01-batra et al

CURRENT CONCEPTS REVIEWActa Orthop. Belg., 2008, 74, 147-160

Congenital scoliosis remains an interesting and chal-lenging diagnostic entity. Vertebral absence, partialformation or lack of segmentation may cause asym-metrical growth and resultant deformity. Because ofthe high frequency of associated anomalies withinand outside the spine, a detailed history and physicalexamination are mandatory. Maternal, perinatal his-tory, family history, and developmental milestonesmust be fully explored.Plain radiographs remain standard for diagnosis ofcongenital anomalies and measuring curve magni-tude, progression and perhaps growth potential ofthe vertebral anomaly. Preoperative CT scan definesthe anatomy and avoids any unexpected intraopera-tive posterior element deficiencies. MRI can excludeassociated conditions of the spine, cranio-cervicaljunction and viscera. The recognition of curves witha bad prognosis at an early stage is pertinent toprevent curve progression and possible neurologicalcomplications.The goal of surgery is to achieve a straight spine anda physiological sagittal profile while maintainingflexibility, to arrest progression of the curve witha short fusion segment preserving as much normalspinal growth as possible. Developments in gene research continue to bepromising and may potentially lead to early detectionof congenital vertebral malformations.

Keyword : congenital scoliosis.

BACKGROUND

Congenital scoliosis (CS) is defined as a lateralcurvature of the spine due to a developmental

abnormality. Scoliosis present at birth that is notassociated with an underlying developmentalanomaly is referred to as infantile scoliosis. Theterm CS implies that there is always spinal curva-ture at birth, but only the vertebral malformationsare present at birth in all patients, while the scolio-sis for some may not develop until later.

An incidence of approximately 0.5 to1/1,000 births has been observed for CS (37). Withan isolated single vertebral malformation, thechance of a first degree relative having a similaranomaly was estimated to be approximately oneout of a hundred in one study (45), but no pattern ofinheritance was found in another report (45). If mul-tiple vertebral anomalies are present, then the riskof similar anomalies in either siblings or childrenof the patient is between 10 to 15% (45).

Genetic inheritance has been shown responsiblefor some congenital vertebral anomalies ; however,there is no clear-cut genetic aetiology of CS to

No benefits or funds were received in support of this study Acta Orthopædica Belgica, Vol. 74 - 2 - 2008

Congenital scoliosis : Management and future directions

Sameer BATRA, Sashin AHUJA

From Llandough Hospital and University Hospital of Wales, Cardiff, United Kingdom

■ Sameer Batra, MRCS, MS (Orth), Spinal Fellow.■ Sashin Ahuja, FRCS (Orth), Consultant Spinal Surgeon.Division of spine surgery, Department of Orthopaedics,

Llandough hospital and University hospital of Wales, Cardiffand Vale NHS Trust, Cardiff, United Kingdom.

Correspondence : Mr Sameer Batra, Specialist Registrar,Department of Trauma and Orthopaedics, Royal Gwent hospi-tal, Newport, NP20 2UB, United Kingdom.E-mail : [email protected]

© 2008, Acta Orthopædica Belgica.

148 S. BATRA, S. AHUJA

date (15). Basic science research evidence in micesuggests that maternal exposure to toxins, such ascarbon monoxide exposure, may cause congenitalscoliosis (16). Associations with maternal diabetesand ingestion of antiepileptic drugs during preg-nancy have also been postulated as possible caus-es (16).

The vertebral column develops from pairs ofsomites, which begin to appear at 3 weeks gesta-tion. The somites are mesenchymal segments,which are on both sides of the neural tube, and theantero-medial wall of the somite is termed a scle-rotome. Cells from the sclerotomes spread out cen-trally to form an unsegmented, cellular perichordalsheath, which eventually forms the centrum of thevertebral column. Next a zone of loose cells in thesclerotome forms superiorly where the interseg-mental and spinal nerve pass through the plate anda dense zone forms inferiorly which goes on tobecome the posterior neural arch of the vertebraand the rib (fig 1). One sclerotome pair forms onelevel of ribs and posterior elements of the spine. Inthe cellular sheath of the notochord, alternatingzones of loose and dense zones also appear, but thesuperior zone of loose cells goes on to form thecentrum of the vertebra, while the inferior densezone goes on to form the intervertebral disk.

Various theories to explain congenital vertebralmalformations have included failure of ossificationas the cause of defects of vertebra formation,osseous metaplasia of the annulus fibrosus as thecause of defects of vertebral segmentation, andvertebral development hindered by persistentnotochord (5).

CLASSIFICATION

Moe et al (35) classified congenital scoliosisaccording to morphologic characteristics on plainAP frontal and lateral images as an embryologicaldefect into formation failure, segmentation failure,and a mixed type. This classification also includesseveral other factors, such as the level of formationfailure and the presence or absence of interverte-bral disc space. Based on this information, the nat-ural history of congenital scoliosis and strategicapproaches to its treatment have been evaluatedand reported (35).

The failures of formation (fig 2) are charac-terised by the deficiency of a portion of a vertebralelement, causing 1) hemivertebra (complete) or2) wedge vertebra (partial).

Wedge vertebrae present with a height asym-metry, with one side being hypoplastic, there are

Acta Orthopædica Belgica, Vol. 74 - 2 - 2008

Fig. 1. — Vertebral centrum development from two adjacent segments.

CONGENITAL SCOLIOSIS 149

two pedicles. Next in severity is a hemivertebra,with the absence of one pedicle and a region of thevertebral body. Hemivertebrae may be furtherclassified on the basis of the presence or theabsence of fusion to the vertebral bodies aboveand/or below (29). An unsegmented hemivertebra isfused to the vertebral body above and below ; apartially segmented hemivertebra is fused to thevertebral body either above or below ; and a fullysegmented hemivertebra is separated from the bodyabove and below by disk space. Hemivertebraemay occur at ipsilateral adjacent levels of the spine,which produces significantly asymmetrical spinegrowth, or one hemivertebra may be counterbal-anced by another hemivertebra on the contralateralside of the spine in the same region, separatedby one or several healthy vertebrae (termed ashemimetameric shift) (38,41). This anomaly occursmost commonly in the thoracic spine.

The defects of segmentation are characterised byabnormal bony connections between vertebrae(fig 3). These bony connections may be bilateraland symmetrical, resulting in a block vertebra.Segmentation defects caused by unilateral bonyfusions are termed bars and may act as a unilateralgrowth tether. Occasionally, a segmentation defectmay span an ipsilateral formation defect, resultingin a unilateral bar and a contralateral hemiverte-bra (24).

The third category or mixed type comprises thecomplex anomalies that include both segmentationand formation errors and at first may be difficult todefine, as the spine is only 30% ossified at birth.

Hemivertebrae usually occur as extra spinalsegments and are often accompanied by an extrarib. They may result from an abnormal cleavage ofthe primary chondrification center. Wedge vertebraedo not present as extra segments or ribs and thusmay involve a unilateral failure of development ofthe chondrification center (31).

Anomalies tend to occur at the apex of a curve. Acurve can be upper thoracic (33% of curves), lowerthoracic (31%), thoracolumbar (20%), lumbar(11%), or lumbosacral (5%) (40). In general, thora-columbar curves tend to have the worst prognosisand the greatest progression, followed by lower tho-racic curves and then upper thoracic curves.

The ribs are formed in close association with thevertebrae, and it is, therefore, not surprising to havea combination of developmental abnormalitiesaffecting both the ribs and the vertebrae. An exten-sive thoracic congenital scoliosis due to mixedvertebral anomalies associated with fused ribs mayaffect thoracic function and the growth of the lungsin young children and lead to a thoracic insuffi-ciency syndrome. An imbalance in the mechanicalthrust of the ribs may also adversely affect spinalgrowth as well as the function of trunk muscles andthe pressure within the thorax.


Fig. 2. — Representation of formation failures : (A) wedgevertebra (B) Fully segmented hemivertebra (C) Partially seg-mented hemivertebra (D) Unsegmented hemivertebra.

Fig. 3. — Representation of segmentation failures : (A) blockvertebrae (B) Partially unsegmented bar. (C) combination ofhemivertebra and unsegmented bar.


However, since only the vertebral body and thepedicle of the vertebral arch in the malformed ver-tebrae can be evaluated by this classification, anevaluation of the morphology of the posterior com-ponents has not been performed. Furthermore, it isdifficult to evaluate the morphology of vertebralbodies in severe kyphoscoliosis, or in cases ofsevere complex malformed vertebrae, using plainAP images alone.

NATURAL HISTORY

Reports in the 1950s by Kuhns and Hormell (12)

estimated that congenital scoliosis had a generallybenign course : only 37% of the children in theirseries had curves greater than 30° at maturity. Thisconclusion was refuted by the independent naturalhistory studies by Winter et al (44) and McMasterand Ohtsuka (26) who reviewed a combined total of485 patients, most until skeletal maturity. They re-ported that the rate of deterioration and the severityof final deformity were predictable according to thetype of anomaly and curve location. Upper thoraciccurves tend to be less severe than thoracolumbarcurves, which tend to be most severe. Curve pro-gression occurs more rapidly during the first 5 yearsof life and, again, during the adolescent growthperiod of puberty ; these two periods represent themost rapid stages of spine growth (12).

These authors found that nearly 75% of patientsstudied required treatment, and that 84% ofpatients who were untreated developed curvesgreater than 40° at maturity. For most anomalies,the deterioration of the curve was steady and unre-lenting, starting at birth, accelerating during thepreadolescent growth spurt, and slowing or stop-ping at skeletal maturity. Children with clinicaldeformities in the first year of life had the worstprognosis and had severe early progression.

Curve progression is caused by unbalancedgrowth of one side of the spine relative to the other.Spine growth comes from the superior and inferiorend plates of each vertebral body. Thus by assess-ing the quality of the disks surrounding the anom-alous segment of the spine, growth potential maybe inferred. Radiographically definable disks signi-fy the presence of vertebral growth plates and,

when asymmetrical or more present on one side ofthe spine than on the other, they have the potentialfor asymmetrical growth in that area of the spine.Thus, fully segmented hemivertebrae with healthy,definable disks above and below have much morepotential to cause a curvature compared withunsegmented hemivertebrae, which are fused to thevertebrae above and below (33). The presence ofeither a bar or fused ribs is a good predictor ofcurve progression. Either can act as a tether, andthe combination tends to produce even more rapidcurve progression.

With recognition of the ‘at-risk’ types of verte-bral malformations, patients with congenital scol-iosis can now be classified according to their risk ofcurve progression and the frequency of follow-upadjusted accordingly. Some deformities, such as aunilateral unsegmented bar, were recommended fortreatment at time of recognition because of theirinvariable tendency for severe progression. Thethoracic curves have the poorest prognosis, withthe worst anomaly being a unilateral unsegmentedbar combined with single or multiple convexhemivertebrae, followed by a unilateral unsegment-ed bar, double convex hemivertebrae, and a singleconvex hemivertebra, with a block vertebra havingthe best prognosis (20).

A unilateral unsegmented bar adjacent to a con-tralateral hemivertebra with an average curve pro-gression in excess of 10 degrees per year in the tho-racolumbar region is the most aggressive anomaly.It may not be radiographically visible until it ossi-fies when the patient is 3 to 4 years of age. Ahemimetameric shift does not always present witha balanced spine, and produces occasionally a pro-gressive deformity especially in thoracolumbar andlumbosacral junctions (20).

Block vertebrae are uncommon and tend to be themost benign anomaly producing curves of less than20°. Complex curves consisting of multiple anom-alies occur in 10% of cases (20). Both conditionsshould be observed until progression is documented.

Mixed deformities are unpredictable, and theirseverity depends on the amount of unbalancedgrowth potential. The combination of mixed defor-mities and rib anomalies can lead to a global loss oftrunk and thoracic height and width, resulting in



severe volume and motion restriction leading torespiratory failure.

ASSOCIATED SURFACE DEFORMITY

A cosmetic deformity occurs with most curves.In upper thoracic curves, elevation of the shoulderon the convexity of the curve, with tilting of thehead into the concavity, may be seen. Structuralcongenital curves tend to have only a mild rota-tional component ; thus, they produce only a mildrib hump. Unbalanced curves in the lower thoracicregion and the lumbar region produce a pelvicobliquity with apparent shortening of the leg on theconcave side of the curve. The trunk tends to listaway from the apex of the curve, and this can causedifficulty in ambulation and balance.

ASSOCIATED ANOMALIES

Another important aspect of congenital scoliosisis the recognition of the presence of concurrentabnormalities of the spinal cord, the kidneys, andthe heart, which may be present in 61% of patientsand in an even higher percentage in patients withmixed defects (2). There is some evidence in the lit-erature to support the development of organ mal-formations at the body segment at which vertebralmalformations occur. The developmental field con-cept described by Opitz (32) is useful in providing aframework for understanding multisystem involve-ment and a defect in one system should promptevaluation of others.

Neural axis abnormalities are present in up to35% of patients, as detected with MRI, patientswith mixed or segmentation defects being at high-est risk (5). These abnormalities include (but are notlimited to) diastematomyelia (split cord), cord teth-ering, Chiari malformations, and intradural lipo-mas. The absence of cutaneous signs of dysraphismand the absence of neurological deficit do not ruleout an intraspinal dysraphism. Diastematomyeliais found in up to 20% of patients with congenitalscoliosis and should be treated by resection beforeproceeding with corrective surgery of the spinebecause of the risk of stretching the spinal cordtethered at the site of the anomaly (46). Otherlesions of the cord seen in congenital scoliosis

include dermoid cysts, epidermoid cysts, ter-atomas, and lipomas as well as tethered spinal cordand they may all need treatment prior to spinalsurgery to minimise risk of subsequent neuralinjury. With such a high incidence of associatedintraspinal anomalies, routine MRI screening ofpatients with congenital scoliosis has been suggest-ed (4). The specific indications for MR imaging forpossible intraspinal anomalies in congenital scolio-sis include : the presence of neurological defectssuch as weakness, sensory loss, bowel or bladderdysfunction ; an associated skin abnormality overthe spine such as a dimple, hairy patches, or nevi ;complaints of back or leg pain ; patients with lum-brosacral kyphosis ; radiographic evidence of inter-pedicular widening, diastematomyelia or presenceof a unilateral congenital bar with a contralateralhemivertebra ; and any patient who is to undergospinal stabilisation surgery (4).

Unrecognised abnormalities of the renal systemmay be present in 25 to 33% of patients and includeunilateral horseshoe kidney, renal agenesis,duplicated kidney/ureters and hypospadias (13). Allpatients with congenital scoliosis warrant a renalevaluation by either intravenous pyelogram or renalultrasound/MRI.

Congenital heart disease is present in 10% ofpatients, ranging from atrial and ventricular septaldefects, which are the most common abnormalities,to complex congenital heart defects, such astetralogy of Fallot and transposition of the greatvessels. A screening echocardiogram and cardio-logic evaluation is needed for patients under-going surgical correction for a congenital spinedeformity.

The development of restricted pulmonary func-tion is a concern. Recent studies imply that curvemagnitude directly affects pulmonary function, butthe development of restrictive lung function occursonly as the curve approaches 90° (30). Severecurves in young children tend to produce the mostsevere pulmonary compromise. This restrictionmay be due more to hypoplastic lung developmentthan to a physical constriction from the curve ;alveolae continue to develop in the growing childup to 8 years of age. Vital capacity screening is rec-ommended for patients with severe curves.



Musculoskeletal anomalies occur frequently inassociation with congenital spine anomalies.Disorders, such as clubfeet, Sprengel deformity,Klippel-Feil deformity, developmental dysplasiaof the hip, and upper and lower limb deformitieswarrant a detailed evaluation and managementprior to spinal surgery.

Vertebral malformations have been shown to beassociated with hemifacial microsomia, Alagille,Jarcho-Levin, Klippel-Feil, Goldenhar, Joubert,and VACTERL syndromes as well as basal cellnaevus, trisomy 18, and diabetic embryopathy (3).

CLINICAL AND RADIOGRAPHICASSESSMENT

Clinical Evaluation

Clinical evaluation begins with a comprehensivehistory and physical examination. A prenatal histo-ry of the mother, including all health problems, pre-vious pregnancies, and medications, is recorded.Birth history of the child should include detailssuch as length of gestation, type of delivery (vagi-nal or caesarean), birth weight, and complications.Given that the presence of cognitive delay has beenshown to correlate with curve progression in somepatients, particular attention should be paid towhether the child has appropriately reached devel-opmental milestones (10).

The physical examination should start with theheight and the weight of the patient, given thatgrowth plays a significant role in curve progres-sion. The skin must be examined for abnormalitiessuch as “café au lait” spots or axillary freckles asseen in neurofibromatosis and midline patches ofhair for any evidence of spinal dysraphism. Spinaldysraphism may also manifest itself in the lowerextremities, and signs would include asymmetricalcalves, cavus feet, clubfeet, vertical tali, and abnor-mal neurological findings. The spinal examinationitself focuses on any evidence of truncal or pelvicimbalance. Rib cage deformities, chest or flankasymmetry, chest excursion and anomalies need tobe evaluated, as does the inspiratory and expiratorycapacity of the chest wall. Limitation in chestexcursion may indicate a syndromic scoliosis and

thoracic insufficiency syndrome (10). In young chil-dren, the Adams forward bending test (looking forprominence of the ribs in the thoracic spine ortransverse processes in the lumbar spine) is notpossible, but the test can be simulated by laying thechild in a prone position over the knee of the exam-iner. Curve flexibility can be assessed by placingthe child in a lateral position over the knee of theexaminer or by suspending the infant over the armof the examiner.

Spinal balance in both the coronal and the sagit-tal planes must be evaluated. Truncal imbalance,head tilt, shoulder inequality, and pelvic balanceshould be addressed. Photographs serve as animportant aid in sequential evaluation of progress.Given the association of neural axis abnormalitiesand the possibility of neurological compromise incongenital spine deformities, a detailed motor, sen-sory and reflex examination, (including abdominalreflexes) is mandatory. Muhonen et al (29)

described the absence of an abdominal reflex as theonly objective finding seen in some patients with aChiari malformation. When the results of reflextesting are abnormal, the reflex is usually absent onthe convex side of the curve.

The vital capacity of patients with congenitalscoliosis has been compared with those with idio-pathic scoliosis. For any given Cobb angle the lossin vital capacity was approximately 15% greater incongenital compared with idiopathic scoliosis. Thisgreater impairment in lung function in congenitalscoliosis has been proposed to be due to the associ-ated rib deformity or to an underlying lung anom-aly (30). Vital capacity screening has been recom-mended for patients with severe curves. A fullspirometry work-up is recommended if surgery isplanned for those with a vital capacity < 60% ofnormal (30).

RADIOGRAPHIC ASSESSMENT

Plain radiographs remain standard for diagnosisof congenital anomalies and measuring curve mag-nitude, progression and perhaps growth potential ofthe vertebral anomaly. Conventional radiographyoften is difficult to interpret because of the patient’ssmall size, the complex nature of the deformity,



superimposed structures obscuring visualisation ofthe anomaly, and spine deformity in a plane otherthan that of the radiograph.

Preoperative CT scan helps define the anatomyand avoid any unexpected intraoperative posteriorelement deficiencies (fig 4). The combination of 3-D images and 2-D coronal and sagittal multiplanarreformatted CT images curved to match the con-tour of the spine can be extremely beneficial in theassessment of hemivertebrae for which it is diffi-cult to determine the degree of segmentation or thepresence of a contralateral unsegmented bar byplain radiography or 2-D CT imaging. We obtainCT with 3-dimensional reconstructions for preop-erative assessment and evaluation of complexdeformities, but not for routine observation or seri-al documentation.

Computed tomography scan helps in evaluationof chest wall deformity and lung volume in con-genital deformities with chest wall anomalies,chest deformity, or thoracic insufficiency. The useof 3-D CT data to define lung volumes in patientswho were too young for pulmonary function testshas been described and subsequently used to mea-sure improvement in lung function followingexpansion thoracoplasty (18).

Magnetic resonance imaging (MRI) has had arevolutionary impact on imaging the neural axisand its role has been widely studied and document-ed in patients with congenital scoliosis and spinaldysraphism (4). The prevalence rate of spinal dys-raphism detected using MRI approaches 30% inpatients with congenital spine deformities. Imagingof the brainstem to the bottom of the sacrum isrequired to exclude associated conditions of thespine, the cranio-cervical junction and the viscera.In our institution, we perform coronal T1, sagittalT1 and T2 weighted scans through the whole spinewith axial gradient echo images through the cervi-cal spine and T2 weighted images through the apexof the curve and through the lumbar spine. Thecoronal T1 weighted images are useful for demon-strating if the cord is single and for assessing theshape of the scoliosis.

MANAGEMENT

The skill in managing a patient with a congeni-tal spine deformity lies not only in the ability toperform major complex salvage surgery in patientspresenting at a late stage with a severe rigid defor-mity, but in recognising those curves with a badprognosis at an early stage to prevent curve pro-gression and possible neurological complications.Meticulous management planning requires anastute knowledge of the natural history of all typesof congenital spine deformity and the methods oftreatment that are available. Once high-risk anom-alies such as unilateral unsegmented bars with con-tralateral hemivertebrae are recognised, treatmentis initiated regardless of age to prevent deformity.

NON-SURGICAL TREATMENT

Congenital vertebral anomalies can be recog-nised fortuitously on radiographs of infants beforeclinical deformity develops. Both these andpatients diagnosed later in life should be followedcarefully at 4 to 6 month intervals until the end ofgrowth. The goals of observation are to recognisethe presence of vertebral anomalies, which requireimmediate treatment, and to detect the progressionof curves.


Fig. 4. — Preoperative three-dimensional reconstructed com-puted tomography scan defines the anatomy of the curve andmay aid in the positioning of the implants.


Bracing

Bracing plays a very limited role in the manage-ment of congenital scoliosis, especially as a prima-ry form of treatment because these curves are typi-cally inflexible.Less than 10% of congenital curvescan be treated by bracing. Bracing is sometimesused to prevent progression of secondary curvesthat develop above or below the congenital curveand are causing balance problems. The Milwaukeebrace is the brace of choice for upper curves.A TLSO (thoracic/lumbar/sacral orthosis) brace isadequate for lumbar curves. The optimum durationof daily brace wear is still under investigation.As long as the curve remains controlled, treatmentshould continue until skeletal maturity. Thebrace may not arrest or correct development of thecurve, but it may slow progression and maintainflexibility, allowing surgery at a later age. Bracingserves no function in short, sharp, rigid curves.

SURGICAL TREATMENT

Principles of surgery

As a guiding principle, it has been widely recog-nised in studies that it is easier to prevent a defor-mity than allow it to progress and attempt correc-tion at a later stage (17). The goal of surgery is toachieve a straight spine, a physiological sagittalprofile while maintaining flexibility, to arrest pro-gression of the curve and as short a fusion segmentas possible, preserving as much normal spinalgrowth as possible. The risk of neurological injurywith surgical treatment of patients with congenitalspinal deformities is greater than in patients withidiopathic spinal deformity (34). The occurrence ofa perioperative neurological deficit may be min-imised in multiple ways by :

(1) A routine use of MRI evaluation of the spinalcord.

(2) Earlier treatment of a spinal cord anomaly.(3) Preventing lengthening the spinal cord intraop-

eratively by avoiding distraction and usingshortening procedures.

(4) Carefully monitored use of controlled hypoten-sion to minimise blood loss to prevent theoccurrence of cord ischaemia, especially dur-ing any corrective manoeuvres.

(5) Motor- and sensory-evoked potential monitor-ing, supplemented by a wake-up test in caseswith intraoperative changes in neurologicalmonitoring that do not return to baseline.

(6) The postoperative monitoring of a patient’sneurological status, given that paraplegia afterdeformity surgery may present in a delayedfashion, especially in the first 72 hours (28).

(7) Early and aggressive treatment of deformitiesbefore they become severe.

The procedures are broadly divided into (a)those preventing further deformity and (b) thosethat correct the present deformity ; with the lattertype further subdivided into techniques that correctthe curve gradually and those that correct the curveacutely.

Prevention of Future Deformity

In Situ Fusion

In situ fusion is a safe technique and a goodchoice for many progressive curves with minimaldeformity involving a relatively short section of thespine.

It is ideally done in patients with unilateralfailure of segmentation such as a unilateralunsegmented bar and should be accomplished veryearly, before a significant curve develops, whichwould then make a more extensive correction nec-essary.

A patient with no trunk imbalance with a curveless than 40 degrees not requiring substantial cor-rection is best fused in situ without instrumentationwith some modest correction by postoperative cast-ing (17). One of the major controversies surround-ing in situ fusion is whether a combined anteriorand posterior fusion is required (24). There is lesspotential for anterior growth in a spine with con-genital scoliosis because the growth plates may notbe properly formed. In an in situ fusion for con-genital scoliosis, a shorter segment is fused andthere is less rotation in the segment, which makes a



true crankshaft phenomenon (continued anteriorgrowth in presence of posterior fusion) unlikely.

McMaster (25) and Winter and Moe (43) recom-mended an early prophylactic in situ fusion to pre-vent the unbalanced growth of the spine in a unilat-eral unsegmented bar, with or without contralateralhemivertebrae with best results reported before theage of two years. A convex growth arrest procedureor simply dividing the unsegmented bar will notcorrect this type of deformity because there is nogrowth potential in the unsegmented bar on theconcavity of the curve. This usually necessitates acombined anterior and posterior fusion to controlthe severe spinal growth imbalance in thesepatients. The argument that an early spine fusionwill stunt the growth of the spine in these youngpatients is of no relevance because the unsegment-ed bar is not contributing to vertical height and onlymaking the spine more crooked.

The decision regarding which curves may befused in situ can be complex. The location of thecurve and likelihood of progression and the impactof the curve on the balance and cosmetic appear-ance of the child must be taken into account.Deformities at the lumbosacral and cervicothoracicjunction are more likely to lead to cosmetically dis-figuring deformities and therefore should be fusedin situ very early or alternate treatments considered.

Convex hemiepiphysiodesis

It is indicated for a unilateral formation failurei.e. hemivertebra. Because there must be adequategrowth potential on the concave side, this proce-dure is contraindicated in segmentation defects,such as bars that have no potential for concavegrowth. The epiphysiodesis is done by removingthe convex lateral half of disks adjacent to thehemivertebra, with no exposure of the spine on theconcave side. Winter et al (42) suggest that this pro-cedure should be reserved for patients younger than5 years of age, with a progressive curve of< 70 degrees involving 5 segments or less, and pre-senting with a pure scoliosis not involving the cer-vical spine with no major kyphosis or lordosis. Intheir series, 38% of patients had an average correc-tion of 10 degrees, 54% had a stabilisation effectwithout significant correction, and the only compli-

cation was curve progression in one patient. Morepredictable alternatives to convex hemiepiphysio-desis for a short segment deformity includehemivertebra excision or wedge resection, andgrowth-oriented procedures, such as VEPTR(Vertical expandable prosthetic titanium rib) orgrowing rods, when a longer segment of spine isinvolved.

Hemivertebra excision

A hemivertebra causing marked truncal imbal-ance and progression of the curve can be managedwith wedge resection in patients whom an isolatedin situ fusion or convex epiphysiodesis would notresult in a balanced spine. It is best performed atapproximately 2 years of age when the child canstill tolerate a cast and fusion is more likely. Itoffers nearly complete correction over a shortfusion segment and may be performed via anterior-posterior surgery or posterior-only surgery (19). Toclose the gap after resection of the hemivertebra,compression instrumentation is necessary andallows for strong fixation into adjacent vertebrae.Ruf and Harms (35) previously argued that hemiver-tebra resection via a posterior-only approach withtranspedicular instrumentation was the ideal proce-dure for correction in young children. They opinedthat this methodology was not only less invasivebut also prevented the development of secondarycurves and severe local deformities.

Suk et al (39) also support the use of a posterior-only approach, arguing that the combined approachrequires prolonged operative time, and that whenused in the treatment of lumbosacral deformity, itposes a risk to the anterior vascular and visceralstructures. Consequently, based on their retrospec-tive review of treatment of fixed lumbosacral defor-mity with posterior vertebral-column resection,they maintained that this technique offers theadvantages of single-stage surgery, the ability toaddress the deformity at the apex, and of controlledshortening across the resection gap.

Bollini et al (5) on the other hand advocate acombined anterior and posterior fusion because ofgreater correction capabilities, including correctionin the sagittal plane secondary to disc excision.This approach gives better visualisation of the



anatomy including the spinal column and nerveroots, as well as affording the surgeon the opportu-nity to apply a corrective force to the spine from ananterior position while simultaneously applyingcompressive instrumentation. It also decreases thelikelihood of pseudoarthrosis and prevents thedevelopment of crankshaft phenomenon by theremoval of the growth plates (6). The procedure isbest performed in the thoracolumbar junction orlower because the spinal cord in the thoracic spineis not as amenable to manipulation, but this viewhas been recently refuted (11).

Correction of the present deformity

Gradual correction is obtained through the useof a hemiepiphysiodesis and/or hemiarthrodesis ofthe spine. Immediate or acute correction may beaccomplished through several different techniques :excision or decancellisation of the hemivertebraand definitive or staged instrumentation using agrowing rod. For patients with thoracic insufficien-cy syndrome, expansion thoracoplasty and stabili-sation via the VEPTR addresses this issue by ribdistraction on the concave side of the curve. Forfixed deformities, it may be necessary to performan osteotomy to balance the spine and correct thecosmetic deformity.

Gradual correction

Hemiepiphysiodesis and Hemiarthrodesis

These techniques are used for failure of forma-tion, relying on the future growth of the spine onthe concave side. In segmentation failure, there isno potential for growth to occur. The procedure isunlikely to correct curves with low concave growthpotential such as those due to a convex hemiverte-bra opposed by a concave unsegmented bar (41).Dubousset (14) has contributed much to growtharrest procedures of the spine by pointing out therelationship between sagittal and coronal growthdisturbance of the spine with congenital deformityand how mechanical instability must be consideredin the treatment. The goal is to arrest the growth onthe convex side of the curve ; the concave side ofthe spine is not exposed surgically, as doing socould lead to spontaneous fusion. Much of the cor-rection associated with hemiepiphysidesis is

achieved acutely at the time of the initial proce-dure, thus postoperatively,a corrective cast is usedto encourage fusion of the spine in a somewhat cor-rected position, which will therefore improve thestarting point and permit the ultimate correction torely less on the growth of the spine The total cor-rection obtained by performing a convex hemiepi-physiodesis varies because the younger the child atthe time of the operation, the more potential thatexists for correction over time.

Growing Rods

Single and Dual Growing Rods

The growth of the spine is greatest during thefirst 5 years of life as sitting height reaches twothirds of the adult level by age 5, while thoracicvolume reaches 30% of the adult size by five yearsof age. Congenital scoliosis is associated with shortstature and diminished trunk height. Long fusionsperformed on younger children may have a furtherdeleterious effect on trunk height and thoracic vol-ume, leading to thoracic insufficiency. In theabsence of congenital rib fusions, early progressivedeformities may best be managed using a growingrod technique, pioneered by Paul Harrington (43) inthe 1960’s. If the child is still very young, the pri-mary congenital curve can be treated with fusion insitu, hemiepiphysiodesis and/or hemiarthrodesis,excision, or osteotomy, and the corrected curve canthen be treated with a growing rod until the child isolder. This avoids fusing the entire curve, whichwill lead to growth retardation and potentiallyharmful pulmonary effects.

The goals of treatment have been to achieve cor-rection of the spinal deformity without extensivefusion, maintain correction during the subsequentgrowth period, allow spinal growth and lung devel-opment, and avoid or eliminate the need for defini-tive fusion of the spine at an early age.

Akbarnia and McCarthy (1) reported their seriesof a dual growing rod construct due to problemsassociated with a single-rod implant, includinghook dislodgement and rod breakage. This is per-formed by subperiosteal dissection of the anchorsites proximally and distally and by placement ofclaw constructs. Rods are then placed subcuta-neously on each side and joined with tandem con-



nectors placed at the thoracolumbar junction,where the lengthening may occur. An apical fusioninvolving the congenital anomaly may be needed ;this will obtain control of the apex of the curve and,if a hemiepiphysiodesis and/or hemiarthrodesis hasbeen performed, will possibly permit improvementof the apical curvature with growth.

However, apical fusion is a must if the apicalcurvature is due to failure of segmentation. Theirpatients showed an average Cobb angle improve-ment from 82 to 38 degrees at follow-up and anaverage growth of the T1-S1 segment of 1.2 cm peryear.

Acute Correction Techniques

Correction and Fusion with Instrumentation

The partial or complete correction of a deformitydepends on the congenital anomaly itself and thedegree of deformity. The choice of fusion levels isdifficult when there are multiple noncontiguousanomalies but the aim should be to achieve a bal-anced spine with the safest possible correction. Thewhole curve needs to be fused, but there may beabnormal segments cephalad or caudad to the curvethat will not be included or two curves may need tobe included in the instrumentation. High thoracicand cervical curves need special precaution asovercorrection of a lower thoracic or thoracolum-bar curve may lead to shoulder imbalance or neckobliquity. Congenital anomalies associated withrelatively normal segmentation, flexibility (asrevealed by radiographs), and less severe truncaldeformity may be managed by means of a standardposterior arthrodesis and instrumentation. Thosecases with well-defined disk spaces seen on imag-ing imply significant remaining growth ; thesepatients may be at risk of a crankshaft phenomenonand should have in addition anterior surgery in theform of either an open or a thoracoscopic anteriorrelease and fusion. The amount of correctionobtained by instrumentation in congenital scoliosisis usually much less than that obtained in idiopath-ic scoliosis and one should not attempt to gain dra-matic correction of large, stiff congenital curves.Complications of treatment have declined as expe-rience has increased with the most common prob-lem being progression of the curve. Occasionally,

intraspinal abnormalities must be dealt with priorto the instrumentation procedure. Computedtomography with three-dimensional reconstructionis very helpful for the assessment of local anatomyand as an aid in pre-operative planning.

The safe use of traction in congenital spinedeformities, particularly in cases involvingpreexisting neurological deficits, was questioned inpast reports by MacEwen et al (22) due to the riskof traction-induced paraplegia in congenital defor-mities. Recently, the use of traction has been popu-larised for severe deformities, including congenitaldeformities (34).Rinella et al (34) reported the use ofHalo gravity traction which can be applied in bedor while in a wheelchair or walker device allowingfor gradual curve correction before or following anassociated anterior release.

Hemivertebra excision

Hemivertebra excision may be an ideal proce-dure for long standing curves with significantdeformity or truncal imbalance. The ideal indica-tion is a hemivertebra at the lumbosacral junction,as this particular deformity often leads to majorimbalance or very serious compensatory curves.

It is possible to combine this procedure with aconvex hemiepiphysiodesis and/or hemiarthrode-sis, with use of a nonfusion rod or a definitiveinstrumentation of a longer curve. The excision ofa hemivertebra may be performed by using com-bined anterior and posterior procedures. Posterior-only hemivertebra excision in growing children hasrecently been reported with successful results. Rufand Harms (35) reported their results on hemiverte-bra excision using a posterior-only approach andsegmental transpedicular instrumentation. Theyreported excellent results in patients younger than6 years, with an average Cobb measurement of45 degrees. At 3.5 years follow-up, the Cobb mea-surement had been maintained at 14 degrees, withno patient having a neurological complication.

Reconstructive osteotomy

Pelvic obliquity, severe truncal decompensation,progressive deformity, and developing neurologicaldeficit are indications for reconstructive osteo-tomies. They are considered for fixed curves with



unacceptable cosmesis and cannot be successfullymanaged with procedures described so far. Anosteotomy may be part of a combined approachthat involves resection of the hemivertebra andinstrumentation and fusion of a more extendedcurve. A cosmetically displeasing but well bal-anced curve is probably best treated with stabilisa-tion alone.

EXPANSION THORACOPLASTYAND VEPTR

Congenital spine deformities with rib fusionsmay be associated with thoracic insufficiency, aterm coined by Robert Campbell et al (8) to explain

poor thoracic and lung parenchymal developmentwith inability to support normal lung function andgrowth. The tethering effect of congenital ribfusions add to the scoliotic effect of the spine toproduce a concave, constricted hemithorax, whichis diminished in height and function.

The growth of the lung, the respiratory tree, andthe alveoli cell multiplication are greatest in thefirst 8 years of life. Fifty percent of the thoracicvolume is attained at 10 years of age ; thus, earlyfusions may have an even more profound effect onthoracic development than on spine height. Havingborderline pulmonary capacity at the end of growthcan be a serious risk factor for developing respira-tory insufficiency in late adulthood becauseapproximately 400 cc’s of vital capacity is lost inthe normal aging process and even further lossesmay occur from obstructive respiratory disease orinfection. The possible effect of early spinal fusionon thoracic development may compound the preex-isting thoracic insufficiency associated with con-genital scoliosis and fused ribs (8).

In children with congenital scoliosis and severecurve progression early surgical curve stabilisationwas based on the belief that ‘a short straight spineis always preferable to a long crooked one’ (21).Therefore, early posterior or combined anterior andposterior spinal fusion, convex anterior epiphy-seodesis or arthrodesis has been performed in thepast (23). The total growth inhibition effect on thethoracic spine due to these procedures was thoughtto be negligible because of the assumption ofdiminished growth potential of the concaveside (21). A recent study (7) challenged this assump-tion by showing an increase in length of the unilat-eral unsegmented bars and equal increases inlength of the concave and convex sides of the tho-racic spine in children with congenital scoliosis,unilateral fused ribs and unsegmented bars treatedby VEPTR. The surgical concept of the expansionthoracoplasty and stabilisation with the VEPTRimplant is based on the expansion of the thorax byrib distraction on the concave side of the curveachieving indirect correction of the curve (fig 5).Spinal fusion may be necessary once skeletal matu-rity is reached and thoracic spinal growth is nolonger an issue.


Fig. 5. — Postoperative radiograph of a four-year-old boy witha volume-depletion deformity resulting from fused ribs andcongenital scoliosis. The VEPTR stabilisation expansionthoracoplasty lengthened and widened the constrictedhemithorax, corrected the scoliosis indirectly without growthinhibition, restored truncal balance, and corrected pelvicobliquity.


DIRECTIONS FOR FUTURE WORK

Congenital scoliosis remains an interesting andchallenging diagnostic entity. Developments ingene research continue to be promising and maypotentially lead to early detection of congenitalvertebral malformations. It is likely that differentapproaches will enable an elucidation of the genet-ic and environmental basis of CS. Since CS is dueto localised alterations in vertebral body develop-ment as opposed to a more widespread distributionof vertebral malformations, a model needs to bedeveloped that incorporates position identity andfailure of segmentation. The association of renal,cardiac, skeletal and spinal cord malformationswith CS may reflect the involvement of differentgenes that are associated with developmental path-ways in several organs. Epidemiologic studies canprovide assistance in determining whether variousenvironmental factors contribute to CS.

John Cobb (9) wrote in 1958 that “In the futurestudy of scoliosis it will be necessary to keep oureye on the patient and not on the curve.” Importantissues to be addressed in the future about thoracicinsufficiency syndrome and VEPTR include char-acterisation of the natural history of the thoracicinsufficiency syndrome by anatomic classification,correlating progressive three-dimensional thoracicdeformity with the onset of respiratory insufficien-cy and whether the procedure can halt or reverseprogressive three-dimensional thoracic deformityafter the onset of respiratory insufficiency.Furthermore, continued advances in surgical tech-nique and instrumentation will provide surgeonswith an array of safe and efficacious modalities forthese patients with complex conditions.

REFERENCES

1. Akbarnia B, Marks DS, Boachie-Adjei O et al. Dualgrowing rod technique for the treatment of progressiveearly-onset scoliosis : a multicenter study. Spine 2005 ;30 : S46-S57.

2. Basu PS, Elsebaie H, Noordeen MH. Congenital spinaldeformity : a comprehensive assessment at presentation.Spine 2002 ; 27 : 2255-2259.

3. Beals R, Rolfe B. Current concepts review VATER associ-ation : a unifying concept of multiple anomalies. J BoneJoint Surg 1989 ; 71-A : 948-950.

4. Belmont PJ Jr, Kuklo TR, Taylor KF et al. Intraspinalanomalies associated with isolated congenital hemiverte-bra : the role of routine magnetic resonance imaging.J Bone Joint Surg 2004 ; 86-A : 1704-1710.

5. Blake NS, Lynch A, Dowling F. Spinal cord abnormalitiesin congenital scoliosis. Ann Radiolol 1986 ; 29 : 237-241.

6. Bollini G, Docquier P, Viehewger E et al. Lumbosacralhemivertebrae resection by combined approach. Spine2006 ; 31 : 1232-1239.

7. Campbell RM, Hell-Vocke AK. Growth of the thoracicspine in congenital scoliosis after expansion thoracoplasty.J Bone Joint Surg 2003 ; 85-A : 409-420.

8. Campbell RM Jr, Smith MD, Mayes TC et al. The char-acteristics of thoracic insufficiency syndrome associatedwith fused ribs and congenital scoliosis. J Bone Joint Surg2003 ; 85-A : 399-408.

9. Cobb JR. Scoliosis – quo vadis ? J Bone Joint Surg 1958 ;40-A : 507-510.

10. Conner AN. Developmental anomalies and prognosis ininfantile idiopathic scoliosis. J Bone Joint Surg 1969 ; 51-B : 711-713.

11. Deviren V, Berven S, Smith JA et al. Excision ofhemivertebrae in the management of congenital scoliosisinvolving the thoracic and thoracolumbar spine. J BoneJoint Surg 2001 ; 83-B : 496-500.

12. Dimeglio A. Growth in pediatric orthopaedics. J PediatrOrthop 2001 ; 21 : 549-555.

13. Drvaric DM, Ruderman R, Coonrad R et al. Congenitalscoliosis and urinary tract abnormalities. J Pediatr Orthop1987 ; 7 : 441-443.

14. Dubousset J. The Pediatric Spine : Principles andPractice, Raven press, Ltd, New York, 1994 ; pp 245-257.

15. Erol B, Tracy MR, Dormans JP et al. Congenitalscoliosis and vertebral malformations : characterizationof segmental defects for genetic analysis. J Pediatr Orthop2004 ; 24 : 674-682.

16. Farley FA, Hall J, Goldstein SA. Characteristics ofcongenital scoliosis in a mouse model. J Pediatr Orthop2006 ; 26 : 341-346.

17. Hall JE, Herndon WA, Levine CR. Surgical treatment ofcongenital scoliosis with or without Harrington instrumen-tation. J Bone Joint Surg 1981 ; 63-A : 608-619.

18. Hedequist D, Emans J. Congenital scoliosis. J Am AcadOrthop Surg 2004 ; 12 : 266-275.

19. Hedequist DJ, Hall JE, Emans JB. Hemivertebra exci-sion in children via simultaneous anterior and posteriorexposures. J Pediatr Orthop 2005 ; 25 : 60-63.

20. Kuhns JG, Hormell RS. Management of congenital scol-iosis. Review of one hundred seventy cases. Arch Surg1952 ; 65 : 250-263.

21. Letts R, Bobechko W. Fusion of the scoliotic spine inyoung children. Effect on prognosis and growth. ClinOrthop 1974 ; 101 : 136-145.

22. MacEwen GD, Bunnell WP, Sriram K. Acute neurolog-ical complications in the treatment of scoliosis. A report of



the Scoliosis Research Society. J Bone Joint Surg 1975 ;57-A : 404-408.

23. McCarthy RE, Campbell RM, Hall JE. Spine : state ofthe art reviews Infantile and juvenile idiopathic scoliosis.Vol. 14, Hanley & Belfus, Philadelphia, 2000, pp 163-180.

24. McMaster MJ. Congenital scoliosis caused by a unilater-al failure of vertebral segmentation with contralateralhemivertebrae. Spine 1998 ; 23 : 998-1005.

25. McMaster MJ, David CV. Hemivertebra as a cause ofscoliosis. A study of 104 patients. J Bone Joint Surg 1986 ;68-B : 588-595.

26. McMaster MJ, Ohtsuka K. The natural history of con-genital scoliosis. A study of two hundred and fifty-onepatients. J Bone Joint Surg 1982 ; 64-A : 1128-1147.

27. Moe JH, Winter RB, Bradford DS et al. (eds). Scoliosisand Other Spinal Deformities, Saunders, Philadelphia,1978 ; pp 131-202.

28. Mooney JF, III, Bernstein R, Hennrikus WL Jr.Neurologic risk management in scoliosis surgery. J PediatrOrthop 2002 ; 22 : 683-689.

29. Muhonen MG, Menezes AH, Sawin PD, Weinstein SL.Scoliosis in pediatric Chiari malformations withoutmyelodysplasia. J Neurosurg 1992 ; 77 : 69-77.

30. Muirhead A, Conner AN. The assessment of lung func-tion in children with scoliosis. J Bone Joint Surg 1985,67-A : 699-702.

31. Nasca R, Stelling F, Steel H. Progression of congenitalscoliosis due to hemivertebrae and hemivertebrae withbars. J Bone Joint Surg 1975, 57-A : 456-466.

32. Opitz JM. The developmental field concept in clinicalgenetics. J Pediatr 1982 ;101 :805-809.

33. Reckles L, Peterson H, Bianco A, Weidman W. TheAssociation of scoliosis and congenital heart defects.J Bone Joint Surg 1975 ; 57-A : 449-455.

34. Rinella A, Lenke L, Whitaker C et al. Perioperative halo-gravity traction in the treatment of severe scoliosis andkyphosis. Spine 2005 ; 30 : 475-482.

35. Roaf R. The treatment of progressive scoliosis by unilat-eral growth-arrest. J Bone Joint Surg 1963 ; 45-B : 637-651.

36. Ruf M, Harms J. Posterior hemivertebra resection withtranspedicular instrumentation :early correction in chil-dren aged 1 to 6 years. Spine 2003 ; 28 : 2132-2138.

37. Shands AR, Eisberg HB. The incidence of scoliosis in thestate of Delaware. A study of 50,000 minifilms of the chestmade during a survey for tuberculosis. J Bone Joint Surg1955 ; 37-A : 1243-1255.

38. Shawen SB, Belmont PJ Jr, Kuklo TR et al. Hemi-metameric segmental shift : a case series and review. Spine2002 ; 27 : 539-544.

39. Suk S, Chung E, Lee S et al. Posterior vertebral columnresection in fixed lumbosacral deformity. Spine 2005 ; 30 :703-710.

40. Terminology Committee of the Scoliosis ResearchSociety. A glossary of scoliosis terms. Spine 1976 ; 1 : 57-58.

41. Winter R. Convex anterior and posterior hemi-arthrodesisand epiphyseodesis in young children with progressivecongenital scoliosis. J Pediatr Orthop 1981 ; 1 : 361-366.

42. Winter R, Lonstein J, Davis F, de la Rosa H. Convexgrowth arrest for progressive scoliosis due to hemiverte-brae. J Pediatr Orthop 1988 ; 633-638.

43. Winter RB, Moe JH. The results of spinal arthrodesis forcongenital spine deformity in patients younger than5 years old. J Bone Joint Surg 1982 ; 64-A : 419-432.

44. Winter RB, Moe JH, Eilers VE. Congenital scoliosis :A study of 234 patients treated and untreated. J Bone JointSurg 1968 ; 50-A : 15-47.

45. Wynne-Davies R. Congenital vertebral anomalies :aetiology and relationship to spina bifida cystica. J MedGenet 1975 ; 12 : 280-288.

46. Zadeh HG, Sakka SA, Powell MP, Mehta MH. Absentsuperficial abdominal reflexes in children with scoliosis.An early indicator of syringomyelia. J Bone Joint Surg1995 ; 77-B : 762-767.


01-batra et al

Documents

Transcript of 01-batra et al