The Israel Center for Familial Dysautonomia at Hadassah University
Hospital in Mount Scopus and Hebrew University in Jerusalem, Israel,
is a national referral center for patients with familial dysautonomia.
One hundred and thirty-six patients who were registered in the center
formed the base of our study. The charts and radiographs of all
of the patients were reviewed. Of ninety living patients, sixty-four
were available for examination by the senior author (E. B.-O.) at
the time of the study.
We recorded data regarding milestones in motor development, gait
disturbances, spine and limb deformities, arthropathies, fractures,
and musculoskeletal infections or decubitus ulcers, as well as any additional
problems that may affect musculoskeletal function.
The records of 136 patients (sixty-five males and seventy-one
females) were reviewed. The mean age of the patients at the time
of the review was sixteen years (range, three months to forty-six years).
Fifty-six patients (41 percent) were in the first decade of life,
forty-one (30 percent) were in the second decade, and thirty-nine
(29 percent) were in the third or fourth decade. At the time of the
review, forty-six patients had died. The mean age at the time of
death was nineteen years (range, three months to forty-four years).
Fourteen (30 percent) of these forty-six patients died in the first decade
of life; ten (22 percent), in the second decade; and twenty-two
(48 percent), in the third or fourth decade.
The most common causes of death were acute aspiration (thirteen
patients; 28 percent) and chronic pneumonia (ten patients; 22 percent).
Additional causes included sudden death during sleep (eight patients;
17 percent), trauma from either a fall or a motor-vehicle accident
(two patients; 4 percent), and various other causes such as drug
overdose or surgical complications (four patients; 9 percent). The
cause of death was not clear for the remaining nine patients (20
percent).
The most common orthopaedic problems were gait abnormalities,
spinal deformities, foot deformities, fractures, and Charcot arthropathy.
Gait Abnormalities
Independent walking was started at a mean age of twenty-seven
months (range, fourteen to eighty-four months). It was delayed beyond
the normal age of eighteen months in 72 percent (ninety-four) of
the 130 patients who were of walking age. All sixty-four patients
who were examined had an ataxic gait, although the degree of ataxia
was highly variable. Despite these gait disturbances, all of these
patients retained the ability to walk and none required assistive
devices.
Foot Deformities
Foot deformities were found in sixteen (12 percent) of the 136
patients. Nine patients had an equinovarus or cavovarus deformity,
five had a planovalgus deformity, and two had polydactyly.
An equinovarus or cavovarus deformity developed during childhood
in eight patients; the remaining patient who had such a deformity
had a congenital clubfoot. No associated spasticity was found in any
patient. Three of the nine patients had operative treatment. The
patient with a congenital clubfoot had a posteromedial release at
the age of five months. He needed a re-release at the age of two years,
but, at the time of the review, correction had been maintained through
the age of thirteen years. The other two patients had surgery for
a progressive deformity. One patient had a plantar and posterior
release at the age of nine years. She required a relengthening of
the Achilles tendon one year later, but, at the time of the review,
correction had been maintained through the age of fifteen years. The
other patient had a posterior release at the age of thirteen years
and wore an ankle-foot orthosis until his death at the age of twenty-nine
years. In the remaining six patients, the deformity was adequately
managed with shoe modifications and did not interfere with walking.
Although all five patients who had a planovalgus deformity had
bothersome pressure points on the medial side of the foot, four
of the patients were adequately treated with shoe modifications.
The fifth patient had a bilateral distal tibiofibular derotation
osteotomy. She subsequently had multiple decubitus ulcers and osteomyelitis
of the os calcis, which were treated with d衲idement.
Both of the patients with polydactyly were treated operatively.
Spinal Deformity
Spinal deformity was found in seventy-eight patients. Thirty-six
of these patients were male and forty-two were female, for a male:female
ratio of 1:1.2. Spinal deformity was diagnosed at a mean age of
eight years (range, four to sixteen years), and baseline radiographs
were made once the deformity was noted clinically. The prevalence
of spinal deformity was 48 percent (thirty-eight of eighty) by the
age of ten years, 86 percent (forty-eight of fifty-six) by the age
of fifteen years, and 85 percent (thirty-three of thirty-nine) by
the age of twenty years or more.
Forty-one patients (53 percent) had scoliosis only, thirty-four
(44 percent) had kyphoscoliosis, and three (4 percent) had kyphosis
without scoliosis. The curve pattern of the scoliosis in the seventy-five
patients with a coronal deformity was highly variable. Forty-nine
patients (65 percent) had a single thoracic or thoracolumbar curve;
twenty-eight of these curves were to the right and twenty-one were
to the left. Twenty-six patients (35 percent) had a double curve;
fourteen had a right thoracic or thoracolumbar curve and a left
lumbar curve, and twelve had a left thoracic or thoracolumbar curve
and a right lumbar curve. Thoracic hyperkyphosis was found in 65
percent (seventeen) of the twenty-six patients with a double curve
but in only 33 percent (sixteen) of the forty-nine patients with
a single curve.
Nonoperative Treatment
Twenty-three patients were treated with a cast or brace. A cast
was used for one patient who was treated early, a Milwaukee brace
was used for sixteen patients, and an underarm thoracolumbar spinal
orthosis was used for eight patients (two of whom had initially
been treated with a Milwaukee brace). Bracing was initiated at a
mean age of ten years and was continued for a mean of three years. The
mean curve was 34 degrees (range, 23 to 51 degrees) at the start
of bracing and 56 degrees (range, 28 to 79 degrees) at the end of
bracing. Bracing was accompanied by multiple pulmonary problems
and pressure sores, causing discontinuation of treatment in four
patients. These problems, as well as the emotional lability associated
with familial dysautonomia, resulted in extremely poor compliance
with the bracing. Three patients were lost to follow-up during brace
treatment, and seventeen patients showed marked progression of the deformity.
In only three patients was bracing continued successfully to maturity,
with documented arrest of curve progression.
Operative Treatment
Twenty-four patients were treated operatively. In general, surgery
was considered when the deformity reached 40 degrees of scoliosis
and/or 65 degrees of kyphosis. However, the exact indication was
dependent on the age and size of the patient, with attempts being
made to postpone arthrodesis until adolescence. The decision was
influenced by a number of other factors, including the patient's general,
pulmonary, and nutritional status as well as parental consent.
The mean age at the time of surgery was thirteen years (range,
five to eighteen years). Fourteen patients had a single thoracic
or thoracolumbar curve: eight curves were to the right and six were to
the left. Ten patients had a double major curve: seven had a right
thoracic and left lumbar curve, and three had a left thoracic and
right lumbar curve. Seventeen patients had a kyphotic curve of more
than 50 degrees,and in three
patients the sagittal deformity was the primary indication for surgery.
Twenty patients had a posterior spinal arthrodesis with use of
Harrington instrumentation (thirteen patients), Harrington-Luque
instrumentation (one patient), or Cotrel-Dubousset instrumentation
(six patients). Four other patients, who had a kyphotic deformity
of 87 to 110 degrees, underwent combined one-stage anterior and
posterior arthrodesis with Cotrel-Dubousset instrumentation. Harrington
or Harrington-Luque instrumentation was used until 1991; since then,
Cotrel-Dubousset has been the instrumentation of choice.
In the group as a whole, scoliosis was corrected from a mean
of 55 degrees (range, 20 to 90 degrees) to a mean of 35 degrees
(range, 20 to 50 degrees), which represented a mean correction of
36 percent. In the group of twenty-one patients in whom the primary
indication for surgery was progressive coronal deformity, the mean
scoliotic curve measured 60 degrees (range, 38 to 90 degrees) preoperatively
and 37 degrees (range, 20 to 50 degrees) immediately postoperatively,
constituting a mean correction of 38 percent (range, 19 to 68 percent).
When these patients were divided into those with and without kyphosis,
we found no differences between the two groups regarding the magnitude
of the preoperative coronal curve or the amount of correction obtained
at the time of the operation.
In the group as a whole, kyphosis was corrected from a mean of
69 degrees (range, 43 to 110 degrees) to a mean of 61 degrees (range
30 to 100 degrees), which represented a mean correction of 12 percent
(range, -29 to 48 percent). In the group of three patients in whom
the sagittal deformity was the primary indication for surgery, the
mean kyphotic curve measured 80 degrees preoperatively and 71 degrees
postoperatively; specifically, the curve decreased from 63 to 44
degrees in one patient, from 87 to 85 degrees in another, and from 90
to 85 degrees in the third.
At the time of the latest examination, performed at a mean of
sixty-five months (range, twelve to 192 months) postoperatively,
the mean scoliosis measured 49 degrees (range, 18 to 62 degrees)
and the mean kyphosis measured 67 degrees (range, 39 to 115 degrees).
Compared with the preoperative values, this constituted a mean correction
of 18 percent (range, -25 to 50 percent) for scoliosis and 3 percent
(range, -64 to 50 percent) for kyphosis. Progression at the site
of the arthrodesis did not occur unless the instrumentation had
failed. Progression of scoliosis occurred both distal (Fig. 1) and proximal
to the site of the arthrodesis, whereas progression of kyphosis
occurred only proximal to the site of the arthrodesis (Fig. 2).
Complications and Reoperations
There were a total of nineteen complications in fifteen patients.
Early complications included severe postoperative hypovolemia in
two patients, one of whom died and one of whom was resuscitated. Four
patients had pneumonia. One case progressed to generalized sepsis
and necessitated a prolonged period of intensive care in the hospital.
One patient had upper gastrointestinal bleeding.
Twelve complications, in nine patients, were related to the hardware
used in the spinal fixation. The complications included hook disengagement
at either end of the instrumentation in six patients (three of whom
had been treated with Harrington instrumentation and three of whom
had been treated with Cotrel-Dubousset instrumentation) (Fig. 3), hardware breakage
in two patients who had been treated with Harrington instrumentation,
screw pullout in one patient who had been treated with Cotrel-Dubousset
instrumentation (Fig. 4), and screw misplacement in one
patient who had been treated with Cotrel-Dubousset instrumentation.
These complications generally occurred in the early postoperative
period, and there were no proven cases of pseudarthrosis. One patient
had a deep infection with Proteus mirabilis, and
one had skin irritation over the hardware. In addition, four patients
had decompensation in the coronal plane. Two of them had revision
surgery, at six and sixteen years after the initial arthrodesis.
Eight (33 percent) of the twenty-four patients who had been treated
surgically needed at least one additional procedure. A total of
ten revision operations were performed.
Fractures
A total of eighty-seven fractures occurred in fifty-five patients
(40 percent). Twenty-seven patients had one fracture, thirteen had
two, and nine had three or more. Eighty-three of the eighty-seven fractures
occurred before skeletal maturity. The anatomical distribution of
the fractures is shown in Table I. Because of decreased pain sensation,
fractures were often diagnosed late, after parents noted local swelling
and warmth. This was especially true of fractures of the hand and
foot. One fracture of the proximal aspect of the femur was treated
with internal fixation. All other fractures were treated nonoperatively.
Fracture-healing often occurred with profuse callus formation. There
were no cases of delayed union or nonunion.
Charcot Joints
Evidence of Charcot arthropathy (joint effusion, instability,
and arthritic changes) was found in two patients. One of the patients
had bilateral ankle involvement, which was treated with observation only,
and the other patient had severe genu valgum (Fig. 5), which was
treated with a distal femoral osteotomy.
Additional Musculoskeletal Problems
Three patients had an obstetrical brachial plexus palsy, two
were being followed for asymptomatic acetabular dysplasia, one had
mild spastic hemiplegia due to neonatal asphyxia, one had Perthes disease
and was treated with a varus derotation osteotomy, and one had bilateral
slipped capital femoral epiphysis at the age of five years. This
last patient had received prolonged corticosteroid treatment for
severe pulmonary problems. One of the hips in this patient was painless
and was treated nonoperatively; the contralateral hip was painful
and was treated operatively with screw fixation (Fig. 6). Although a
pseudarthrosis developed on the side that had been treated nonoperatively,
the patient regained the ability to walk and the pain on the contralateral
side reversed after screw fixation.
There are two known major concentrations of patients with familial
dysautonomia: one is in North America, and the other is in Israel.
In 1971, Yoslow et al.24 reviewed
the orthopaedic problems in a group of sixty-five patients at New
York University Medical Center. Rubery et al.20 recently
reported their experience with surgical treatment of spinal deformity
in twenty-two patients from the same center. The treatment of spinal
deformity in some of the patients followed in the Jerusalem center
was reviewed in 1984 by Robin19 and
more recently by Kaplan et al.12.
Because of the rarity of familial dysautonomia, several reports
regarding the musculoskeletal aspects of the disease have included
small numbers of patients or have grouped them with patients who
had various other neuromuscular disorders that predominantly affect
sensory pathways or motor function1,10,11.
However, the problems found in patients with familial dysautonomia
differ importantly from those found in patients with other neuromuscular
conditions2. In addition, patients
with familial dysautonomia have a variety of medical problems that
have a direct effect on orthopaedic management. Although the deformities
seen in these patients seem to be of neurogenic origin, the neurological
defect that is associated with the disease affects the afferent pathways
and the agonist-antagonist muscle balance is normal and without
spasticity. The gait disturbance is mainly due to ataxia, probably resulting
from defective proprioception with the additional factors of unbalanced
spinal deformity and generalized weakness. Although the onset of gait
was delayed in most of our patients, they retained independent walking
ability throughout life. Muscle balance around the foot and ankle
is normal, and the reason for the increased prevalence of foot deformities
in patients with the disease remains unclear.
Joint problems were rare in our study. Only two (1.5 percent)
of the 136 patients had acetabular dysplasia compared with eleven
(17 percent) of the sixty-five patients in the study by Yoslow et
al.24. However, because pelvic
radiographs were not made routinely in the present study and because the
dysplasia in both patients was asymptomatic and discovered incidentally,
we cannot comment on the true prevalence of hip dysplasia in our
series.
One of our patients had Perthes disease, and two had Charcot
joints. Our findings may be comparable with those of Mitnick et
al.16. Nine of their 180 patients
had avascular necrosis, and three of the nine had necrosis of the
hip. Distinguishing between Charcot joints and avascular necrosis
in peripheral joints may be difficult. Slipped capital femoral epiphysis
has not been previously described in association with familial dysautonomia,
to our knowledge. It may have occurred in one of our patients because
of prolonged steroid treatment.
The prevalence of fractures in patients with familial dysautonomia
is not known. Yoslow et al.24 reported
that nineteen (29 percent) of sixty-five patients with familial
dysautonomia had sustained a total of twenty-eight fractures and
stated that this prevalence was higher than would be expected in a
normal population. However, in a large epidemiological survey of
fractures in normal children, Landin14 found
a prevalence of 42 percent in boys and 27 percent in girls in the
general population of patients who were sixteen years old or less.
In the present study, 40 percent (fifty-five) of the 136 patients
(including 43 percent [twenty-eight] of the sixty-five male patients
and 38 percent [twenty-seven] of the seventy-one female patients)
sustained a fracture before skeletal maturity, a rate similar to
the normal value found by Landin. Our findings, however, only include
diagnosed fractures, many of which were detected late, and we presume
that there were additional fractures that healed unnoticed.
In addition, the number of fractures per patient was found to
be substantially higher than in the normal population. Whereas Landin
found 8682 fractures in 7961 children, for a mean of 1.1 fractures
per child sustaining a fracture, we found a total of eighty-three
fractures sustained before skeletal maturity in fifty-five children,
for a mean of 1.5 fractures per child. When one considers that patients
with familial dysautonomia are less involved in outdoor activities
than healthy children are, we agree with Yoslow et al. that there
is an increased prevalence of fractures in patients with familial
dysautonomia. The reason for this is unclear, but it may be due
to a reduction in the protective effect of pain sensation as a result
of the sensory neuropathy. In our patients, fracture-healing appeared
to be normal and was accompanied by profuse callus production. We
did not encounter any cases of delayed healing or nonunion of fractures.
Of special interest is the pathophysiology of the spinal deformity
in patients with familial dysautonomia. The atypical curve patterns
and their natural history are similar to those found in patients with
various other neuromuscular conditions. However, patients with familial
dysautonomia do not have the muscle weakness, imbalance, and lack
of coordination that often are associated with these conditions.
The reason for the development of the spinal deformities may be
the disturbance in the afferent sensory pathways, and, indeed, an
association between altered proprioceptive and vibratory sensation
and idiopathic scoliosis has been reported22,23.
The severity of the spinal deformity in patients with familial dysautonomia
manifests not only in a high prevalence but also in early and severe
progression. As the disease itself causes obstructive changes in
pulmonary function, the effects of severe spinal deformity can be
extremely harmful, adding a restrictive component, particularly
as patients survive into adulthood with better medical management.
In our series, bracing was not found to be effective for preventing
or delaying curve progression, and its negative side effects - both cutaneous
and pulmonary - were considerable. We believe that surgical intervention
should not be postponed once notable curve progression is documented.
Earlier intervention may increase the amount of correction that
can be achieved. In the present study, the rate of complications
was high (79 percent; nineteen of twenty-four), with one-third (eight)
of the twenty-four patients having a total of ten additional surgical
procedures; however, it compares favorably with the rates reported
in other series. Albanese and Bobechko1 reported
complications in all sixteen of their patients. Rubery et al.20 reported that fifteen (68 percent)
of their twenty-two patients had a major complication and that a total
of twenty-nine reoperations were performed to treat complications.
The postoperative loss of correction during the follow-up period
is of concern and probably has several causes. There was a high
rate of fixation failure, resulting from both hardware breakage
and hook and screw pullout. The deforming force is great, as evidenced
by the severe preoperative progression. Many patients have poor
bone quality, and increased early motion postoperatively (resulting
from a lack of guarding created by the sensory changes) also may
play a part in weakening the fixation (Fig. 3 and Fig. 4). An additional cause of loss of
correction was continued progression of the deformity proximal and
distal to the site of the arthrodesis, especially in the sagittal
plane. The development of severe kyphosis proximal to the site of
the arthrodesis has also been observed by other authors1,20. We therefore agree with Rubery
et al.20 that the arthrodesis
should be extended as far proximally as possible. Although curve
progression can also occur distally and although extending the arthrodesis
distally might prevent such progression, we believe that lumbar
motion segments should be preserved whenever possible because most
patients with familial dysautonomia are usually fully able to walk.
The problem of instrumentation failure may be partially solved
by improved constructs combining hooks, wires, and pedicle screws.
However, we believe that the possibility of instrumentation failure
should be anticipated and that parents should be counseled accordingly.
Although the postoperative loss of correction is of concern, deformity
in the arthrodesed segments did not progress beyond preoperative
values unless the instrumentation failed. Considering the natural
history of severe progressive deformity in untreated patients6, we believe that preventing this
progression is a reasonable operative goal that can be achieved with
posterior spinal arthrodesis in most patients. The complication
rate, although high, is decreasing because of improved medical,
surgical, and anesthetic management3,4,21.
The small number of patients treated with combined anterior and
posterior arthrodesis does not allow us to draw clear conclusions
regarding the efficacy of this approach.