Although the hips generally are considered to be abnormal at birth in
subjects who later have development of cerebral palsy, these patients are at
risk for subluxation or dislocation of the hip as a result of spastic muscle
forces1,2.
The reported prevalence of spastic hip subluxation or dislocation among
individuals with cerebral palsy has been reported to range from 3% to
75%3,4.
Mathews et al.3
followed 1243 patients with cerebral palsy over a six-year period. The
severity of the cerebral palsy varied greatly. Thirty-two patients (3%) were
found to have hip subluxation or dislocation. In the study by Hoffer et
al.4, twenty
patients with severe cerebral palsy and painful hips were evaluated, and
fifteen (75%) had either unilateral or bilateral hip dislocation.
Higher rates of hip subluxation or dislocation are expected in children
with more severe
involvement5-7.
These individuals are unable to walk; thus, the presence of a dislocated hip
does not affect mobility. Conversely, a dislocated hip may result in sitting
imbalance with a possible risk of pressure
ulcers8-10.
Dislocation of the hip also limits the range of motion of the hip, which may
increase difficulty with perineal
care2,8
and also may increase the rate of femoral
fractures9-11.
Perhaps the greatest concern in the adult population is the potential for
the development of osteoarthritis and progressive hip pain. Pediatric patients
with cerebral palsy and spastic hip subluxation or dislocation are more
predisposed to the development of hip problems than adults are because of
abnormal mechanical forces in the hip and growth. As these patients age, their
function and long-term care can be compromised not only by pain secondary to
osteoarthritis but also by sitting imbalance, poor perineal hygiene, pressure
ulcers, and femoral
fracture5,6,8-10.
There have been no consistent reports on the prevalence of hip pain in
patients with cerebral palsy. The presence or absence of pain depends on
different variables such as the level of cognitive function and the ability of
each subject to perceive pain if
present5,8,10,12,13.
When faced with progressive subluxation or dislocation of the hip in
immature patients, pediatric orthopaedic surgeons typically recommend either
nonoperative modalities, such as physical therapy or orthotics, or operative
intervention to correct the deformity or to prevent its
progression5,14,15.
Most investigators have agreed that treatment is indicated for patients with
painful hip subluxation or
dislocation9,13,14;
however, controversy persists regarding the treatment of painless hip
dislocation or progressive subluxation. Previous investigators have
recommended operative treatment for younger patients with so-called hips at
risk for spastic subluxation or dislocation and for patients who present with
early spastic
subluxation5,15-19.
In these instances, relatively low-risk procedures such as muscle releases or
lengthenings may be effective for maintaining hip
position9.
It is controversial whether subjects with severe cognitive impairment and
documented painless hip dislocation should undergo treatment. Some authors
have advocated reduction of the hip in order to prevent eventual
osteoarthritis and
pain5,14-19,
whereas others have recommended no treatment because pain may not
develop9,10,20.
The latter approach may have merit because operative reconstruction in older
children and adolescents usually involves soft-tissue releases, open
reduction, and femoral and pelvic
osteotomy9,14.
These procedures are extensive and may result in complications such as wound
infection, osteomyelitis, malunion, nonunion, recurrent subluxation, and
continued pain as well as the need for hardware
removal2,4,15,17-19.
Regardless of one's philosophy toward painless dislocation of the hip in
patients with severe cerebral palsy, there are no consistent data to support
or refute either approach.
The purpose of the present study was to assess a group of mature
individuals with severe cerebral palsy and cognitive impairment. We evaluated
hip function, the prevalence of spastic hip subluxation or dislocation, and
radiographic signs of osteoarthritis. Primarily, we endeavored to determine
whether hip subluxation or dislocation and osteoarthritis are correlated with
a history of hip pain or limitations in hip-related activities of daily
living. The purpose of the present study was to determine whether children
with severe cerebral palsy and painless subluxation or dislocation of the hip
should undergo treatment to prevent problems as adults.
The medical directors from six residential centers identified a total of
seventy-two clients with an age of more than twenty-one years and a primary
diagnosis of cerebral palsy. In addition, five other individuals who remained
under the care of a family member while living at home were also recruited.
All individuals were considered to be severely affected by cerebral palsy. All
had severe muscle spasticity and contracture that precluded good walking
ability, and the majority lacked head and trunk control and/or had severe
mental retardation. The Indiana University Institutional Review Board approved
our study, and the guardian for each subject provided informed consent.
We attempted to correlate standard radiographic measures with subjective
and objective variables that were evaluated on the basis of a physical
examination, a review of the medical history, and an interview of the
predominant caregiver. The medical history, interview, and physical
examination were performed by one us (J.J.), and the radiographs were
evaluated by two of us (K.J.N. and J.P.) who had no prior knowledge of the
information obtained by the first investigator.
The medical history was obtained through a review of available records and
the interview with the caregiver. The patients were classified into four
groups on the basis of a new, unvalidated functional classification system
(Table I). We believe that this
classification system accurately depicts the severity of function in patients
with limited or no ability to walk and provides a good understanding of the
severity of involvement. The mean number of hours spent sitting in a chair or
a wheelchair each day was recorded. The investigator who performed the medical
review documented any history of seizure disorder, current use of a
gastrostomy feeding tube, recent pulmonary complications, and the presence of
pressure ulcers. Previous orthopaedic procedures and current orthotic and
wheelchair use were also recorded.
The caregiver of each subject was specifically asked if, in his or her
opinion, the hips of the subject were painful. There were six potential
answers: (1) the left hip was painful, (2) the right hip was painful, (3) both
hips were painful, (4) at least one of the hips was painful but the caregiver
was unsure which hip was painful, (5) the caregiver was not sure if the
subject had any pain, and (6) neither hip was painful. In addition, the
frequency of pain was recorded. The frequency of pain and the use of
anti-inflammatory medication to treat pain also were recorded. Finally, the
caregiver was asked whether there was difficulty in providing perineal
care.
A physical examination was performed for each patient. Vital signs were
assessed and the Face, Legs, Activity, Cry, and Consolability (FLACC)
behavioral pain score (see Appendix) was calculated both before and after the
physical examination, and the differences were recorded. The FLACC scale,
described by Merkel et
al.21, quantitates
pain on the basis of a 10-point system in which 0, 1, or 2 points are assigned
for described changes in the face, the legs, patient activity, crying, and
consolability. Maximum pain is recorded as 10 points, and no pain is recorded
as 0 points.
Each patient was examined for the presence of pressure ulcers, which were
classified as none, healed, superficial, or deep. Pelvic posture was graded as
neutral, windswept to the left (abduction of the left leg and adduction of the
right leg), or windswept to the right. The total abduction of the hips in
flexion and the presence of a flexion contracture of >30° were
recorded. The popliteal angles were measured as the number of degrees short of
full knee extension with the hip flexed to 90°. During examination of the
knee, muscle tone and spasticity were graded on a scale from 0 to 4 points as
described by
Ashworth22. On this
scale, no increase in tone on passive movement is graded as 0 points, a slight
increase in tone or catching is graded as 1 point, a marked but still easily
flexed knee is graded as 2 points, a considerable increase in tone making
passive movement difficult is graded as 3 points, and a rigid limb is graded
as 4 points.
An anteroposterior pelvic radiograph was made for each subject, and several
measurements were made. A hip was considered to be dislocated if there was no
contact between the femoral head and the true acetabulum. A hip was classified
as subluxated if there was at least partial contact between the femoral head
and the acetabulum and if a break in the Shenton line had occurred. The hip
migration percentage was calculated as described by
Reimers23. This is
accomplished by first measuring, in the horizontal plane, the part of the
femoral head extending lateral to a vertical line drawn from the lateral edge
of the roof of the acetabulum (the Perkins line). The migration percentage is
then calculated by dividing this value by the entire width of the visible part
of the femoral head. The center-edge angle and the acetabular depth were
measured for each hip, and the Sharp angle was determined by measuring the
angle formed by a horizontal line connecting the inferior aspects of the
acetabular teardrop with a line from the inferomedial part of the acetabulum
to superolateral part of the
acetabulum24.
Osteoarthritis was determined with use of the Kellgren-Lawrence
system25, in which
grade 0 indicates normal findings, grade 1 indicates doubtful joint-space
narrowing with possible osteophytes, grade 2 indicates definite joint-space
narrowing with osteophytes present, grade 3 indicates moderate osteoarthritis,
and grade 4 indicates severe osteoarthritis.
Statistical Methods
Descriptive statistics were performed on all variables. In the present
study, we sought to determine if physical examination findings or radiographic
measures were correlated either with a history of hip pain or with differences
in the ability to perform daily perineal care. Statistical comparisons between
nominal or categorical variables were performed with use of a chi-square test
or, in cases in which the sample size was small, with use of the Fisher exact
test. A t test or one-way analysis of variance was used to determine
significant differences between two or more subgroups. Pearson product-moment
correlations were performed to determine significant correlations. The level
of significance was set at p = 0.05.
Demographic Evaluation
The mean age of the seventy-seven subjects was forty years (range,
twenty-two to eighty-one years). Thirty-eight subjects were male, and
thirty-nine were female. Sixty-six (86%) of the seventy-seven individuals had
spastic quadriplegia (equal involvement of all four extremities); six (8%) had
severe spastic diplegia (with the lower extremities being more involved than
the upper extremities); and four (5%) had mixed spastic-athetoid quadriplegia.
The remaining subject had predominantly athetoid involvement of all four
extremities. Using our functional classification system, we determined that
fifty-six (73%) of the seventy-seven subjects were unable to communicate or
walk and had poor head and trunk control (Group 1), nine subjects (12%) were
unable to communicate or walk but had good head and trunk control (Group 2),
and twelve subjects (16%) were either able to communicate or had some ability
to walk (Group 3). Ten of the twelve subjects in Group 3 were able to walk in
the home, and the other two were able to walk in the community with
assistance. Only one patient was able to communicate clearly during the
interview and examination process. Each subject spent a mean of eleven hours
(range, zero to sixteen hours) sitting in either a chair or wheelchair. The
caregivers believed that the hips prevented good sitting balance in seventeen
(22%) of the seventy-seven subjects. Difficulty in performing good perineal
care was reported for only six subjects (8%).
On the basis of the interviews of the primary caregivers, we determined
that thirty-eight subjects (49%) had pain in at least one hip and thirty (39%)
did not have any pain in either hip. For the remaining nine subjects (12%),
the caregivers were unsure if there was pain in either hip.
Caregivers were asked to determine which hips were painful. As such,
sixteen right and twelve left hips were definitely painful and thirty-three
right and thirty-six left hips were definitely not painful. Questionable pain
was present in nineteen contralateral hips in subjects with or without
definite hip pain. In nineteen additional subjects, pain was present in one
hip but the caregiver was unable to determine which hip. When both left and
right hips were considered, sixty-nine (45%) of the 154 hips were definitely
not painful and twenty-eight (18%) were definitely painful. Fifty-three (69%)
of the seventy-seven subjects did not receive any pain or anti-inflammatory
medications, eighteen (23%) received such medications on an occasional basis,
and six (8%) received such medications on a daily basis.
Physical Examination
We detected minimal changes in vital signs when the measurements that had
been made before the physical examination were compared with those that had
been made after the physical examination. The mean change in the FLACC score
was 0.7 point (range, —1 to 6 points) when the pre-examination score was
compared with the post-examination score. We concluded that changes in these
measures could not be used to identify whether or not a hip was painful.
Seventy-one (92%) of the seventy-seven subjects were free of pressure
ulcers, three had recently healed ulcers, and three had active pressure
ulcers. Examination of hip motion demonstrated that the mean combined hip
abduction (in flexion) was 60° (range, 0° to 170°). Unfortunately,
in eleven subjects we were unable to quantify hip posture or the presence of a
hip contracture because of severe spasticity or positioning. As such,
thirty-seven subjects (48%) had neutral pelvic posture, sixteen (21%) had a
deformity in which the hips were windswept to the left, and thirteen (17%) had
a deformity in which the hips were windswept to the right. Thirteen subjects
(17%) had a hip flexion contracture of >30° in one or both hips, and
fifty-three (69%) did not have any hip contracture.
Twenty-five subjects (32%) had hamstring muscle contractures or spasticity
that resulted in a popliteal angle of at least 90° in one or both knees.
Spasticity was assessed during the physical examination according to the
method described by
Ashworth22.
Nineteen (25%) of the seventy-seven subjects had either no or mild spasticity
(0 or 1 point), forty-three (56%) had moderate spasticity (2 or 3 points), and
the remaining fifteen (19%) had severe spasticity (4 points).
Radiographic Analysis
Forty-one (27%) of the 154 hips were subluxated or dislocated;
specifically, eighteen were subluxated and twenty-three were frankly
dislocated. Two hips had had a proximal femoral resection, and one hip could
not be fully assessed with regard to complete dislocation secondary to poor
radiographic technique and poor film quality. The remaining 110 hips (71%)
were concentrically located within the acetabulum. In the group of 151 hips
that could be accurately evaluated, the mean Reimer migration percentage was
32% (range, 0% to 100%), the mean center-edge angle was 26° (range,
—120° to 66°), the mean acetabular depth was 17 mm (range, 6 to
33 mm), and the mean Sharp angle was 38° (range, 12° to 57°).
With use of the Kellgren-Lawrence radiographic classification
system25, we
determined that 108 (70%) of the 154 hips had normal findings, eleven (7%) had
doubtful joint-space narrowing, nine (6%) had definite joint-space narrowing,
twelve (8%) had moderate osteoarthritis, and fourteen (9%) had severe
osteoarthritis or had had a proximal femoral resection. Therefore, thirty-five
(23%) of the 154 hips had some sign of osteoarthritis.
Statistical Analysis
Subject-Specific Analysis
Increased spasticity at the knee was associated with pain in the hip
(Table II). Eleven (73%) of
fifteen subjects with severe spasticity had hip pain, compared with twenty-one
(49%) of forty-three subjects with moderate spasticity and six (32%) of
nineteen subjects with no or mild spasticity (p = 0.04). In addition, subjects
with hip pain tended to have stiffer hips or abnormal hip positioning (defined
as any abduction, adduction, or flexion contracture of >30°). For
instance, the mean combined abduction was 50° in subjects with hip pain
compared with 75° in those without hip pain (p = 0.01). The prevalence of
hip pain among subjects with a flexion contracture in at least one hip was
higher than that among subjects without a flexion contracture (eleven [85%] of
thirteen compared with twenty-seven [51%] of fifty-three; p = 0.07).
Similarly, the prevalence of hip pain among patients with windswept hips was
higher than that among patients with neutral hip posture (twenty-two [76%] of
twenty-nine compared with sixteen [43%] of thirty-seven; p = 0.02). Pain
medication was used at least occasionally for eight (62%) of the thirteen
subjects with hip flexion contractures of >30°, compared with only
sixteen (30%) of the fifty-three subjects without a hip flexion contracture (p
= 0.04). With the numbers available, no other subject-specific variable such
as age, gender, individual function (in terms of either the functional
classification or walking ability), time spent sitting, scoliosis, knee
contracture, pelvic obliquity, dislocation in either hip, or osteoarthritis in
either hip had an effect on subject-specific hip pain, the frequency of pain,
or the frequency of use of pain medication.
We identified an association between deformity and the presence of recent
or active pressure ulcers. The mean combined hip abduction was 25° in the
six subjects with physical signs of current or recent pressure ulcers,
compared with 62° in the seventy-one subjects without ulcers (p = 0.02).
Ulcers were noted in zero of thirty-seven subjects with neutral hip
positioning, compared with five (17%) of twenty-nine subjects with windswept
hip deformities (p = 0.01). Caregivers noted increased problems with perineal
care when combined hip abduction was <30° (p = 0.01).
Finally, we found a positive relationship between patient age and the rates
of hip dislocation and radiographic signs of osteoarthritis in either hip. The
mean age was forty-four years for subjects with hip dislocation, compared with
thirty-eight years for subjects without hip dislocation (p = 0.04). Similarly,
the mean age was forty-four years for subjects with osteoarthritis of the hip,
compared with thirty-eight years for subjects without osteoarthritis (p =
0.05).
Hip-Specific Analysis
When we combined the radiographic measurements of the left and right hips,
it became apparent that increased femoral head displacement was correlated
with increased rates of osteoarthritis (as indicated by Kellgren-Lawrence
grade-2, 3, and 4 changes). The mean migration percentage was 74% in the hips
with osteoarthritis, compared with 21% in those without osteoarthritis (p =
0.0001). The mean center-edge angle was 8° (range, —40° to
44°) in hips with osteoarthritis, compared with 29° (range,
—120° to 66°) in hips without osteoarthritis (p = 0.0001). Hip
subluxation (disruption of the Shenton line) also heralded radiographic signs
of osteoarthritis. The Shenton line was disrupted in thirty of thirty-three
hips with osteoarthritis but in only eleven of 115 hips without osteoarthritis
(p = 0.0001). Similarly, nineteen of twenty-three dislocated hips and eleven
of eighteen subluxated hips had osteoarthritis; these rates were significantly
higher than the rate of osteoarthritis in located hips (thirteen of 110) (p =
0.0001). Measures of acetabular dysplasia had variable correlations with the
rate of osteoarthritis. The mean acetabular depth was 17.5 mm in hips without
osteoarthritis, compared with 14.3 mm in hips with osteoarthritis (p =
0.01).
We found no consistent associations between hip displacement and a history
of hip-specific pain. With the numbers available, there was no detectable
relationship between increased migration percentage, disruption of the Shenton
line, or dislocation and an increased prevalence of definite hip pain. The
rates of definite hip pain in dislocated hips (five of twenty-three) and
subluxated hips (five of eighteen) were not significantly different from the
rate in hips that were not dislocated (eighteen of 110 hips) (p = 0.38). With
the numbers available, there was no difference in the prevalence of
hip-specific osteoarthritis in hips that were definitely painful compared with
hips that were definitely not painful.
Interestingly, we found variable associations between increased definite
hip pain and increased amounts of acetabular dysplasia. With the numbers
available, acetabular depth was not prognostic of hip pain. On the other hand,
the mean Sharp angle was 42° (range, 35° to 47°) in hips that were
definitely painful compared with 37° (range, 12° to 56°) in those
that were definitely not painful (p = 0.01). Furthermore, the mean center-edge
angle was 19° (range, —40° to 44°) in hips that were
definitely painful compared with 30° (range, —19° to 66°) in
hips that were definitely not painful (p = 0.02).
The prevalence of cerebral palsy in the general population has been
reported to range from 0.1% to
0.7%26,27.
This rate is not likely to decrease soon, as low-birth-weight babies are
surviving premature delivery with sequelae of immature brain development or
injury that can potentially result in eventual cerebral
palsy28.
Previous studies have demonstrated variations in the rates of eventual hip
pain as a result of hip dislocation in patients with cerebral
palsy5,8,10,20.
Cooperman et al.5,
in their study of patients who had cerebral palsy with spastic quadriplegia,
found that twenty-four (57%) of forty-two dislocated hips were painful. Moreau
et al.8 followed
twenty-one patients who had twenty-four dislocated hips and found that eleven
patients had hip pain. In addition, two of nine patients with hip subluxation
had hip pain. Conversely,
Pritchett10 noted
that seventy-six of eighty patients with unstable hips had either no or minor
hip pain. Similarly, O'Brien and
Sirking20 found
that hip pain was not a major problem in their patients with spastic hip
subluxation or dislocation.
In the current study, we asked caregivers if pain was present in the hips
of their clients as well as if they could tell which hip was painful. It could
be argued that such an assessment is flawed because we can never be sure that
such pain exists; however, one should recognize that this assessment is
probably the only available indicator of pain in these profoundly affected
individuals. Furthermore, it is also the most practical indicator because it
is the assessment by each caregiver that may lead to further evaluation and
treatment. We determined that thirty-eight (49%) of the seventy-seven subjects
had some hip pain; unfortunately, the caregivers could only localize the pain
to a particular hip in twenty-eight (18%) of the 154 available hips. Overall,
45% (sixty-nine) of the 154 hips were definitely not painful, 18%
(twenty-eight) were definitely painful, and 37% (fifty-seven) may or may not
have been painful. In addition, the amount of time spent sitting was not a
reliable indicator of pain as it was not correlated with a history of pain,
dislocation, or radiographic evidence of osteoarthritis.
In our subject-specific analysis, it appeared that abnormal hip motion and
position were associated with generalized hip pain. Hips that were windswept
with decreased abduction or with residual flexion contractures were more
likely to be painful than hips with better motion and more neutral
positioning. Although the overall rate of pressure ulcers was low, we also
noted that the only subjects who had ulcers were those who had limited
abduction or windswept hip positioning. These findings are consistent with
those of Moreau et
al.8, who also found
higher rates of pressure ulcers in patients with hip contractures. Spasticity
also seemed to have important implications in this cohort as we found
increased rates of pain in subgroups of subjects with more spasticity. As a
result of these findings, perhaps greater emphasis should be placed on
maintaining neutral sitting position and motion and perhaps better control of
spasticity would improve the quality of life for these severely affected
subjects.
Very little data are available on the rates of radiographic signs of
osteoarthritis in individuals with spastic hip subluxation or dislocation as a
result of cerebral palsy. In the present study, radiographic signs of
osteoarthritis were found in thirty-five (23%) of the 154 hips. We found a
very strong correlation between radiographic evidence of osteoarthritis and
hip displacement as quantified on the basis of dislocation, subluxation,
increased migration percentage, and decreased center-edge angle (p < 0.001
for all). In the hip-specific analysis, we found that hip dislocation, hip
subluxation, and radiographic evidence of osteoarthritis were not associated
with the caregiver's assessment of hip pain. This finding is similar to that
of Pritchett9, who
reported the rates of hip pain for two separate groups of fifty severely
affected patients. In one group, no attempt at hip reduction was made and the
prevalence of hip pain was 42% (twenty-one of fifty patients). In the other
group, successful hip reduction was performed in 58% (twenty-nine) of the
fifty patients and yet hip pain was present in 38% (nineteen) of the fifty
patients. The rate of pain in the group of patients who were treated
surgically was not significantly different from that in the group of patients
who were not treated surgically.
In the present study, we recorded historical and physical examination
parameters in a blinded fashion, separate from the measurements made on each
subject's radiographs. Despite this advantage, problems in study design were
present. For instance, this series is not a true natural-history study because
some subjects had had previous surgical procedures as evidenced by
appropriately placed scars in twenty individuals. In addition to the potential
for reporting biases or inaccuracies by the caregivers, our physical
examination utilized physical measures of spasticity and range of motion that
do not have excellent
reliability29.
Finally, a certain amount of variability in the radiographic measurement of
subtle acetabular dysplasia must
exist23.
Despite the limitations described above, we have presented some important
findings regarding this study population after relatively long-term follow-up
(mean, forty years). The cohort was substantially older than those in similar
studies in which the mean age was twenty-five to twenty-six years of
age5,8-10.
However, in reviewing our results, it is important to remember that these
findings should be applied only to profoundly affected individuals with severe
cerebral palsy. These findings probably do not apply to more functional
patients or patients who are able to accurately express pain and
limitations.
The primary goal of the present study was to determine if radiographic
evidence of hip dislocation was correlated with sitting problems and hip pain.
We could not demonstrate such an association. We did note that hip
displacement was associated with radiographic evidence of osteoarthritis, but
the osteoarthritis did not seem to be related to hip pain. As such, if the
goal is to reduce potential or future hip pain, aggressive surgical methods
designed to reduce or to maintain hip location in profoundly affected
asymptomatic individuals might not be justified. Conversely, one might argue
that subjects in the present study with radiographic evidence of
osteoarthritis must have pain despite a contrary or inconclusive assessment by
the appropriate caregiver. However, attempts at reconstructive hip surgery in
such individuals should be weighed against their potential complications and
the concept that reducing spasticity and maintaining symmetric motion may be
more important than having a located hip as documented on a radiograph.
?
The details of the FLACC pain-assessment tool are available with the
electronic versions of this article, on our web site at
(go to
the article citation and click on "Supplementary Material") and on
our quarterly CD-ROM (call our subscription department, at 781-449-9780, to
order the CD-ROM).