Clinical etiological classification of scoliosis: report of 1289 cases

Abstract

Objective:  With the development of diagnostic techniques and in‐depth understanding of lateral curvature of the spine (scoliosis), it is possible to differentiate idiopathic scoliosis from other forms with various known etiologies. The present study was to analyze data collected at the authors' center according to the current etiological spectrum and classification of scoliosis.

Methods:  One thousand, two hundred and eighty‐nine consecutive patients with different forms of structural scoliosis were reviewed. The average age at first visit was 18 years, ranging from 4 months to 79 years. Corrective surgery was performed on patients aged from 9 to 28 years; their clinical data were retrieved for independent statistical analyses, and further compared with those obtained from the whole group.

Results:  The prevalence of non‐idiopathic scoliosis was 25.3% in the whole series, but it increased to 34% in the surgical group aged from 9 to 28 years. Thirty‐nine percent of patients with congenital scoliosis presented at least one developmental spinal cord malformation.

Conclusion:  The current study has shown that the etiological distribution of scoliosis has changed a lot from what was true decades ago. Developmental malformation related to scoliosis is one of the risk factors for neurological complications during corrective surgery, so it is important to make an accurate diagnosis and take appropriate prophylactic measures to avoid relative neurological complications.

Introduction

Scoliosis, or abnormal curvature of the spine, is a common spinal deformity Scoliosis is a descriptive term rather than an etiological diagnosis. Although the most common classification is idiopathic (of unknown cause), a definite cause for scoliosis can be found in 20%–25% of scoliotic patients ¹ . An etiologic classification is convenient when it comes to discussing, diagnosing and treating scoliosis. To date, the most commonly accepted classification system has been provided by the Scoliosis Research Society (SRS), and was modified in 1973 ² . With the advancement of diagnostic techniques, especially with the assistance of magnetic resonance imaging (MRI), electrophysiology and laboratory examination, the etiological diagnosis of non‐idiopathic scoliosis has become more available and accurate ³ , ⁴ . Therefore, although the nature of scoliosis remains unchanged, the etiological spectrum or distribution of scoliosis may be amended accordingly. Recently, little attention has been paid to the clinical etiological classification, which is important in clinical treatment, especially in the establishment of the appropriate surgical strategy. This study, in which the 1289 patients who have presented with structural scoliosis since 1998 were consecutively reviewed, aimed to make a detailed clinical etiological classification of scoliosis treated at our center.

Materials and methods

Diagnostic criteria and classification methods

In the current study, the term scoliosis means lateral deviation of the spine on the coronal plane with Cobb angle larger than 15°. Adolescents with 10°–14° Cobb angle at the first examination were excluded from this series if either spontaneous alleviation or non‐progression occurred. Functional scoliosis can be corrected by active or passive ipsilateral bending or supine positioning, and has the potential to resolve completely after early correction of the primary curve. Transient structural scoliosis may be grounded on a sciatic, hysterical, or inflammatory basis. The criterion for inclusion of patients in this study was scoliosis without any functional or transient structural forms.

Patients

From August 1998 to August 2003, a total of 1289 patients with structural scoliosis were seen in the authors' center, including 358 males and 931 females with a male to female ratio of 1:2.6. Mean age at the first visit was 18 years, ranging from 4 month to 79 years.

Detailed physical examinations were performed by experienced spinal surgeons to rule out any interesting signs including skin patches, back central line structure, laxity of joints and others. Neurological examination consisted of an evaluation of motor, sensory, and reflex functions of the upper and lower extremities, as well as an evaluation of abdominal reflexes. Anteroposterior and lateral standing radiographs of the whole spine were taken at the time of presentation to identify congenital and developmental anomalies and to evaluate the severity of the deformity. Because there is a high prevalence of neural axis abnormalities in patients with infantile or juvenile‐onset scoliosis, all patients with early onset scoliosis received an MRI examination from the skull to the coccyx for neural axis abnormalities. The same MRI protocol was applied to those with positive neurological signs, developmental anomalies or a left thoracic curve. Clinical features, results of radiological, laboratory, electrophysiology examinations and biopsies were incorporated in the process of differentiating the diagnosis.

Structural scoliosis was divided into five classes according to general causes: idiopathic, neuromuscular, congenital, developmental and secondary/miscellaneous, as in Table 1, according to the classification method proposed by SRS in 1969 and modified in 1973 ² .

Table 1.

Detailed classification of non‐idiopathic scoliosis ^†

Causes	Surgical patients (9–28 years old)		All patients (4 months–79 years old)
	Number (patients)	Percentage (%)	Number (patients)	Percentage (%)
Congenital scoliosis
Osteopathic vertebral/skeletal anomaly
Failure of formation	45	5.5	46	3.5
Failure of segmentation	12	1.5	19	1.6
Miscellaneous	31	3.8	35	2.7
Neuropathic
Spinal bifida	8	1.0	8	0.6
Tethered cord syndrome	11	1.4	12	0.9
Myelomeningocele	2	0.2	3	0.2
Syringomyelia (congenital)	19	2.3	21	1.6
Chiari I malformation	16	2.0	20	1.6
Neuromuscular
Neuropathic
Upper motor neuron
Cerebral palsy	12	1.5	12	0.9
Friedreich ataxia	2	0.2	2	0.2
Dandy‐Walker syndrome	4	0.5	5	0.4
Athetosis	2	0.2	2	0.2
Lower motor neuron
Poliomyelitis	8	1.0	12	0.9
Spinal muscular dystrophy	2	0.2	2	0.2
Myopathic
Muscular dystrophy
Duchenne	2	0.2	5	0.4
Becker	3	0.4	4	0.3
Other	8	1.0	8	0.6
Developmental
Skeletal dysplasia
Osteogenesis imperfecta	4	0.5	4	0.3
Spondyloepiphyseal dysplasia	12	1.5	13	1.0
Mucopolysaccharidoses	4	0.5	4	0.3
Skeletal dysostosis
Ehler‐Danlos Syndrome	2	0.2	2	0.2
Marfan or Marfanoid Syndrome	16	2.0	17	1.3
Neurofibromatosis	21	2.6	23	1.8
Others
Tuberculosis	11	1.4	18	1.4
Postradiation spinal deformity	3	0.4	4	0.3
Ankylosing spondylitis	12	1.5	16	1.2
Degenerative scoliosis	4	0.5	9	0.7
Total	276	34.0	326	25.3

^† This classification is based on the SRS Principles.

The detailed etiologic classification for non‐idiopathic scoliosis is listed in Table 1. For patients presenting complex deformities or malformations, the following principle was observed according to SRS: if patients presented with both congenital vertebral anomalies and spinal cord developmental malformations, they were sorted into the congenital and neuropathic subgroup rather than into the neuromuscular group. For example, if patients presented with both congenital scoliosis and spina bifida, although the scoliosis may have been secondary to neuromuscular problems, they were listed in the congenital group. If multiple spinal cord malformations occurred in one patient, classification to the most important and severe one took precedence. For example, patients with simultaneous syringomyelia and occult spinal tethered cord syndrome were sorted into the syringomyelia subgroup. Calculation error resulting from a patient being documented more than once was thus eliminated by using this principle.

Results

Management

Operation

A total of 812 patients were treated operatively, among which 801 received posterior three‐dimension CD (Cotrel & Dubousset) or TSRH (Texas Scottish Rite Hospital) instrumentation. Non‐instrumentation correction surgery included spinal arthrodesis in 45 patients, occipital foramen enlargement for Chiari malformations in 20, occipital‐atlas in‐site arthrodesis in 6, cerebral spinal fluid shunt in 20, and semi‐vertebral resection in 10.

Orthotic

In total 338 patients received orthotic treatment. Different types of orthotics such as Milwaukee or Boston braces were selected according to age, skeletal maturity and curve type.

Observation

This group consisted of patients who fulfilled indications for surgery but did not receive it for various non‐medical reasons, such as economic or transportation factors.

Clinical follow‐up

1200 patients were followed up for an average of 38 months (range 12 to 48).

There were 963 patients with idiopathic scoliosis (74.7% of the whole group) and 326 with non‐idiopathic scoliosis (25.3%). The patients fulfilling surgical indications were aged between 9 and 28 years, so an independent data analysis was carried out on the surgical group. There were 812 patients in the surgical group (63% of total): 536 cases of idiopathic scoliosis and 276 cases of non‐idiopathic scoliosis, accounting for 66% and 34% of the surgical group respectively (Table 2).

Table 2.

Percentage of idiopathic and non‐idiopathic scoliosis

Scoliosis	Surgical patients (9–28 years old)		All patients (4 months–79 years old)
	Number (patients)	Percentage (%)	Number (patients)	Percentage (%)
Idiopathic	536	66.0	963	74.7
Non‐idiopathic	276	34.0	326	25.3
Total	812	100.0	1289	100.0

Only 20 patients had positive family histories, which accounted for 2.1% of all idiopathic scoliosis patients. Among these 12 patients with a familial hereditary background and four pairs of twins suffering from idiopathic scoliosis were noted.

Discussion

Etiopathogenesis of adolescent idiopathic scoliosis (AIS)

Recently, several theories have been advanced concerning the etiology or pathogenesis of adolescent idiopathic scoliosis. The most popular theories include an asymmetric growth pattern between anterior and posterior columns or concave and convex sides of the spine, central and peripheral equilibrium system disorder, soft tissue abnormalities, hereditary factors and a neuroendocrine theory ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ . The ‘neuroendocrine’ theory proposes that a decrease in serum melatonin may be an initial factor in scoliosis, and that the serum concentration correlates with scoliosis progression ¹⁰ . Though much work has been done in this area, direct evidence to prove that a decrease of melatonin can lead to scoliosis is rare. However, due to the complexity of its clinical features, there is now a consensus that adolescent idiopathic scoliosis is a result of multiple combined factors ³ . A hereditary factor is one that deserves our attention: monozygotic twins are more liable to suffer scoliosis simultaneously than are dizygotic twins ¹¹ . In our study, 20 patients had a positive family history which accounts for 2.1% of the series, 12 patients had a familial hereditary background and four pairs of twins suffered idiopathic scoliosis simultaneously. Currently, there is no effective prophylactic method to prevent the deformity, therefore it is more important to diagnose correctly and early. Children with a positive family history should be closely monitored to ensure timely and appropriate treatment.

Etiological classification of scoliosis and its differential diagnosis

SRS defines scoliosis as lateral curvature of the spine of more than 10° from the coronal plane with the Cobb measurement method on the standing radiography. Many clinicians tend to use a Cobb angle 15° and Risser sign less than 1 degree as the starting standard for orthotic treatment, while for a curvature between 10° and 14° short‐term observation is recommended ¹² . In order to make an accurate analysis of the distribution of scoliosis which truly needs treatment, only patients with curvature greater than 15° were included in our study. The percentage of all scoliosis that is idiopathic varies in the literature, reports ranging from 75%–80% ¹³ , ¹⁴ , ¹⁵ . Some authors even report that idiopathic scoliosis may account for 95.5% of all adolescent scoliosis ¹⁵ . In studies with a small sample size such variation may be due to the limitations of the numbers, or to the patients not being representative, while for studies with a large sample size it may relate to accuracy of differential diagnosis.

In our study, idiopathic scoliosis accounted for 74.7% of the whole patient pool which is similar to literature reports (75%–80%) ¹³ , ¹⁴ , ¹⁵ . But when the surgically treated patients, aged between 9–28 years, were retrieved from the whole pool for specific analysis, the percentage of non‐idiopathic scoliosis increased markedly from 25.3% to 34%. Thus 1/3 of patients were etiologically non‐idiopathic in the surgical group. Shi and Wu respectively studied 177 and 218 patients with scoliosis who received surgical intervention ¹⁶ , ¹⁷ . They found that non‐idiopathic scoliosis accounted for 30% and 34% of all their patients respectively, which is similar to our finding. Increase in the percentage of non‐idiopathic scoliosis may due to the following factors: (i) careful physical examination revealed some occult physical signs leading to a diagnosis of non‐idiopathic scoliosis. Sixteen cases presented depression of the abdominal reflex, which had previously been ignored, with no other positive neurological signs. Further MRI examination confirmed the diagnosis of Chiari malformation and syringomyelia; (ii) because these radiological features are atypical of idiopathic scoliosis, patients with left thoracic curvature or thoracic kyphoscoliosis were subjected to additional MRI investigation besides routine X‐ray and physical examination. As a result, some spinal cord malformations such as syringomyelia were found in these patients; and (iii) it is well known that the progression of idiopathic scoliosis tends to slow down when people reach adulthood, patients with remarkable curve progression after skeletal maturity or patient with abnormal arthrochalasis, hepatomegaly and splenomegaly deserve additional attention, for these patients may have neurological, metabolic or interstitial abnormalities.

Importance of recognizing non‐idiopathic scoliosis

Surgical correction of scoliosis involves potential neurological risk, and some severe neurological complications may be inevitable if certain neurological developmental malformations are ignored and left untreated. In patients with Chiari malformations, spina bifida and filum terminale tethered cord syndrome, the spinal cord is under tensile force and vulnerable to injury during surgical procedures. To ensure the safety of surgery, decrease the incidence of neurological complications and achieve satisfactory spinal reconstruction, certain first‐stage preparatory surgeries are necessary. These surgeries include grand occipital foramen enlargement for Chiari malformations, cerebral fluid shunt for syringomyelia, neurolysis in tethered syndrome etc. In our study, 58 patients received such preparatory surgeries before second‐stage correctional surgery. For patients with minor neurological malformations for which no neurosurgery intervention is needed, halo‐femur traction before spinal correction is recommended (35 cases in current study). If no neurological changes are found during continuous traction till optimal correction of deformity, corrective surgery is safe. Decompensation, or loss of correction on both sagittal and coronal planes, is often seen in neurological scoliosis post‐operatively. Carefully considered selection of fusion level and second stage anterior strut grafting in these patients can effectively reduce the incidence of such long‐term complications. Massive and uncontrollable bleeding during surgery can occur in myogenic scoliosis or Ether‐Danlos diseases, so prophylactic work is mandatory for these patients. As to Marfan syndrome or Friedreich ataxia, adequate acquaintance with such disease plays a very important role in evaluating the prognosis and surgical risk.