Childs Nerv Syst (2006) 22: 1154–1157

DOI 10.1007/s00381-006-0090-y ORIGINA L PA PER

Ranjeev Bhangoo
Spyros Sgouros
Scoliosis in children with Chiari I-related
syringomyelia

Received: 25 October 2005

Published online: 16 March 2006
# Springer-Verlag 2006
R. Bhangoo . S. Sgouros (*)
Department of Paediatric Neurosurgery,
Birmingham Children’s Hospital,
Steelhouse Lane,
Birmingham, B4 6NH, UK
e-mail: S.Sgouros@bham.ac.uk
Tel.: +44-121-3338075
Fax: +44-121-3338081
Abstract Objective: To study the
relationship between scoliosis and
Chiari I malformation, with reference
to the possible role of cranio-vertebral
decompression in preventing the
need for scoliosis correction.
Material and methods: Out of a
total of 36 patients with symptomatic
Chiari I, who underwent primary
cranio-vertebral decompression by a
single paediatric neurosurgeon with
an interest in Chiari malformation
between 1998 and 2003, 13 had
clinically detected scoliosis. Of these,
ten had no other structural spine
abnormality, which could influence
the natural history of scoliosis and
were included in this study. Results:
In all but one patient, syringomyelia
improved significantly after cranio-
vertebral decompression. Of the ten
patients, eight had levoscoliosis (left
convexity), all single curves, and two
had curves to the right (both double
curves). Six patients did not require
corrective scoliosis surgery after
cranio-vertebral decompression. The
mean Cobb angle for those not
requiring scoliosis correction was 29°
in contrast to a mean of 76° for those
requiring correction [p=0.001, one-
way analysis of variance (ANOVA)].
The mean age of patients requiring
corrective surgery was 158 months
against 125 months for those not
requiring correction (p=0.084, one-
way ANOVA). These findings were
confirmed by multivariate analysis,
which also confirmed that symptom
duration, syrinx length and site were
not significant in predicting the need
for corrective surgery following
cranio-vertebral decompression.
Conclusions: Cranio-vertebral de-
compression for Chiari I may prevent
the need for corrective scoliosis sur-
gery when performed before the age
of ten and below a Cobb angle of 30°.
Keywords Chiari malformation .
Hindbrain hernia . Syringomyelia .
Scoliosis . Cranio-vertebral
decompression

Introduction Chiari I, seen in up to 50 to 75% of such patients [2–8].

Numerous abnormalities of the spine and spinal cord have
been associated with the Chiari I malformation. In the
vertebral column, these include the Klippel–Feil deformity,
atlanto-axial assimilation, retroflexion of the odontoid
process and thickening of the ligamentum flavum [1]. The
abnormality that has exercised perhaps most discussion
amongst neurosurgeons and paediatricians is the associa-
tion of scoliosis in patients with syringomyelia and
Most surgeons who intend to surgically correct scoliosis
prefer that the cranio-vertebral junction is decompressed
first to avoid intra- and post-operative acute neurological
deterioration from acute impaction of the cerebellar tonsils
in the cranio-vertebral junction. However, there remains
considerable debate concerning the value of cranio-verte-
bral decompression in preventing further deterioration or
progression of already established scoliosis. In addition,
while most neurosurgeons regard scoliosis as a symptom of

the hindbrain-related syringomyelia, there is a debate
whether scoliosis only in the absence of neurological
deficit constitutes an indication for cranio-vertebral
decompression.
This study aims to identify the role of cranio-vertebral
decompression in preventing the need for scoliosis correc-
tion in Chiari I malformation.
Materials and methods
Data were collected prospectively on all Chiari I patients
referred to a single paediatric neurosurgeon at Birmingham
Children’s Hospital, U.K. (S.S.), with an interest in Chiari
malformation between 1998 and 2003. All patients with a
symptomatic Chiari I malformation, syringomyelia and
scoliosis with no other bony abnormality of the spinal
column that could influence the natural history of scoliosis
and a minimum follow up of 6 months were included in the
study. Data collected included the age and sex of the
patients at presentation, duration of symptoms, length of
syrinx (measured in length of vertebral segments), age at
which cranio-vertebral decompression was performed
(“age at operation”), the Cobb angle of the scoliosis at
the time of presentation, progression of the curve and the
need for corrective surgery and/or bracing to the scoliotic
curve.
While data collection was prospective, data analysis was
retrospective. Statistical comparison of Cobb angle and age
at operation was performed using one-way analysis of
variance (ANOVA), comparing mean values between those
who required spinal corrective surgery after cranio-verte-
bral decompression and those who did not.
Results
Of the 36 patients who underwent primary cranio-vertebral
decompression for hindbrain hernia (Chiari I malforma-
tion) during the study period by the senior author, a total of
20 patients had Chiari I malformation and syringomyelia.
All operations were performed with a standard technique
involving a 3×3 cm occipital craniectomy including the
foramen magnum, opening of the arachnoid and dissection
of adhesions at the outlets of the fourth ventricle and
leaving the dura of the posterior fossa widely open creating
a pseudomeningocele as described by Williams [9], but
without performing tonsillectomy. Of these, a total of 13
Table 1 Comparison of Cobb
angle and age for patients
with syringomyelia and scolio-
sis according to the need for
scoliosis correction Cobb angle at presentation
Age at cranio-vertebral operation (months) 158
1155
patients had a clinically detected scoliosis and 10 of these
patients had no other bony abnormality of the spine, which
could influence the evolution of scoliosis and thus were
eligible for this study. Some patients presented with
scoliosis to spinal surgery unit and were referred for
neurosurgical assessment. In others, the scoliosis, in the
presence of syringomyelia, was regarded as an indication
for cranio-vertebral decompression even in the absence of
neurological deficit. It should be noted that at the same time
period, the senior author has been managing in a non-
operative fashion other patients with hindbrain hernia and
syringomyelia without neurological deficit or scoliosis.
The mean age of the patients at operation (cranio-vertebral
decompression) was 138 months (range: 106–191 months),
mean post-surgical follow up was 33.6 months (range:
6–65 months). Of these ten patients, only four have to date
been required corrective surgery for their scoliotic curves
after successful cranio-vertebral decompression.
The mean Cobb angle of the curves at presentation was
48° with a range of 20 to 100° and a median of 36°; eight
patients had a levoscoliosis (a curve with the convexity to
the left), all single curves. The two remaining patients had
double curves. The length of the syringes at presentation
varied from two to 20 segments with a mean of 9.1
segments and a median value of 6.5 segments. In nine of
the 10 patients, the syringomyelia improved significantly
after cranio-vertebral decompression. The one patient who
did not experience improvement of the syringomyelia
cavity developed recurrent arachnoiditis in the cranio-
vertebral junction as a result of infection.
One-way ANOVA, looking at the impact of cranio-
vertebral decompression on the need to perform curve
correction surgery, showed that those patients who required
curve correction surgery after successful cranio-vertebral
decompression had presented with a mean Cobb angle of
76° and that those who did not require surgery to correct
their curves had presented with mean Cobb angle of 29°.
This finding was statistically significant with a p value of
0.001. The only other variable that approached significance
was the age of the patient at the time of cranio-vertebral
decompression surgery. Those who required scoliosis
corrective surgery had mean age of 158 months at the
time of surgery (13.1 years of age) and those who did not
require surgery had mean a age of 125 months (10.4 years
of age), with a p value of 0.084 (see Table 1). Multivariate
analysis confirmed that symptom duration, syrinx length
and site were not significant in predicting the need for
corrective scoliosis surgery.
Curve correction Curve correction Significance
required not required (one-way ANOVA)
76° 29° p=0.001
125 p=0.084
1156

Discussion correction surgery was most likely to be avoided if

Our study is one of the few prospective studies to address
the issue of the relationship between the Chiari I malfor-
mation, syringomyelia and scoliosis. As is expected in
patients with Chiari I, the scoliotic curves seen were all
associated with syringomyelia, and the majority of the
curves (eight out of 10) were single and had their convexity
to the left, the so-called levoscoliosis [1]. Retrospective
studies have suggested that not only scoliosis can be a
presenting feature of Chiari I but that surgical intervention
can prevent further deterioration of the scoliotic curve and
indeed prevent the need for corrective scoliotic surgery
[2–8]. Others have felt that cranio-vertebral decompression
does not provide an effective intervention with regard to
halting the progression of scoliotic curves [10] and
certainly the likelihood of avoiding surgery for scoliosis
curve correction in adult Chiari I patients remains
poor [11].
In an attempt to identify those patients who may avoid
curve correction surgery after successful cranio-vertebral
decompression, some studies have specifically looked for
particular clinical features associated with such patients.
Ozerdemoglu et al. [3] stated that in patients without
myelomenigocele or congenital scoliosis but with Chiari
malformation and syringomyelia, suboccipital craniectomy
gave the best chance for syrinx reduction and scoliosis
improvement, particularly in children younger than
10 years. This finding is remarkably concordant with the
findings of our prospective study where there was a result
approaching statistical significance with regard to age.
Those patients undergoing surgery to the cranio-vertebral
junction around the age of 125 months did not require
curve correction surgery and those around 158 months of
age required scoliosis correction. Eule et al. [4] identified
an age of 8 years as the key cutoff, while Sengupta et al., as
a result of their experience, identified a cutoff age of
10 years [6]. In support of this concept of age as a key
discriminating factor in identifying those scoliotic patients
who may benefit most from cranio-vertebral decompres-
sion, a study of scoliotic curves in adult Chiari I patients
concluded that the surgical outcomes were poor [11].
The other feature identified as discriminating those
patients who may avoid curve correction surgery after
cranio-vertebral decompression has been the Cobb angle of
the scoliotic curve. Eule et al. [4] stated that curve
cranio-vertebral decompression was performed before the
“curve was severe”. Ghanem et al. noted that scoliosis
progressed, despite suboccipital foraminotomy, in those
cases with a severe preoperative deformity, thus requiring
deformity surgery [8].
Nagib [12] more precisely defined a Cobb angle of 30°
as identifying those patients who were most likely to avoid
curve correction surgery. Our study showed that the Cobb
angle was the only significant factor that allowed identi-
fication of those patients who would avoid deformity
surgery as a result of cranio-vertebral decompression.
Those patients with curves around 29° at the time of
surgery did not require curve correction subsequently,
while those who had a Cobb angle of around 76° required
scoliosis surgery at a later date.
We would suggest therefore that cranio-vertebral de-
compression, when performed below a Cobb angle of 30°
and before the age of 10 years, may prevent the need for
corrective scoliosis surgery. We guard this conclusion by
acknowledging that our study though prospective, like
many of the other studies in this field, is as a result of the
relative rarity of the defining condition small. Another
compromise of this study is that follow-up of these patients
has not extended to the age of complete skeletal maturity,
therefore, it cannot be regarded as conclusive on the topic,
as it is possible that more patients in the group that did not
require scoliosis surgery may require such surgery later on
in life.
Conclusion
On the basis of this prospective study looking at the role of
cranio-vertebral decompression surgery in Chiari I patients
with syringomyelia and scoliosis, we would propose that
hindbrain decompression performed before the age of
10 years and a Cobb angle of 30° may prevent the need for
scoliosis correction surgery. Larger series with follow-up
well past skeletal maturity would provide the definitive
answer on the topic.
Acknowledgement This paper was presented at the XIX Biennial
Congress of the European Society for Pediatric Neurosurgery, Rome,
6–9 May 2004.

References
1. Oakes WJ, Tubbs RS (2004) Chiari
malformations. In: Winn HR (ed)
Youman’s neurological surgery, 5th
edn. Saunders, Philadelphia, Pennsyl-
vania, pp 3347–3361
2. Ozerdemoglu RA, Denis F, Transfeldt
EE (2003) Scoliosis associated with
syringomyelia: clinical and radiological
correlation. Spine 28(13):1410–1417
3. Ozerdemoglu RA, Transfeldt EE, Denis
F (2003) Value of treating primary
causes of syrinx in scoliosis associated
with syringomyelia. Spine 28(8):
808–814

4. Eule JM, Erickson MA, O’Brien MF,
Handler M (2002) Chiari I malforma-
tion associated with syringomyelia and
scoliosis: a twenty-year review of sur-
gical and non-surgical treatment in a
paediatric population. Spine 27
(13):1451–1455
5. Loder RT, Stasikelis P, Farley FA
(2002) Sagittal profiles of the spine in
scoliosis associated with an Arnold–
Chiari I malformation with or without
syringomyelia. J Pediatr Orthop 22
(4):483–491
10. Farley FA, Puryear A, Hall JM,
1157
6. Sengupta DK, Dorgan J, Findlay GF 11. Ono A, Ueyama K, Okada A, Echigoya
(2000) Can hindbrain decompression N, Yokoyama T, Harata S (2002) Adult
for syringomyelia lead to regression of scoliosis in syringomyelia associated
scoliosis? Eur Spine J 9(3):198–201 with Chiari I malformation. Spine 27
7. Hanieh A, Sutherland A, Foster B, (2):E23–E28
Cundy P (2000) Syringomyelia in 12. Nagib MG (1994) An approach to
children with primary scoliosis. Childs symptomatic children (ages 4–14 years)
Nerv Syst 16(4):200–202 with Chiari type I malformation.
8. Ghanem IB, Londono C, Delalande O, Pediatr Neurosurg 21:31–35
Dubousset JF (1997) Chiari I malfor-
mation associated with syringomyelia
and scoliosis. Spine 22(12):1313–1317
9. Williams B (1993) Surgery for hind-
brain related syringomyelia. In: Symon
L (ed) Adv Tech Stand Neurosurg
20:108–164
Murasko K (2002) Curve progression
in scoliosis associated with Chiari I
malformation following subocciptal
decompression. J Spinal Disord Tech
15(5):410–414