Charcot-Marie-Tooth disease is the most common inherited sensorineuropathy,
affecting one in 5000
individuals1 each
being identified by its genetic etiology. There are several forms of the
disease, with2. The
gene for the most common form—type 1, or the hypertrophic demyelinating
form—has been localized to an area on chromosome 22 that encodes for
peripheral myelin protein
223. Inheritance is
autosomal dominant, although the phenotypic penetrance of the condition varies
among family members.
Foot deformity, specifically cavovarus deformity, is the most common
indication for orthopaedic intervention in these patients, but other
musculoskeletal problems can arise. Hip dysplasia and scoliosis occur more
commonly in patients with Charcot-Marie-Tooth disease than they do in the
general population. The prevalence of scoliosis in patients with
Charcot-Marie-Tooth disease has been estimated to be
38%4. Specific
genetic mutations seen in subtypes of Charcot-Marie-Tooth disease have been
linked with a higher prevalence of
scoliosis5,6.
While a few authors have reported the prevalence of scoliosis in patients with
this condition, little is known about the response of these curves to
treatment4,7,8.
The purpose of this study was to report the prevalence of scoliosis in a
large clinical population of children and adolescents with various subtypes of
Charcot-Marie-Tooth disease, to characterize the spinal deformities, and to
assess the effect of orthotic and surgical treatment of curves that required
such treatment.
Amedical record search for all patients with a diagnosis of
Charcot-Marie-Tooth disease followed at a large neuromuscular clinic from 1981
until 2002 was performed. Each chart was reviewed to confirm the diagnosis,
which was made by a neurologist. While earlier cases generally were diagnosed
on the basis of an abnormal nerve-conduction velocity or results of a sural
nerve biopsy, more recently seen patients underwent molecular genetic testing
for known chromosomal markers of the disease. The subtype of
Charcot-Marie-Tooth disease was noted when available.
The medical record was reviewed for a history of scoliosis. When scoliosis
was noted by the examining physician, radiographs were made to confirm the
presence of a scoliotic curve, but radiographs were not routinely made for
patients in whom a spinal deformity had not been observed clinically.
Scoliosis was measured with the Cobb method, and kyphosis was quantified in
patients with lateral radiographs. The location and direction of the curve
pattern were noted. We evaluated radiographs made at the time of the initial
diagnosis, at the time of orthotic prescription, immediately preoperatively,
immediately postoperatively, and at the time of final follow-up.
Additional information gathered from the charts of the patients with
scoliosis included the age when the scoliosis was diagnosed, brace use, and
the need for spinal fusion surgery. The type of brace was noted, and, when
surgery had been performed, the age at the time of the surgery, type of
surgery, number of levels fused, intraoperative complications, and need for a
reoperation were recorded. The results of intraoperative neurologic monitoring
were noted as well. Finally, whether the patients were able to walk, and if
they remained able to do so postoperatively, were evaluated.
We identified 298 patients with Charcot-Marie-Tooth disease who had been
seen at the neuromuscular clinics at our hospital. Forty-five patients had
scoliosis, a prevalence of 15% (see Appendix). Eighteen patients with
scoliosis were female, and twenty-seven were male. The age at the time of the
diagnosis of the scoliosis averaged 12.9 years, with a range of eight to 17.8
years. The duration of follow-up averaged 3.6 years (range, zero to 9.8
years). Of the twenty-nine patients for whom the subtype of
Charcot-Marie-Tooth disease was known, thirteen had type 1; nine, type 1A;
five, type 2; one, type 2B; and one, type 3. The subtype of
Charcot-Marie-Tooth disease was not noted for sixteen patients.
Twenty-nine patients had thoracic scoliosis, and fifteen of them had a left
thoracic curve with the apex of the deformity ranging from T5 to T10. Three
patients had a double major curve, and in all three the thoracic curve was
right-sided. Seven patients had a thoracolumbar curve. Only four patients had
a primary lumbar curve.
The average curve magnitude at the time of the diagnosis of the scoliosis
was 27.6°, with a range of 10° to 60°. At the time of diagnosis,
fourteeen curves (31%) were =20°, seventeen (38%) were 21° to
30°, and fourteen (31%) were >30°. Many of the curves were
hyperkyphotic in the thoracic area. Nineteen (49%) of the thirty-nine patients
in whom it was measured had a kyphosis of =45°, and the average
thoracic kyphosis was 48.0° (range, 12° to 97°).
Curve progression was analyzed only when the curve had been followed for
more than one year and the final Risser sign was =3. (One curve that had a
final Risser sign of 0 was included because it had progressed.) Six patients
presented after they had reached skeletal maturity and had radiographic
follow-up of less than one year, one patient underwent surgery at the time of
the initial presentation, and four patients had a Risser sign of 0, 1, or 2 at
the time of the last radiographic follow-up and were excluded. Twenty-four
(71%) of the thirty-four curves that had been followed for at least one year,
to maturity, or to surgery progressed >5°, whereas ten curves did
not.
Curve progression was seen with equal frequency in female and male
patients. Progression was documented in 69% of the female patients and 71% of
the male patients who were followed radiographically for more than one year
and until the Risser sign was at least 3.
The curve magnitude at the time of presentation correlated with the
likelihood of progression. Only four of the nine curves that measured 20°
at presentation and nine of the thirteen curves that measured 21° to
30° progressed, whereas eleven of the twelve curves that initially
measured >30° progressed.
Three patients were unable to walk as a result of the disease at the time
of the diagnosis of the scoliosis, and one had a limited ability to walk about
the house. All four of these patients had substantial curve progression,
averaging 28.8°, and all eventually underwent spinal fusion (Figs.
1-A,
1-B,
1-C,
1-D, 1-E). Three of these
patients had type-1 Charcot-Marie-Tooth disease, and one had type-2
disease.
An orthosis was prescribed for the management of the scoliosis in eighteen
patients (40%). The average curve magnitude at the initiation of brace
treatment was 36.2° (range, 23° to 47°). Fourteen patients used a
Boston brace, one used a TLSO brace, and three used a Charleston nighttime
orthosis, but two of the patients were lost to follow-up shortly after bracing
began. The outcome was known for sixteen patients treated with bracing.
Six patients had a curve of =40° when the bracing was prescribed;
five of them had curve progression of >10° that later required surgical
treatment and one (a thirteen-year-old boy) had a curve of 42°, which did
not progress. Ten patients had a curve of <40° when bracing was begun.
Six of those curves progressed >10° while still undergoing brace
treatment and were treated, or scheduled to be treated, with surgery. Four
patients who wore a brace for a curve of <40° did not have surgery, and
two of those curves did not progress >5°. Therefore, only three of the
sixteen curves that were treated with bracing did not progress, and only five
of the sixteen did not undergo, or were not scheduled for, surgery. The curve
magnitude at the time of the initial use of the brace averaged 33.2°
(range, 25° to 42°) for the patients who did not require surgery and
38.3° (range, 23° to 47°) for those who did have surgery.
Fifteen patients (33%) were scheduled for spinal fusion, with one of them
dying of unrelated causes before it could be performed. The average age at the
time of surgery was 13.6 years (range, 11.5 to 15.75 years). Nine (33%) of the
twenty-seven male patients and six (33%) of the eighteen female patients had
(or were scheduled for) surgery. Of the fourteen surgical procedures, twelve
consisted of posterior spinal fusion alone, one was a posterior spinal fusion
preceded by preoperative halo traction, and one was a combined anterior and
posterior spinal fusion. An average of 13.1 spinal segments (range, ten to
fifteen levels) were fused. Only three patients had a fusion ending cephalad
to L3, three fusions were to L3, seven were to L4, and one was to L5.
Posterior instrumentation was used in all fusions. One patient underwent
posterior spinal fusion with Harrington instrumentation and sublaminar wires,
seven had hook/rod instrumentation, two had hook/rod constructs with
additional sublaminar wires, three had a combination of hooks and lumbar
pedicle screws, and one had hooks, pedicle screws, and sublaminar wire
fixation. The average curve magnitude at the time of surgery was 64.3°,
ranging from 50° to 82°. The immediate postoperative curve magnitude
averaged 33.1° (range, 17° to 67°), representing a 47.6%
correction. The curve magnitude at the time of final follow-up averaged
38.4° (range, 25° to 60°). Eleven of the fourteen patients who had
a spinal fusion had thoracic kyphosis measuring >45° (average, 58°;
range, 16° to 97°). The duration of postoperative follow-up averaged
3.9 years (range, 2.3 to 5.7 years).
There were no intraoperative or perioperative neurologic complications. One
of the fourteen patients underwent a reoperation to remove instrumentation
because of a delayed infection that developed 1.7 years after posterior spinal
fusion, and one patient underwent partial removal of the implant because of
prominence secondary to proximal junctional kyphosis. There were no documented
pseudarthroses or failures of instrumentation. All four patients who had been
unable to walk preoperatively continued to use a wheelchair postoperatively.
One patient who was able to walk preoperatively was unable to do so one year
following surgery. No immediate postoperative neurologic changes were noted in
this patient, and when she was discharged from the hospital her ambulatory
status was the same as it had been preoperatively; however, progressive
lower-extremity weakness developed over the first postoperative year. An
extensive neurologic evaluation did not reveal a cause other than the
progressive neuropathy attributed to Charcot-Marie-Tooth disease.
Intraoperative neurologic monitoring was attempted in twelve patients
undergoing spinal fusion, but somatosensory evoked potentials could not be
obtained in nine of those twelve patients. Only three patients had sufficient
responses to permit intraoperative monitoring. In these three, decreased
amplitudes of the somatosensory evoked potential responses were seen
throughout the operation. Normal amplitudes in adolescents vary but usually
range between 0.3 and 1.0 mV. The three patients who had responses had
amplitudes as low as 0.09 mV. Two of the three patients also had prolonged
latencies of up to 60 msec (normal latency in adolescents, 25 to 30 msec). One
patient also had monitoring of motor evoked potentials with spinous process
electrodes, and that patient had amplitudes of between 0.54 and 0.66 mV
(normal, >1.0 mV) and latencies of 24 msec (normal, 15 to 20 milliseconds).
Intraoperative changes in evoked potentials did not occur in the three
patients who were monitored throughout the surgical procedure. We did not
monitor cortical motor evoked potentials or perform pedicle screw
electromyographic stimulation, as it was not our practice to do so. Four
patients who could not be monitored had a wake-up test, which confirmed intact
lower-limb movement in all of them. The surgeons treating the remaining
patients who could not be monitored chose to not perform a wake-up test, in
many cases because the patient had extensive preoperative lower-extremity
weakness.
Scoliosis in patients with Charcot-Marie-Tooth disease has been studied
before. Hensinger and MacEwen described seven children with scoliosis in a
population of sixty-nine patients with Charcot-Marie-Tooth disease (a rate of
scoliosis of
10.1%)7. Walker et
al. reported that, in a population of eighty-nine patients with
Charcot-Marie-Tooth disease pooled from three Shriners hospitals, thirty-four
patients had either scoliosis or kyphoscoliosis (a prevalence of
38%)4. The
prevalence of scoliosis of 15.1% (forty-five patients) in our study of 298
patients with Charcot-Marie-Tooth disease seen at a neuromuscular clinic falls
between the prevalences reported in the above two studies. Only patients in
whom spinal deformity is suspected on the basis of a physical examination of
the spine undergo radiographic evaluation in our institution, so there may be
more patients with minor curves whom we did not identify.
There was a predominance of male patients in this series (representing 60%
of the group with scoliosis), which differs from the prevalence of idiopathic
scoliosis. Additionally, the radiographic appearance of the majority of the
curves was distinctly different from that seen in patients with idiopathic
scoliosis. First, there was a 33% prevalence of left thoracic curves, which
are atypical for patients with idiopathic scoliosis and which have been linked
with posterior cerebral fossa and spinal canal abnormalities such as Chiari
malformations and
syringomyelia9-11.
The presence of a left-sided thoracic curve frequently prompts the acquisition
of a spinal magnetic resonance imaging scan to rule out such
pathology12.
Second, 16% of our patients had a thoracolumbar curve. Finally, almost half of
the curves were hyperkyphotic, a finding that is clearly different from the
typical apical thoracic lordosis seen in adolescent idiopathic scoliosis.
Hyperkyphosis is also seen more often with scoliosis in patients with
syringomyelia13,14.
While our series included six ambulatory girls who had a typical right
thoracic curve, an atypical appearance of the curve should remind the
attending physician to once again carefully perform a detailed neurologic
examination.
Certain factors appear to be correlated with an increased prevalence of
curve progression in patients with Charcot-Marie-Tooth disease. First, the
patients who were unable to walk all had substantial curve progression and
underwent spinal fusion. Second, scoliosis that was associated with
hyperkyphosis was more likely to progress than was scoliosis with normal
sagittal plane contours or thoracic lordosis. A curve magnitude of >30°
was associated with curve progression, and only one curve measuring
>30° did not progress =6°. Conversely, eleven of twelve curves
of >30° at presentation progressed. Only two curves that initially
measured <30° progressed enough to require surgery.
It should be noted that eleven of the forty-five patients with scoliosis in
our study were excluded from the analysis of progression. Six patients were
excluded because they had been followed for less than one year, but all were
almost mature (a Risser sign of 4 or 5) when their last radiograph had been
made, and their curves ranged from 13° to 30°. One patient presented
with a 60° curve and was scheduled to have surgery at the time of
presentation. Four patients were excluded because the Risser sign was <3 at
the time of their final radiograph, and three of the four had a curve
measuring <20° at that time. Two patients were lost to follow-up after
the start of bracing for the scoliosis. To summarize, the 71% prevalence of
curve progression may be falsely high as a result of the exclusion of many
patients who had a small curve and either were seen only once at skeletal
maturity or were not followed because the curve was very small.
To our knowledge, this is the largest study evaluating the results of
orthotic treatment of scoliosis in patients with Charcot-Marie-Tooth disease.
Of the sixteen curves that were treated with bracing and not lost to
follow-up, only three did not progress >5°. Additionally, eleven of the
sixteen patients went on to require spinal fusion. It is clear that the
response of scoliosis to bracing in a patient with Charcot-Marie-Tooth disease
differs from that in patients with idiopathic scoliosis.
The lack of success of the brace treatment of the scoliosis in our patients
with Charcot-Marie-Tooth disease cannot be attributed solely to the magnitude
of the deformity when the brace was prescribed. In a previous report of the
outcomes of bracing of idiopathic scoliotic curves of between 35° and
45° at our center, Katz and Durrani found that only 31% progressed and
required surgery15,
compared with eight of ten curves of similar magnitude at the time of
initiation of bracing in the present study. Additionally, Katz and Durrani
found that 61% of the idiopathic curves were successfully managed with an
orthosis without substantial progression, compared with only one of the ten
curves of similar magnitude in our series. As the same orthotic department
treated the patients in both studies and the time periods during which the
patients were assessed overlap, the differences were most likely due to the
neurologic nature of scoliosis in Charcot-Marie-Tooth disease. Compliance with
brace wear was not quantified in this study. Many of these patients also used
lower-extremity orthoses to assist gait. Whether patients with
Charcot-Marie-Tooth disease will tolerate the addition of a spinal orthosis
remains an unanswered question.
A second question that we sought to answer was whether the risks of spinal
fusion surgery in patients with Charcot-Marie-Tooth disease are different from
those in patients with idiopathic scoliosis. A study by Daher et al. included
four patients treated with spinal fusion, and pseudarthrosis developed in two
of them8. None of
the fourteen surgically treated patients in our series had either
postoperative symptoms or radiographic evidence of pseudarthrosis of the
fusion. While it is impossible to be certain of successful fusion without
reexploration, we currently have no evidence of an increased risk of
pseudarthrosis in patients with Charcot-Marie-Tooth disease who undergo spinal
fusion for the treatment of scoliosis. Perhaps the use of modern principles of
instrumentation has improved the surgical outcomes in these patients.
With the use of spinal instrumentation to achieve correction comes the
increased risk of neurologic injury during spinal fusion surgery. Krishna et
al. reported on five patients with Charcot-Marie-Tooth disease in whom
somatosensory evoked potential monitoring was impossible to perform during
spinal surgery16.
Intraoperative neurologic monitoring has become standard practice in our
institution, but we found that it is rarely possible to monitor either
somatosensory or motor evoked potentials during surgery in patients with
Charcot-Marie-Tooth disease because of the demyelinating polyneuropathy. Only
three of our fourteen surgically treated patients had a sufficient
somatosensory evoked potential response to allow intraoperative monitoring,
and even in those patients the electrical potentials were of decreased
magnitude throughout the procedure. Because of the difficulty with monitoring,
one should seek a consultation with the evoked-potential-monitoring team and
the anesthesiologist before performing spine surgery in patients with
Charcot-Marie-Tooth disease. It is worthwhile to attempt monitoring, but when
it is unsuccessful at the initiation of the case, preparations for an
intraoperative wake-up test should be made when possible. However, as a result
of the neurologic involvement of the lower extremities in Charcot-Marie-Tooth
disease, and the lower-extremity weakness, patients may be unable to move the
ankles or feet preoperatively and an intraoperative wake-up test may not be
possible.
While intraoperative monitoring was successful in only a quarter of our
patients, none had a postoperative neurologic deficit despite an average 48%
curve correction. Neither Hensinger and
MacEwen7 nor Daher
et al.8 described
neurologic injury following spinal surgery in their patients. It might appear
that the neurologic risk is not measurably increased in this patient group,
but the number of surgically treated patients was too small to prove this.
Surgery should be approached cautiously because of the inability to use
neuromonitoring in most of these patients.
In conclusion, scoliosis was found in approximately 15% of our patients
with Charcot-Marie-Tooth disease. The curve progressed in 71% of the patients
who were followed with serial radiographs, and although bracing was attempted
in 40% of the patients it rarely prevented either curve progression or the
need for surgery. Surgical instrumentation and fusion was necessary in 33% of
the patients with scoliosis, and it was not associated with increased
intraoperative risks or postoperative complications. Intraoperative monitoring
is rarely possible in patients with Charcot-Marie-Tooth disease; therefore,
preoperative planning regarding the intraoperative assessment of potential
neurologic compromise with a wake-up test should be considered when
possible.