilaterally
at a point ten centimeters proximal to the level of the superior pole
of the patella. Knee stability was evaluated clinically and was
described either as present or as decreased with or without an end
point in the medial, lateral, and anteroposterior directions.
An anteroposterior radiograph of the pelvis was made on the day
of testing. Because the purpose of the current study was to determine
the final result and not to follow the patients sequentially over time,
radiographs were made at various times after the operation but not
as part of the study. All patients had instrumented gait analysis
and isokinetic muscle strength-testing of the knee and the ankle.
Both extremities were tested only at the time of the latest follow-up
visit.
Kinematic data were gathered as each patient walked at a self-selected
speed along a fifteen-meter runway; the volume of the space for
calibration was 2.5 meters. Each walk was videotaped simultaneously
in the frontal and sagittal planes. Thirty-eight reflective markers
were placed on the body to track motion during gait of twelve linked segments:
the head, the arms, the thoracic spine, the lumbar spine, the pelvis,
the thighs, the legs, and the feet. Three-dimensional kinematic
data were recorded with a six-camera sixty-hertz VICON system (Oxford
Metrics, Oxford, England). Data were recorded for each extremity
of each patient for four representative cycles. Lumbar, hip, knee,
and ankle-joint angles were calculated in the sagittal, coronal,
and transverse planes with use of BodyBuilder software (Oxford Metrics).
These angles included lumbar flexion and extension, right and left
lateral tilt, and axial rotation; pelvic anterior and posterior
tilt, right and left obliquity, and axial rotation; hip flexion
and extension, adduction and abduction, and internal and external rotation;
knee flexion and extension, varus and valgus, and internal and external
rotation; ankle plantar flexion and dorsiflexion and internal and external
rotation; and foot progression angle. The three-dimensional relationships
between the movements of the pelvis and the lumbar spine during
walking were evaluated on the basis of the model described by Whittle
and Levine17. Hip, knee, and ankle-joint
centers were located with use of the methods described by Davis
et al.5. Cadence, velocity, step
length, and the percent of the gait cycle spent in single and double-limb
support were calculated. Maximum and minimum values for joint angles
were identified with use of custom software written in the movement-science laboratory
at the Texas Scottish Rite Hospital for Children. A computer was
used to determine maximum and minimum values in order to improve precision
and eliminate human error.
Five normal subjects who had never had an orthopaedic procedure
had gait analysis with use of the thirty-eight-marker model, and
the data were analyzed with the BodyBuilder software. These data then
were compared with those in reports on lumbar spine motion in the
literature4,17 in order to verify
our model. We did not perform similar comparisons for motion of
the pelvis, hips, knees, and ankles because that data has been well documented10. We used standard data that are
employed by gait-analysis laboratories, and we were confident that the
data on our patients with hip fusion could be compared reliably
with the aforementioned normal data. In contrast, there is much
less data on lumbar spine motion; three-dimensional motion, calculated
with use of photogrammetry, has been reported in only two studies3,17, to our knowledge. We chose to
duplicate the methods of Whittle and Levine17 because
those authors indicated that markers attached to lightweight wands
and not directly to the skin allowed for more accurate and reliable
data. Our data on the normal subjects showed ranges and patterns
of motion similar to those reported by Whittle and Levine.
Isokinetic muscle strength was measured with a Cybex-II machine
(Ronkonkoma, New York); the knee flexors and extensors were tested
with the patient in the seated position, which required allowing
the patient to recline slightly to accommodate the fused hip. The
ankle plantar flexors and dorsiflexors were tested with the patient
in the prone position, with allowances made for the flexed position
of the hip. The quadriceps and hamstrings were tested at 60 degrees
per second, and the anterior tibialis and gastrocnemius-soleus were tested
at 30 degrees per second. The peak torque per body weight was selected
from one of five trials. All testing was performed by an experienced exercise
physiologist. Asymmetry of more than 20 percent between the two
sides was considered clinically relevant13.
All nine patients completed a modified Harris hip-score questionnaire7. The Harris hip score, as described
by Harris in 1969, allots a maximum of 100 points, 91 of which are
related to pain and function and 9, to range of motion and the presence
of deformity. We omitted the areas of the score pertaining to motion
and deformity and did not ask our patients about their use of public
transportation. This left a possible maximum score of 90 points,
similar to that used in previous studies of patients who had had
a hip arthrodesis2,9. We used
the Harris hip score because it had been used by Benaroch et al.2 and we wanted to compare our data
with theirs.
Anteroposterior radiographs were reviewed to assess the healing
at the site of the fusion. Adduction and abduction of the extremity
were measured as the angle between the perpendicular to the horizontal
line connecting the inferior parts of the teardrops of both acetabuli
and a line drawn down the anatomical shaft of the femur beginning
at the site of the subtrochanteric osteotomy. The perpendicular
line is drawn from the junction of the line along the shaft and
the horizontal line. We selected the inferior ends of the teardrops
because they were seen on all of the radiographs and because the
triradiate cartilage was closed as these patients were adults or
older teenagers. The top of the iliac crest was not used as a marker
because it could have been distorted by harvest of the bone graft.
Statistical analysis of the data was performed with t tests to
determine the difference in the range of motion between the control
subjects and the patients who had had an arthrodesis. Pearson correlation
was used to determine the relationship between the position of the
fused hip and the dependent variables, which included the range
of motion and the maximum and minimum values for joint angles of
the lumbar spine, pelvis, hip, knee, and ankle in the sagittal,
coronal, and transverse planes. Pearson correlation also was used
to determine if there were any relationships between the clinical
data, including pain, and the motion data or the duration of follow-up.
At the time of the examination in the gait laboratory, the position
of the fused hip in the nine patients was an average (and standard
deviation) of 26 ± 10 degrees (range, 10 to 45 degrees) of flexion,
1 ± 6 degrees of abduction (range, 7 degrees of abduction to 10 degrees
of adduction), and 6 ± 7 degrees of internal rotation (range, 18
degrees of internal rotation to 0 degrees of external rotation).
The additional procedures on the hip (Cases 2, 3, and 5) and the
fractures sustained after the arthrodesis (Cases 2, 3, and 6) may
have influenced the ultimate position of the fusion.
Limb-length discrepancy averaged 2.7 centimeters (range, 1.0
to 5.0 centimeters). The extremity on the side with the fused hip
was shorter in all patients. All patients had muscle atrophy in
the thigh, which ranged from 0.5 to 4.0 centimeters (average, 2.6
centimeters). Knee laxity was present in eight patients. All eight
had laxity in the mediolateral plane, and seven also had anteroposterior
laxity. However, all eight had good end points to stress. Passive
range of motion of the ipsilateral knee was notable for hyperextension
in all patients.
Force-plate evaluation, performed with the patient standing with
both feet on the ground, revealed that six patients put more than
50 percent of their weight on the extremity on the side with the
fused hip. As the limb on that side was shorter in all patients,
this may have simply been an attempt to maintain a plantigrade position
of the feet. Nonetheless, the patients did not avoid weight-bearing on
the side with the fused hip.
Cadence parameters were identical to those of the patients in
the studies by Benaroch et al.2 and
Waters et al.16. Five of the patients
in the current study had been included in the study by Benaroch
et al. at the time of an earlier follow-up. The cadence of the patients who
had a fused hip averaged 110 steps per minute compared with 116
steps per minute for the normal subjects; this difference was significant
(p < 0.05). The walking velocity of the patients who had a fused
hip averaged 1.1 0.1 meters per second compared with 1.3 meters
per second for the normal subjects; this difference was also significant (p < 0.05).
Step lengths, which averaged 0.6 meter, were reduced symmetrically
in the group with a fused hip compared with that (0.66 meter) in
the normal group. The patients spent slightly less time in single-limb
stance on the fused side than on the contralateral side (35 and
39 percent of the gait cycle, respectively).
Kinematic findings included decreased dorsiflexion of the ipsilateral
ankle during the stance phase in all nine patients; only one patient
had an ankle that was in frank equinus during stance (Fig. 2). Knee flexion-extension
curves in the sagittal plane showed hyperextension on the ipsilateral side
during the stance phase in eight patients. By the time of toe-off,
all patients had knee flexion that was greater than normal on the
fused side (Fig. 2).
The knee has a greater dynamic range of motion during gait to accommodate
for the lack of hip motion in the forward movement of the body over
the stance-phase foot and also to assist in clearance of the foot
since the hip cannot flex and pull the foot off the ground as swing
is initiated. The dynamic range of motion correlated with the increased
passive range of motion and passive hyperextension seen in all nine
patients (r = 0.78, p < 0.05). Additionally, knee hyperextension
tended to increase with increasing flexion of the fused hip (r =
0.76, p < 0.05). The eight patients with medial and lateral
joint laxity had excessive genu valgum during gait. Normally, the
knee has minimal rotation during gait. The patients had an average
of 7± 4 degrees of genu valgum. All nine patients exhibited excessive
rotation through the ipsilateral knee during gait (p < 0.05).
The average pelvic and lumbar ranges of motion for the five normal
subjects were very similar to the values reported by Crosbie et
al.4 and by Whittle and Levine17. In the sagittal plane, the pelvis
tilts posteriorly at the same time as the hip is in maximal flexion,
and it tilts anteriorly during single-limb stance. In the coronal
plane, the pelvis normally drops slightly during the swing phase.
In the transverse plane, the pelvis rotates internally slightly
as the swinging limb advances and it rotates externally in the stance
phase.
The average pelvic motion in the sagittal plane in the nine patients
was 12.8 ± 2 degrees (Fig. 3), which was greater than the normal
value of 3.5 ± 1.5 degrees over one gait cycle; this difference was
significant (p < 0.05). Hips that were fused in more flexion
were associated with more anterior pelvic tilt (r = 0.81, p < 0.05).
(One hip was fused in 45 degrees of flexion; one, in 35 degrees;
one, in 30 degrees; two, in 25 degrees; three, in 20 degrees; and
one, in 10 degrees. No hip was fused in extension.) Normal anterior
pelvic tilt is approximately 10 degrees, whereas the patients had
between 18 and 42 degrees of anterior pelvic tilt. No patient had
an increase in pelvic motion in the coronal plane, but six had greater-than-normal
pelvic rotation in the transverse plane. The normal range of pelvic
rotation in the transverse plane during gait is 10.1 ± 2.5 degrees.
The average range of pelvic rotation in the transverse plane was
13.0 ± 2 degrees (range, 9 to 19 degrees) in the patients (Fig. 4). In the current
study, one standard deviation of normal was considered to be beyond
normal limits.
The position of the lumbar spine is quite variable, but motion
in the sagittal plane is minimal in normal subjects, as was noted
both in the current study and in published series4,17.
Lumbar motion averaged 7.5± 2 degrees in the nine patients. All
patients had greater-than-normal lumbar motion (range, 5.7 to 9.8
degrees) in the sagittal plane (Fig. 3). Lumbar lordosis tended to increase
with increased flexion of the fused hip (r = 0.74, p < 0.05).
We determined the degree of kinematic lumbar lordosis on the
basis of the motion data and chose the average of the lumbar flexion-extension
graph as the degree of lordosis. Lumbar tilt in the coronal plane
was not increased in the patients, whereas lumbar rotation was increased,
averaging 12 ± 3 degrees compared with 10 degrees in the normal subjects
(Fig. 4).
This 2-degree difference was not significant; however, the increase
in lumbar spinal motion in the sagittal and transverse planes is
clinically relevant since we believe that it explains why patients have
back pain after an arthrodesis of the hip. If the hip cannot move
during gait, the spine compensates with abnormally increased motion,
which, over time, leads to pain. Six of the nine patients had increased
lumbar rotation. Patients in whom the hip was fused in more flexion
did not exhibit more pelvic and lumbar rotation when advancing the
limb compared with patients in whom the hip was fused in more relative
extension (r = 0.78, p < 0.05).
Isokinetic muscle-strength testing showed that six of the nine
patients had stronger plantar flexors in the ipsilateral ankle (p < 0.05);
however, the ipsilateral quadriceps was weaker in all nine, with
an average difference between the involved and uninvolved sides
of 18 percent (range, 3 to 40 percent). Six patients had asymmetry
of more than 20 percent in the isokinetic strength of the quadriceps. Hamstring
weakness with a difference between sides of more than 20 percent
was present in five patients. The gastrocnemius-soleus muscle was stronger
on the side with the fused hip in six patients. Comparisons were
not made with the extremities of the normal subjects as strength
is extremely variable both among individuals and in relation to
gender and age.
Seven patients complained of back pain, which averaged 6.1 points
on a scale of 0 to 10 points. The same seven patients also noted
pain in the ipsilateral knee, which averaged 3.1 points on a scale
of 0 to 10 points. Two patients rated the pain in the ipsilateral
knee as 1 point; one, as 2 points; one, as 3 points; and three,
as 5 points. The two patients who did not have pain in the back
or the knee were the youngest patients in the series; they were
less than fifteen years old at the time of the latest follow-up,
and they also had the shortest duration of follow-up.
According to the Harris scale, two patients had no hip pain,
three had slight discomfort, one had mild pain, and three had moderate
pain. One patient had pain in the contralateral hip, and two had
pain in the contralateral knee. All patients could walk at least
one mile (1.6 kilometers), although one often used a cane. Only
one adult patient was unemployed, although three patients thought
that the fused hip limited their employment choices.
The modified Harris hip score averaged 69.5 points (range, 47
to 90 points) of a maximum of 90 points. We used the rating scale
of Benaroch et al.2 to classify
the results of the questionnaire. According to this scale, 82 to
90 points indicated an excellent result; 73 to 81 points, a good
result; 64 to 72 points, a fair result; and less than 64 points, a
poor result. One patient had an excellent result, four had a good
result, one had a fair result, and three had a poor result. There
was a trend toward worsening of the modified Harris hip score with increasing
durations of follow-up (r = 0.76).
Pearson correlation coefficients demonstrated a relationship
between abduction of the fused hip and knee instability (r = 0.76,
p < 0.05). There was a good correlation between the severity
of back pain and the duration of follow-up after the arthrodesis
(r = 0.75, p < 0.05) (Fig. 5).
We found that back pain was present in seven of nine patients
who had a fused hip. The two patients who reported no back pain
were both fourteen years old at the time of the study and had only
two and three years of follow-up. The onset of the back pain was
much earlier than has been reported in most series3,12, including an earlier series from
this institution2. Callaghan et
al.3 evaluated twenty-eight patients
at an average of thirty-five years after an arthrodesis of the hip.
The average age at the time of the arthrodesis was twenty-five years
(range, ten to fifty-eight years). Sixty percent of their patients
had back pain within twenty-five years after the arthrodesis. Those
authors used their own questionnaire to evaluate the patients but
did not include it in their paper. Roberts and Fetto12 evaluated ten patients at an average
of eight and a half years (range, one year and six months to forty-four
years) after a hip arthrodesis. The average age at the time of the
operation was thirty-four years (range, twenty-three years and six
months to seventy-eight years). Back pain was present in four of their
patients, and pain in the ipsilateral knee was noted in two.
Sponseller et al.15 studied
fifty-three patients who were thirty-five years old or less at the
time of an arthrodesis. At an average of thirty-eight years (range,
twenty to fifty-four years), 57 percent had some low-back pain. Those
authors devised their own scale to evaluate the patients. A careful
review of their study shows that only 16 percent of their patients
had no back pain and an additional 27 percent had rare episodes of
pain.
Benaroch et al.2 studied thirteen
patients who had an average age of twenty-two years (range, nineteen
to twenty-seven years) at the time of the study and fifteen years
at the time of the arthrodesis. Ten patients had occasional back
pain, and seven had pain in the ipsilateral knee. Barnhardt and
Stiehl1 reported back pain in
all six of their patients at an average of twelve years (range,
five to twenty-eight years).
The correlation between the duration of follow-up and back pain
in the current study suggests that pain develops earlier than has
traditionally been thought. We noted an earlier onset of back pain
because our patients were younger and more active and therefore
were more likely to have symptoms in association with activities.
We did not use the same scales as those employed by other authors with
the exception of Benaroch et al.2.
We asked our patients to quantify their pain numerically because
it is possible that a patient who is able to participate fully in
activities and who has only occasional pain would not report pain
otherwise. We found no correlation between the degree of flexion
or adduction of the fused hip and the presence of back pain. Others
have found that hip fusion in between 25 and 30 degrees of flexion
and in neutral to slight adduction led to fewer symptoms related
to the lumbar spine6,11. Price
and Lovell11 studied fourteen
children at an average of five years (range, one to ten years) after
a hip arthrodesis and reported no back pain in any child and knee
pain in only one child. Only four patients were followed for more
than five years. Those authors recommended an optimum position of
30 degrees of flexion, neutral adduction, and neutral rotation;
however, they did not attempt to correlate the subjective result
with the position of the fused hip. In the current study, no patient
had pain during the first seven years after the operation.
On the basis of our data, it is possible to provide a biomechanical
explanation for the increasing prevalence of back pain. We documented
abnormally increased motion of the pelvis, which correlated with
increased motion of the lumbar spine. Motions in the sagittal and
transverse planes were the most abnormal. It follows that increased
flexion and extension of the pelvis and, therefore, of the lumbar
spine is necessary to advance the body forward in gait, replacing
the role of hip motion. Increased pelvic and lumbar rotation are
also recruited to help advance the extremity. The increased lumbar
and pelvic motion comes at a price, with low-back pain as the result.
Seven of our patients complained of pain in the ipsilateral knee,
and we documented increased laxity and hyperextension of the knee
in all nine patients (Fig. 6). Kinematically, the pattern of
knee motion was abnormal, with an increased dynamic range of motion
in the sagittal plane. Since the iliopsoas is ineffective in pulling
the foot on the side with the fused hip off the ground in the early
swing phase, the knee must hyperflex to initiate swing and then to
allow for foot clearance during the swing phase. Although abduction
of the fused hip was linked to increasing knee instability, a series
of nine patients is insufficient for us to precisely define the
optimal position for hip fusion so as to protect the knee. Clearly,
increased abduction or adduction will result in abnormal varus and
valgus moments across the knee in the stance phase and will place
it at risk for instability or degenerative arthritis due to asymmetrical
loads6. Gore et al.6 stated that more than 10 degrees
of adduction should be avoided, since hips that were fused in more
adduction caused the patient to walk slowly and with greater irregularity
in forward progression. Sponseller et al.15 found
that 10 degrees of adduction of the fused hip led to greater valgus
angulation at the knee and that less than 10 degrees led to genu
varum.
Muscular support of the ipsilateral knee is compromised by weakness
of the quadriceps and the hamstrings. Despite weakness of the knee,
the plantar flexors of the ipsilateral ankle were stronger on isokinetic
testing in the present study. As the iliopsoas and the superficial
hip flexors are nonfunctional in a patient with a fused hip, the
gastrocnemius-soleus must push the extremity off the ground and
into the swing phase. Neutral rotation of the extremity is important
in order to place the ankle and gastrocnemius-soleus in optimal
alignment for push-off.
We did not collect metabolic data on oxygen consumption or oxygen
cost in this group of patients. Waters et al.16 measured
oxygen consumption in an older group of eleven adult patients who
had had a hip arthrodesis at an average age of forty-six years (range, twenty-three
to fifty-seven years) and found a 32 percent increase in oxygen
consumption compared with that in normal adults. The average interval between
the operation and the testing was 7.5 years (range, 1.5 to thirty
years). The average rate of energy expenditure was 14.9 milliliters
per kilogram of body weight per minute compared with 12.1 milliliters
per kilogram of body weight per minute for normal individuals. The
heart rate of the patients who had had a hip arthrodesis averaged
112 beats per minute compared with ninety-nine beats per minute
for the controls. The oxygen cost averaged 0.223 milliliter per
kilogram per meter for the group that had had a hip arthrodesis compared
with 0.166 milliliter per kilogram per meter for patients who had
had an ankle arthrodesis. Waters et al. concluded that patients
who had had a hip arthrodesis were 53 percent as efficient during
gait as the normal controls.
Limb-length discrepancy remains an issue for patients who have
a fused hip. Benaroch et al.2 correlated
back pain, quadriceps weakness, and cadence abnormalities with limb-length
inequality. If epiphyseodesis of the distal aspect of the femur
of the unaffected extremity is possible on the basis of skeletal
age, then this procedure can be performed to decrease or correct
the discrepancy. Benaroch et al. recommended equalization for patients
who had a limb-length discrepancy of more than two centimeters.
Gore et al.6 found that limb-length
discrepancy was associated with a decrease in cadence and step length
and emphasized the importance of minimizing shortening on the side
of the fusion. Price and Lovell11 did
not make any specific recommendations, but several of their patients
used a shoe-lift and one patient had an epiphyseodesis. All of our
patients had less ankle dorsiflexion on the side with the hip fusion
in the stance phase, which is a strategy used to reduce the rise
and fall of the center of mass by children with limb-length inequality14. Attention should be given to equalization
of limb lengths when feasible in patients with a fused hip. The
utility of prescribing a shoe-lift was not evaluated in the current
study.
In general, we found that patients with less flexion at the fusion
site walked better than those with increased hip flexion. Less pelvic
anterior tilt and less sagittal plane motion were seen and the lumbar
spine moved less in the transverse plane if the hip was less flexed.
However, patients in whom the hip was fused in more extension complained
of difficulty in sitting. Videotapes of these patients made during
the transition from sitting to standing revealed thoracolumbar kyphosis
when the patient was seated. There does not appear to be a position of
flexion that allows easy sitting and pain-free walking; therefore,
on the basis of these preliminary data, we recommend the commonly
used position of moderate flexion of the hip (20 to 25 degrees).
Since knee instability was correlated with hip abduction, a position
of neutral hip abduction-adduction at the time of the final follow-up,
after postoperative adduction drift has occurred, should be attempted.
Our findings are disturbing, as abnormal lumbar spine motion
was evident in all nine patients and back pain was already present
in seven patients between seven and fourteen years after the arthrodesis.
Only the two patients who had been followed for two and three years
did not have back pain, but they demonstrated abnormal kinematic
movement of the lumbar spine. However, there are patients for whom
no other treatment can provide pain relief. On the basis of our
data, we believe that if an osteotomy can relieve pain in an adolescent
with hip-joint destruction then it may be reasonable to consider
performing this procedure instead of an arthrodesis of the hip.
Increased pelvic and lumbar motion and abnormal knee motion are
seen in young adults following an arthrodesis of the hip performed
during adolescence. This abnormal motion predisposes the patient
to back pain over time. The onset of back symptoms occurred earlier
in our patients than has been noted previously. Arthrodesis should
remain a last resort in the treatment of an adolescent who has severe
degeneration of the hip joint. The position of choice for an arthrodesis
of the hip is 20 to 25 degrees of flexion, neutral adduction-abduction,
and neutral rotation.
Note: The authors acknowledge the assistance of Cecilia Concha,
B.S., in data acquisition and model development, and Cindy Smith,
B.S., in patient testing.
Barnhardt, T., and Stiehl, J.
B.: Hip fusion in young adults. Orthopedics, 19: 303-306, 1996.
Benaroch, T. E.; Richards, B. S.;
Haideri, N.; and Smith, C.: Intermediate follow-up of a simple method
of hip arthrodesis in adolescent patients. J. Pediat. Orthop., 16:
30-36, 1996.
Callaghan, J. J.; Brand, R. A.; and
Pedersen, D. R.: Hip arthrodesis. A long-term follow-up. J. Bone
and Joint Surg., 67-A: 1328-1335, Dec. 1985.
Crosbie, J.; Vachalathiti, R.; and
Smith, R.: Patterns of spinal motion during walking. Gait and Posture,
5: 6-12, 1997.
Davis, R. B.; Ounpuu, S.; Tyburski,
D.; and Gage, J. R.: A gait analysis data collection and reduction
technique. Human Movement Sci., 10: 575-587, 1991.
Gore, D. R.; Murray, M. P.; Sepic,
S. B.; and Gardner, G. M.: Walking patterns of men with unilateral
surgical hip fusion. J Bone Joint Surg, 57-A: 759-765, Sept.
1975.
Harris, W. H.: Traumatic arthritis
of the hip after dislocation and acetabular fractures: treatment
by mold arthroplasty. An end-result study using a new method of
result evaluation. J Bone Joint Surg, 51-A: 737-755, June
1969.
Mowery, C. A.; Houkom, J. A.; Roach,
J. W.; and Sutherland, D. H.: A simple method of hip arthrodesis.
J. Pediat. Orthop., 6: 7-10, 1986.
Murrell, G. A. C., and Fitch, R. D.:
Hip fusion in young adults. Using a medial displacement osteotomy
and cobra plate. Clin. Orthop., 300: 147-154, 1994.
Perry, J.: Gait Analysis: Normal and
Pathological Function. New York, McGraw-Hill, 1992.
Price, C. T., and Lovell, W. W.:
Thompson arthrodesis of the hip in children. J Bone Joint Surg,
62-A: 1118-1123, Oct. 1980.
Roberts, C. S., and Fetto, J. F.:
Functional outcome of hip fusion in the young patient. Follow-up study
of 10 patients. J. Arthroplasty, 5: 89-96, 1990.
Sapega, A. A.: Current concepts review.
Muscle performance evaluation in orthopaedic practice. J. Bone and
Joint Surg., 72-A: 1562-1574, Dec. 1990.
Song, K. M.; Halliday, S. E.; and
Little, D. G.: The effect of limb-length discrepancy on gait. J. Bone
and Joint Surg., 79-A: 1690-1698, Nov. 1997.
Sponseller, P. D.; McBeath, A. A.;
and Perpich, M.: Hip arthrodesis in young patients. A long-term follow-up
study. J Bone Joint Surg, 66-A: 853-859, July 1984.
Waters, R. L.; Barnes, G.; Husserl,
T.; Silver, L.; and Liss, R.: Comparable energy expenditure after
arthrodesis of the hip and ankle. J Bone Joint Surg, 70-A:
1032-1037, Aug. 1988.
Whittle, M. W., and Levine, D.: Measurement of
lumbar lordosis as a component of clinical gait analysis. Gait and
Posture, 5: 101-107, 1997.