Data on the patients were obtained retrospectively from the files
of the Dysautonomia Treatment and Evaluation Center at New York
University Medical Center. Of the 480 patients with familial dysautonomia
who were listed in the Center's registry, 165 had been born prior
to 1976 and had been examined at the Center. Only the 123 patients who
had reached the minimum age of twenty years were included in the
study. At the time of the survey, seventy-four of these patients
were still alive.
Thirty-three patients were at least twenty years old when they
were referred to our institution. Twenty-two of these patients had
been diagnosed with spinal deformity before the time of referral;
the mean age at the time of diagnosis was nine years (range, four
to seventeen years). Although these twenty-two patients had been
diagnosed at another institution, the compilation of data for the
study began at the time of the diagnosis. The magnitude of the initial
curve could not be determined for nine of these patients (six female
and three male) because the early radiographs could not be obtained.
According to the records, the deformity had been diagnosed at a
mean age of eight years and all nine patients had been treated with
observation only.
Data on the spinal curves were obtained from computer records
and radiographs. For some patients, only serial chest radiographs
were available, but these radiographs were adequate for the detection and
measurement of thoracic spinal deformity. Scoliosis was defined
as a Cobb angle7 of 11 degrees
or more in the coronal plane, and kyphosis was defined as an angle
of 41 degrees or more in the sagittal plane. Scoliosis was classified as
mild (11 to 20 degrees), moderate (21 to 40 degrees), or severe
(41 degrees or more). In the present study, kyphosis was limited
to the thoracic spine and, in most patients, the apex of the kyphosis
was at the fifth thoracic level.
In addition to the measurement of spinal curvature, we also evaluated
the type, duration, and effect of treatment as well as the age at
which treatment was initiated.
Statistical Analysis
A statistician evaluated all of our data. Variables were tested
with one-way analysis of variance followed by Tukey's HSD (honestly
significant difference) test for multiple comparisons. A p value of
less than 0.05 was considered significant.
Survivorship Analysis
Kaplan-Meier survivorship analysis9 was
performed to estimate the probability of development of scoliosis
or kyphosis. The end point was defined as the age at which the diagnosis
of spinal deformity was made. Therefore, in our study the term "survivorship"
does not refer to the patients' being alive but to their having
not been diagnosed with spinal deformity.
There were sixty-nine female patients and fifty-four male patients.
Most of the patients were first seen at the Dysautonomia Treatment
and Evaluation Center during the first year of life, but some were
not referred until adolescence or even adulthood. The mean age at
the time of the initial evaluation at our institution was sixteen
years (range, six months to forty-four years), and the mean duration
of follow-up was 16.5 years (range, one to twenty-six years).
Epidemiology of the Spinal Deformity
Of the 123 patients with familial dysautonomia, 102 (83 percent)
had been diagnosed with spinal deformity and twenty-one (17 percent)
had not. Sixty-nine patients (56 percent) had scoliosis only, thirty-one
(25 percent) had scoliosis as well as kyphosis, and only two (2
percent) had kyphosis without scoliosis. The mean scoliotic curve
was 30 degrees, and the mean kyphotic curve was 79 degrees. Scoliosis
was diagnosed at a mean age of nine years (range, six months to
seventeen years), and kyphosis was diagnosed at a mean age of eight years
(range, three to sixteen years).
Both genders were affected equally. The male:female ratio in
the population of patients who had scoliosis was 0.82 (forty-five
male to fifty-five female patients). This was similar to the ratio
in the overall study population, which was 0.78 (fifty-four male
to sixty-nine female patients).
Kaplan-Meier analysis, with the end point defined as the age
at which the diagnosis of spinal deformity was made, revealed that
sixty-four (52 percent) of the 123 patients with familial dysautonomia
had scoliosis by the age of ten years (Fig. 1) and twenty-six (21 percent) had
kyphosis (with or without scoliosis) by the same age (Fig. 2).
Type and Severity of Scoliosis
Unlike idiopathic scoliosis, in which most curves are right thoracic
and left lumbar, forty of the 100 scoliotic curves in our series
were left thoracic. Nine were thoracic curves only, thirty were
associated with an additional right lumbar curve, and one was a
left thoracolumbar curve. At the time of presentation, eighteen
of the 100 curves were rated as mild; twenty-two, as moderate; and
sixty, as severe.
Treatment
The results of conservative treatment were assessed in ninety-four
of the 102 patients with spinal deformity. These ninety-four patients
were divided into two groups: the observation group (twenty-nine
patients) and the bracing group (sixty-five patients). The remaining
eight patients, who had a mean curve of 70 degrees, had surgery as
the primary treatment at a mean age of eighteen years.
Observation Group
Eleven of the twenty-nine patients in the observation group had
a severe scoliotic curve. In nine of these eleven patients, the
initial curve could not be documented because they had been born
prior to 1953 and had first been seen at our center after the age
of twenty years. In the other two patients, the curve had progressed
from 19 to 55 degrees over ten years and from 70 to 106 degrees
over eleven years. The pulmonary status of the eleven patients with
a severe curve was believed to be so frail that any type of treatment
was contraindicated. In the eighteen patients with a mild-to-moderate
scoliotic curve, the mean curve progressed from 14 to 22 degrees
over seventeen years, for a mean progression of 0.5 degree per year.
Bracing Group
Sixty-five patients were treated with bracing at a mean age (and
standard deviation) of 12 ± 4 years (range,
three to twenty years). The brace was worn for a mean of five years
(range, six months to fourteen years). The Milwaukee brace was worn
by fifty-six patients, and a thoracolumbosacral orthosis was worn
by nine. Initially, the brace was to be worn full time, but we noted
exacerbation of gastroesophageal reflux and aspiration when the
brace was worn at night. This was probably because the patients
were in a recumbent position, they were being fed at night through
a gastrostomy tube, and their gag reflex was already compromised.
Therefore, bracing was discontinued at night. Currently, our patients
(except for those with excessive kyphosis of more than 50 degrees)
are treated with a thermoplastic thoracolumbosacral orthosis.
Fifty-eight (89 percent) of the sixty-five patients had progression
of the curve despite bracing. The thoracic curves progressed a mean
of 5 degrees per year (from 25 to 50 degrees), the lumbar curves progressed
a mean of 4 degrees per year (from 19 to 38 degrees), and kyphosis
increased a mean of 5 degrees per year (from 53 to 77 degrees).
In contrast, seven (11 percent) of the sixty-five patients who wore
a brace did not have progression of the curve. In these seven patients,
the mean thoracic curve was 21 degrees (range, 11 to 44 degrees)
at the beginning of bracing and 21 degrees (range, 9 to 49 degrees)
at the end of bracing. In this group, bracing was started at a mean
age of twelve years and was continued for a mean of four years;
this was not significantly different from the duration of bracing
in the group of patients who had progression. We found no association
between the magnitude of the curve and the change that occurred during
bracing.
Of the fifty-eight patients for whom bracing had failed, twenty-four
underwent spinal arthrodesis. The remaining thirty-four patients
were not operated on because of their precarious cardiopulmonary
status, because they had been seen for consultation only, or because
they did not want an operation.
Because the data in our study originated from the Dysautonomia
Treatment and Evaluation Center (a comprehensive care facility for
patients with this disorder) and because the sample size was larger
than it has been in previous studies, we believe that we are able
to provide an extensive analysis of the prevalence of spinal deformity
in this population. Other retrospective studies on the prevalence
of scoliosis in patients with familial dysautonomia were from orthopaedic
referral centers. In 1971, Yoslow et al.15 reported
on sixty-five patients with familial dysautonomia who ranged in
age from nine months to twenty-nine years. Thirty-nine (60 percent)
of the patients were found to have scoliosis. Five patients had
severe scoliosis measuring over 50 degrees, but none were treated
operatively. In 1987, Albanese and Bobechko1 reviewed
the cases of sixteen patients with familial dysautonomia. Nine of
the patients had spinal deformity: two had scoliosis, four had scoliosis and
kyphosis, and three had predominantly kyphosis. Seven patients were
treated operatively with spinal arthrodesis and Harrington compression
instrumentation following the failure of bracing. In 1995, Rubery
et al.14 described their experience
with twenty-two patients who had operative treatment for the correction
of spinal deformity. In that study, it was estimated that progressive
scoliosis would develop in 95 percent of patients with familial
dysautonomia and that kyphosis would develop in 53 percent.
In an attempt to define the probability of a spinal deformity
developing during growth in patients with familial dysautonomia,
we reviewed the data on all such patients who had reached the age
of twenty years or more. We found that 102 (83 percent) of the 123
patients had a spinal deformity: sixty-nine (56 percent) had scoliosis
only, thirty-one (25 percent) had scoliosis associated with kyphosis,
and two (2 percent) had kyphosis only. Forty of the 100 scoliotic
curves were convex to the left. This is in agreement with the findings
reported in the 1997 study by Kaplan et al.10,
in which 51 percent (twenty-six) of the fifty-one curves in patients
with familial dysautonomia were convex to the left.
These deformities develop early and progress rapidly during the
early years of life. In our study, scoliosis had been diagnosed
in half of the patients by the age of ten years (Fig. 1).
A limitation of our study was the lack of access to the initial
radiographs of nine untreated patients who already had severe curves
at the time of presentation to our institution. The progression
of these curves could not be assessed.
The etiology of spinal deformity in patients with familial dysautonomia
is not clear. The curves tend to be severe and progressive, resembling
deformities found in patients with other neuromuscular diseases8.
The treatment of spinal deformity in these patients is difficult
and frustrating. As the life expectancy of patients with familial
dysautonomia has increased, such treatment has become a major issue related
to their medical care. In 1970, Brunt and McKusick6 predicted that only 50 percent of
all children born with familial dysautonomia would survive beyond the
age of five years. In our study, 123 (75 percent) of 165 patients
had survived to the age of at least twenty years, with the oldest
patient being fifty-one years old.
In the present study, bracing was found to be of limited effectiveness
for the treatment of spinal deformity in patients with familial
dysautonomia. Even with bracing, most of the curves continued to progress
quite rapidly, at a rate of about 4 to 5 degrees per year. Bracing
prevented progression of the curve in only seven (11 percent) of
sixty-five patients. Most of the curves that did not progress had
been mild (mean, 21 degrees) at the onset of bracing. We found no
other variable to be predictive of which curves would stabilize.
These findings are in agreement with those of Kaplan et al.10, who reported a progression rate
of 2.8 degrees for scoliosis and 3.5 degrees for kyphosis. In their
series, all ten patients who were managed with a Milwaukee brace
had progression of the curve and three of the seven patients who
were managed with a thoracolumbosacral orthosis needed spinal arthrodesis.
It was not clear whether the curves in the other four patients had
stabilized or whether they simply had not progressed enough to necessitate surgery.
Forty of the 100 scoliotic curves in our study were left thoracic.
This is in contrast to curves in patients with adolescent idiopathic
scoliosis, which are usually right thoracic with the apex at the
seventh or eighth thoracic vertebra.
Our protocol for the treatment of spinal deformity in immature
patients with familial dysautonomia calls for the use of a custom-molded
spinal orthosis for curves of at least 20 degrees that have progressed
a total of 5 degrees or more since the initial evaluation. We believe
that a frank discussion with the family should be carried out in
order to outline the poor response to bracing; however, we continue
to recommend bracing as the last resort in an attempt to avoid surgery
and its complications in this medically fragile population. We still
believe that bracing may help to slow the progression of mild curves.
Although bracing was initially used in a small number of skeletally
mature patients because it was thought to be the only option that would
stop curve progression, this is not currently done. When a curve
progresses to 45 degrees or more in an immature patient, the patient
becomes a candidate for spinal arthrodesis with instrumentation.
In summary, the prevalence of spinal deformity in patients with
familial dysautonomia who had lived for at least twenty years was
83 percent. By the age of ten years, 52 percent of the patients
had scoliosis and 21 percent had kyphosis with or without scoliosis.
Bracing was found to be of limited effectiveness as a definitive
treatment for spinal deformity. With the increasing longevity of
these patients, aggressive operative treatment may provide the only
means of controlling severe deformities despite the high risk of
serious complications. We could not establish a means of predicting
the progression of a given curve with our data. Improved technology
and increased knowledge of the disease will facilitate our approach
to the treatment of patients with familial dysautonomia.