Adolescent idiopathic scoliosis (AIS) is defined as a coronal plane radiographic angulation greater than 10°. AIS is the most common spinal deformity encountered in children, with 2% to 3% of the population meeting this criterion. Only approximately 0.1% to 0.3% of the population have curves greater than 30°, with 0.03% having severe scoliosis and undergoing surgery for deformities greater than 45° to 50°.

The etiology of AIS is largely unknown, although genetics, environment, and bone health have all been implicated.¹ Relative anterior column overgrowth is commonly appreciated in AIS, with the subsequent buckling of the spine resulting in the complex three-dimensional deformity of AIS: coronal curvature, axial rotation, and relative lordosis² (Figure 1). Other causative conditions, such as congenital malformations, neuromuscular diseases, and syndromes, need to be ruled out to reach a diagnosis of AIS.

Like other asymptomatic pediatric conditions, understanding the natural history of untreated AIS informs management strategy. Small curves may progress in severity as patients grow; thus, skeletally immature patients have more potential for deformity progression. Curves that reach greater than 45° to 50° are thought to progress even in patients without remaining growth, presumably from degenerative mechanisms. Severe, progressive curves appear to be associated with morbidity (possible pain, pulmonary restriction, and appearance concerns) when not addressed.³ Nevertheless, these large untreated curves notably are not associated
with physical disability, unemployment, or mortality. These presumed truths that underpin modern scoliosis management are derived from best-available evidence, although limitations exist.

Clinical evaluation for spinal deformity should screen for other causes and thoroughly assess the deformity. Painful conditions (herniated nucleus pulposus, some tumors), chest wall deformity, and compensatory posture (limb-length difference) can also present as a scoliotic deformity. Severe back pain is not explained by the presence of a minor deformity and warrants further workup. Minor to moderate deformities have classically not been thought to cause back pain, although recent work has demonstrated an association between back pain and curve severity and psychosocial factors.⁴ Deformity characteristics, such as pelvic and shoulder heights, trunk shift, and truncal rotation, are important to recognize when initiating treatment. Assessment of skeletal maturity through serial height measurements, menarche status (in females), or radiographic assessment is essential in AIS.

Because growth is a critical component of planning treatment, methods for determining growth need to be accurate, reproducible, and convenient. Skeletal maturity scoring using a hand radiograph (also referred to as Sanders scoring) has become popular for its close association with peak height velocity during the adolescent growth spurt⁵ (Figure 2). The proximal humerus ossification system has also garnered recent interest for prediction of growth remaining.⁶ The Risser classification (grading of iliac apophysis ossification) is a canonical scheme for maturity determination, although its accuracy has been questioned in a 2020 study.⁷

Mild scoliotic curves can be observed clinically or radiographically for progression. Progressive curves between 20° and 40° are indicated for brace wear in patients with growth remaining. Patients with moderate curves who have Risser sign of 2 or less and Sanders score of 5 or less are considered candidates for brace wear because of significant growth remaining. Although previously controversial, there is now strong evidence that brace wear prevents curve progression to a surgical threshold. Treatment success increases from 48% to 72% with proper bracing.⁸ The effect of brace wear is dose dependent—at least 12.9 hours of daily brace wear is needed to have an effect. Scoliosis-specific exercises are being explored as an adjunctive treatment for treatment of moderate curves in addition to brace wear. Early studies show some promise to these therapies, but more evidence is needed.⁹

Attention to detail is important when starting brace treatment. There are numerous brace designs, but there is no strong evidence that any one brace is superior to others. It is likely that brace comfort and correction achieved are more critical than brace design. The psychological effects of brace wear are debated. The duration of brace wear has not been linked to differences in quality of life and body image.¹⁰ Still, avoiding unnecessary brace treatment is a common patient concern. There is no consensus on when brace wear can be discontinued, although many surgeons will cease brace wear at Sanders 7 or Risser 4. Recent publications have demonstrated continued curve progression with these criteria, but it
is unclear whether a longer bracing interval would have prevented this progression.¹¹,¹² New clinical tools that consider three-dimensional deformity parameters can better predict those at risk of progression and may help guide initial treatment and reshape bracing criteria.¹³,¹⁴

Curves progressing to 45° to 50° are considered severe scoliosis and may be offered surgery based on the aforementioned natural history of progression, even if the patient is asymptomatic. Posterior spinal fusion (PSF) with segmental instrumentation using predominantly pedicle screws is the most common treatment for AIS in the modern era (Figure 3, A and B), although anterior techniques may still have a role in treatment. Successful fusion surgery will prevent curve progression of the involved vertebra while also resulting in significant deformity correction. There are likely advantages to having scoliosis surgery as an adolescent compared with having surgery as an adult. A 2019 matched comparison study of adults versus adolescents undergoing PSF found increased levels fused (including 36% fused to the pelvis) and increased major complications (25% versus 5.4% at 2 years) in the adult patients with deformity.¹⁵ In a 2019 study, health-related quality-of-life scores in patients who have undergone spine fusion demonstrate that pain, activity, and self-image scores were improved at 5 years postoperatively in comparison with an untreated AIS group; they were also similar to those of healthy population control, except for decreased function subscores.¹⁶ Long-term outcomes data are not yet available for modern segmental fixation, but average 24.5-year data on fusions performed with nonsegmental Harrington rod fixation demonstrate health-related quality-of-life scores to be similar to those of the general population.

Emerging technologies are reshaping surgical paradigms for surgery and recovery. Enhanced recovery after surgery programs have reduced average length of stay to 2.2 days after spinal fusion, stressing multimodal pain regimens and early mobilization.¹⁷ CT navigation technologies allow visual augmentation of freehand screw placement techniques and are becoming more popular in complex deformity as well as AIS spine fusions. Navigation appears to result in more accurate screw trajectory, although the clinical importance of screw malposition is debated. A 2019 series has demonstrated that navigation technology decreased rate of return to the operating room without increasing surgery
times or blood loss.¹⁸ Anterior vertebral body tethering (VBT) is an emerging nonfusion strategy for select patients with AIS with growth remaining. This technology uses screws thoracoscopically placed in the vertebral body and connected with a flexible polyethylene tether that corrects the curve through tensile forces but still allows movement between spinal segments. The differential growth created theoretically results in growth modulation and continued correction of the scoliotic curve (Figure 3, C and D). Minimum 2-year comparative results comparing PSF with VBT demonstrate that PSF results in more sustained correction and less
reoperation (zero versus 30%). However, 74% of VBT patients successfully avoided fusion in a 2020 study.¹⁹ Further research is required to determine durability of these results and how these approaches compare long term in terms of function, flexibility, health-related quality of life, and adjacent segment disease.

When EOS progresses, but the patient is too young for definitive PSF, the surgical options are usually referred to by terminology such as growth friendly, growth preserving, or growing spine instrumentation. This category can generally be divided between guided growth or distraction-based options (Figure 4), with the latter being more popular. There is currently no specific definition of too young for PSF. One proposed goal is to achieve 22 cm of thoracic height, which is the average height of a 10-year-old child, to optimize lung function.²⁰ However, there seems to be a trend toward earlier PSF (as early as 7 to 8 years old) to avoid the physical, mental, and socioeconomic burden of growth-preserving implants.²¹,²²,²³

The traditional casting technique is referred to as Mehta casting or elongation-derotation-flexion casting and was performed on a Mehta table. Some have simplified the process by applying smaller casts that do not extend over the shoulders, which does not seem to sacrifice outcomes.²⁴ Others have simplified cast application by suspension of the spine with a single fulcrum. Casting is most beneficial when aggressive: beginning very early (eg, before age 3 years) and avoiding cast holidays.²⁵

Historically, the most commonly used distraction-based implant has been traditional growing rods (TGRs). TGRs are anchored to the spine (or ribs or pelvis) proximally and distally, and the spine between the anchors is left unfused so it can continue to grow. Patients with TGR typically undergo surgery every 6 months to distract the implants. TGRs carry a high rate of complications, especially infection, as well as financial and psychosocial burden.²⁶ Distraction-based treatment is also plagued by the law of diminishing returns—the spine stiffens and sometimes autofuses during treatment. As a result, each lengthening is progressively shorter and obtaining additional correction during the eventual fusion procedure may require more aggressive releases or osteotomies compared with a primary fusion.

The vertical expandable prosthetic titanium rib functions similarly to TGRs. It was initially developed as a rib-to-rib construct to expand the concave side of the chest to help lung development in patients with severe thoracic insufficiency syndrome. Subsequently, though,
it has been used similarly to TGRs with the ability to connect to either the ribs or the spine.²⁷ Magnetically controlled growing rods have been developed as a promising alternative to TGR (Figure 5, A and B). They are lengthened through the skin (Figure 5, C), which may reduce the physical, mental, logistical, and financial toll on the patient and family.²⁶ However, the law of diminishing returns still ultimately tends to impede the process, although it may occur more gradually than seen with TGR.²⁸

It remains unclear whether all patients with EOS undergoing distraction-based treatment require eventual revision to PSF. Most surgeons elect to perform definitive fusion, but many spines have autofused at that point. Thus, avoiding final fusion seems to be a viable option as long as the deformity is balanced and acceptable and the implants are intact. Certain implants, such as magnetically controlled growing rods, may carry additional risk if left implanted.

Guided growth procedures are designed to allow the spine to keep growing as straight as possible. Compared with traditional distraction-based treatment, guided growth can reduce the number of surgeries. Unfortunately, they do not typically seem to be as effective, possibly because distracting the spine actually accelerates growth beyond the physiologic rate through the Hueter-Volkmann principle. A historic example of guided growth for EOS was the Luqué trolley; rods were anchored to the spine proximally and distally and wrapped together with sublaminar wires. With growth, the wires would theoretically slide along the rods, holding the spine straight while allowing expansion. Unfortunately, the results were not reliable. Some authors still advocate for the Luqué trolley to treat patients with myelomeningocele with a gibbus deformity after kyphectomy. A modern version of the Luqué trolley has been described with pedicle screws rather than sublaminar wires at the proximal and distal ends of the construct, but results still remain in question.²⁹

A more recent concept for guided growth is the Shilla technique. Rather than anchoring to short, fused segments proximally and distally, the apex of the curve is fused and rods extend proximally and distally to screws with open eyelets that can slide along the rods as the spine grows. This has the theoretical benefit of straightening and fusing the part of the curve with the most deformity (the apex). Although some early reports compared favorably with TGR,³⁰,³¹ more recent, larger studies have demonstrated less spinal growth and curve correction.³² Another subcategory of guided growth is compression-based growth modulation, such as staples or anterior VBT. Perhaps in the future, these techniques will be refined to accommodate the EOS population.