Early Onset Scoliosis (EOS)

Early-onset scoliosis (EOS) is a time-based definition: scoliosis with onset before age 10. It spans a heterogeneous group of diagnoses, and the primary concern is the lung. Unlike adolescent idiopathic scoliosis, EOS carries a high mortality if untreated, through thoracic insufficiency syndrome (TIS), which Campbell defined as the inability of the thorax to support normal respiration or lung growth.

The lung is still growing. Alveoli multiply rapidly until age 8. Severe deformity restricts lung volume, and the cascade runs from restrictive lung disease to pulmonary hypertension, cor pulmonale and early death. Fusion does the same harm: fusing more than 4 thoracic segments before age 9 stunts lung development and leads to TIS.

Infantile and juvenile idiopathic curves. Infantile idiopathic scoliosis is rare, under 1% of all idiopathic scoliosis in most series, with a higher incidence historically reported in Europe than in North America. Juvenile idiopathic curves are right thoracic, like AIS, and their progression is common and often relentless without treatment.

The rib-vertebral angle difference (RVAD). Mehta's angle is measured at the apical vertebra, as the angle between the rib neck and the vertebral body on each side. The RVAD is the concave-side angle minus the convex-side angle, and it separates the two infantile curves:

• Under 20° - resolving, with an 80% chance of spontaneous resolution
• Over 20° - progressive, with an 80% chance of progression

The rib-head phase. In Phase 1 there is a prominent gap between the rib head and the vertebral body. In Phase 2 the rib head overlaps the vertebral body, which indicates progression. Alongside the RVAD and the phase, a Cobb angle greater than 30° marks an infantile curve at risk of progression.

View larger

C-EOS (2014). The classification describes the curve on four axes:

• Aetiology - congenital (failure of formation or segmentation), neuromuscular (muscle imbalance), syndromic (connective tissue disorders) or idiopathic (a diagnosis of exclusion)
• Cobb angle - the current magnitude of the major curve
• Kyphosis - hyperkyphosis (greater than 50) is a major negative predictor for pulmonary function
• Progression modifier - P0 stable (under 10° a year), P1 progressive (over 10° a year), P2 malignant (over 20° a year)

Stiff or flexible. Flexibility drives the surgical decision. A flexible deformity can be managed with distraction-based techniques such as growing rods or VEPTR. A stiff, ankylosed deformity, such as one with congenital bars or fused ribs, requires release (osteotomy) or resection (vertebrectomy) before distraction.

History. Ask about the birth (prematurity, NICU admission), about development (walking age and milestones, for a neuromuscular cause) and about a family history of neurofibromatosis or Marfan syndrome.

Examination. Plagiocephaly is often associated with infantile scoliosis. The examination covers four areas:

• Skin - café-au-lait spots (NF1), a hairy patch (spinal dysraphism), laxity (Ehlers-Danlos)
• Neurology - a full examination; the abdominal reflexes are critical for syringomyelia
• Spine - flexibility: can the curve be corrected with traction or bending?
• Chest - pectus deformity, rib hump

Radiographs. Whole-spine PA and lateral films measure the Cobb angle and kyphosis, and bending or traction films assess flexibility. In an infantile curve the RVAD is calculated on the apical vertebra.

Whole-spine MRI. MRI is mandatory for all patients, and it is obtained before any deformity surgery rather than only when there are signs. A substantial minority of children with EOS have a neural axis abnormality, a Chiari malformation or syrinx, despite a normal neurological examination. Dobbs (2002) found an abnormality in 21.7% of infantile idiopathic curves of 20° or more (10 of 46, of whom 8 needed neurosurgery), and the global Pediatric Spine Study Group registry (Williams 2020) found an abnormal pretreatment MRI in 24% across all EOS types (13% idiopathic, 39% neuromuscular). The scan looks for Chiari malformation, syrinx, tethered cord, diastematomyelia and intraspinal tumours.

Genetic testing. Consider a microarray if syndromic features are present.

The goals are to maximise thoracic volume, control the deformity and allow vertical growth, which means delaying fusion.

Serial casting (Mehta/Cotrel). The indication is a progressive, flexible, idiopathic infantile curve. Under general anaesthesia the spine is put under traction and derotated and a plaster jacket applied, changed every 2-3 months. Cure is possible if casting starts young, under 2 years; otherwise the aim is to delay surgery.

Bracing (TLSO). A brace is used for juvenile cases, older children, or maintenance after casting. It is less effective than casting for true infantile curves.

Growth-friendly surgery. The indications are failed casting, a curve greater than 50-60°, or threatened lung function. The principle is distraction without fusion, and the options are:

• Traditional growing rods (TGR)
• Magnetically controlled growing rods (MCGR)
• VEPTR (rib-based)
• Shilla (self-sliding)

Early fusion. A last resort, for a severe stiff deformity. It stops lung growth and leaves a short trunk.

Algorithm

View larger

The device follows the curve. A large, progressive curve needs the corrective power of distraction. A curve where repeated surgery is undesirable may suit growth-guidance, and a small, flexible curve with lots of growth left is the candidate for compression or growth modulation. Name the category and the behaviour follows.

Magnetically controlled growing rods. Proximal and distal anchors (screws or hooks) span the deformity and are connected by a rod with a magnetic actuator. Through limited exposures at the top and bottom, the rods are passed sub-muscularly. Lengthening is done in clinic with an external magnet, so there are no repeated operations to lengthen; the costs are metal artefact on MRI and actuator failure.

VEPTR (vertical expandable prosthetic titanium rib). The indications are thoracic insufficiency syndrome, fused ribs or absent ribs. Anchored rib-to-rib or rib-to-spine (lamina), it expands the chest wall and so addresses lung volume directly. Its price is a high rate of rib migration and cut-out, and prominent hardware that can break down the skin.

Congenital hemivertebra. Fusing the convex side (hemiepiphysiodesis) lets the concave side keep growing, potentially correcting the curve over time. Resecting the hemivertebra entirely is more aggressive and gives better correction.

Failure modes follow the foundation. Anchor pull-out happens in small or osteoporotic bone, rib migration or cut-out with rib cradles, and PJK when the proximal foundation is too low. This is why a foundation uses multiple anchor points, and why salvage of failed rods often means moving the foundation, to the ribs proximally or to the pelvis distally.

The law of diminishing returns. With traditional growing rods, every open lengthening leaves the spine stiffer (auto-fusion). By the time of the final fusion the spine may already be fused in a less-than-perfect position. MCGR helps avoid this by avoiding open surgery.

The crankshaft phenomenon. If the posterior elements are fused, or auto-fuse, while the anterior vertebral body growth plates remain open, the spine twists and rotates as it grows anteriorly. To prevent it, definitive fusion in young children (under 10) often requires anterior and posterior fusion to stop all growth centres.

Pulmonary. Survival is the main goal. Early fusion (under age 5) results in severe TIS and death, and delaying fusion until age 10-12 significantly improves pulmonary volume.

Deformity and final fusion. Complete correction is rarely the goal; control is. Final fusion is usually performed at skeletal maturity, around age 12-14 in girls and 14-16 in boys, to lock in the correction.

Global epidemiology

• EOS is a heterogeneous, time-based group (onset under age 10). Idiopathic infantile scoliosis accounts for under 1% of idiopathic curves in most series, with historically higher reported rates in Europe than North America. Congenital and neuromuscular aetiologies dominate the structural/progressive end of the spectrum worldwide.

Society guidance, side by side

Registry and multicentre evidence

• Practice is driven largely by multicentre prospective EOS registries (e.g. the Pediatric Spine Study Group / Growing Spine Study Group datasets) rather than national arthroplasty-style registries, given device diversity and small numbers. These cohorts underpin the data on dual-rod growth rates, MCGR diminishing returns, and complication profiles.

High- vs limited-resource practice variation

• High-resource settings: ready access to whole-spine MRI, MCGR (avoiding repeated open lengthenings), VEPTR, intraoperative neuromonitoring and multidisciplinary paediatric pulmonology/spine teams.
• Limited-resource settings: greater reliance on serial casting and traditional growing rods (lower implant cost but repeated GA), with MRI and MCGR access constrained. Halo-gravity traction is a valuable low-cost adjunct for severe rigid curves. Late presentation with established thoracic insufficiency is more common.

• RVAD threshold: The classic 20-degree Mehta cut-off is challenged — long-term data suggest an optimal threshold nearer 17 degrees and warn against relying on a single index measurement (Lloyd et al, 2020). Serial RVAD and Cobb trends are more reliable than any one value.
• MCGR metal ions and reliability: Magnetically controlled rods reduce open lengthenings, but elevated serum titanium, actuator/pin failure, and progressive loss of distraction (a residual "law of diminishing returns") generate ongoing debate about long-term safety and true cost-effectiveness.
• Casting vs early surgery: When to abandon casting for a growth-friendly construct is not standardised. Casting can be curative if started under age 2, but late or stiff curves respond poorly.
• Rib-based vs spine-based anchors: VEPTR (rib-based) directly expands the thorax but has high migration/cutout rates; spine-based anchors give better deformity control but less direct chest-wall benefit. The optimal hybrid is unsettled.
• Graduation strategy: Whether to perform definitive fusion, "watchful waiting" with retained rods, or remove implants at skeletal maturity remains individualised, with no consensus on the best endpoint.
• Crankshaft prevention: The need for anterior growth arrest (combined anterior/posterior fusion) in very young children is debated in the era of all-pedicle-screw posterior constructs.

The Mehta Technique Min Mehta revolutionized the treatment of infantile scoliosis.

• Principle: The spine grows rapidly in the first 2 years. If you can hold it straight, the growth will correct the deformity (Hueter-Volkmann law).
• Technique:
  • General Anaesthesia.
  • Risser Table (Traction frame).
  • Derotation: Correct the rotation, not just the lateral bend. Mold over the apical rib hump.
  • Window: Cut a large anterior "mushroom" window for belly breathing (infants are diaphragmatic breathers).
• Result: Can result in a permanent cure for idiopathic curves if started early.