Idiopathic Scoliosis

The anatomy relevant to idiopathic scoliosis is that of the thoracic and lumbar spine. Key points regarding developmental anatomy of the spine are outlined below. The anatomy specifically relevant to anterior and posterior surgical approaches to the spine is discussed further elsewhere (see Treatment, Surgical Therapy).

Developmental anatomy

Significant growth, development, and differentiation occur as a single-cell zygote progresses to become an approximately 100 trillion–cell adult human. Identifiable spine development has begun by week 3 of gestation. First, the neural tube forms. Later, paired somites appear (at 4.5 weeks' gestation), and spinal nerves are present by gestational week 6. A discernible cartilage model of the spine is present by gestational week 7.

The bone and cartilage of the spine are mesodermal derivatives, as are significant portions of the cardiovascular and urogenital systems. This explains the frequent coexistence of congenital spine anomalies with congenital cardiac and kidney defects. Thus, gestational weeks 3-7 are very important in the development of all of these major body systems.

Postnatal spinal growth also must be understood and appreciated. Dimeglio showed that the majority of spinal-canal diameter (about 90%) has been achieved by age 5 years; by age 10 years, approximately 80% of sitting height has also been achieved.[25, 26] During adolescence, radiographic evidence of ossification of the growth cartilage of the vertebral bodies occurs. Before this, these completely cartilaginous growth plates remained nestled between their respective vertebral bodies and intervertebral disks.

Much has been written regarding the potential influence of melatonin on the development of idiopathic scoliosis.[27, 28] This has largely originated from studies in which the pineal gland was removed in chickens and scoliosis developed. These same studies suggested that the melatonin deficiency following pinealectomy might be the underlying reason for the development of scoliosis.

Bagnall et al studied pinealectomized chickens to which they administered therapeutic doses of melatonin.[29] They were unable to demonstrate any ability of the melatonin to prevent the development of scoliosis. It is fair to say that no final answer is yet available.

Some authors have suggested that a posterior-column lesion within the central nervous system (CNS) might be present in patients who have idiopathic scoliosis.[30, 31] Such CNS dysfunction was hypothesized to be manifested as decreased vibratory sensation.

McInnes et al later pointed out that the vibration device used in earlier studies (a Bio-Thesiometer) did not demonstrate sufficient reliability characteristics to allow valid conclusions.[32] This line of research might be attractive to those who feel that a postural disturbance is the root cause of scoliosis.

The precise etiology of idiopathic scoliosis remains unknown, but several intriguing research avenues exist.

A primary muscle disorder has been postulated as a possible etiology of idiopathic scoliosis. The contractile proteins of platelets resemble those of skeletal muscle, and calmodulin is an important mediator of calcium-induced contractility. Kindsfater et al studied the level of platelet calmodulin in 27 patients with adolescent idiopathic scoliosis.[33] Using a direct measurement technique, they showed that patients with a progressive curve (>10° progression) had statistically higher platelet calmodulin levels (3.83 ng/μg vs 0.60 ng/μg).[33] If these data are reproduced in larger studies, they hold the potential to allow clinicians to identify patients at higher risk for curve progression.

An elastic fiber system defect (abnormal fibrillin metabolism) has been offered as one potential etiologic explanation for idiopathic scoliosis.[34] Such abnormal connective tissue has not been found universally in patients with idiopathic scoliosis. No clear cause-and-effect relation has been established. Further research in this area is clearly warranted.

Disorganized skeletal growth, probably with its root cause at a gene locus or group of loci, has been discussed as a possible etiologic explanation for idiopathic scoliosis. This theory is simply that a rather localized primary growth dysplasia leads to a cascading Hueter-Volkmann effect on a much larger portion of the spine.[35] The Hueter-Volkmann principle states that compressive forces tend to stunt skeletal growth and that distractive forces tend to accelerate skeletal growth. A possible, yet unproven, association with such a growth disturbance is the osteopenia that has been identified in patients with idiopathic scoliosis.[36]

Aronsson et al conducted a series of experiments exploring this mechanical modulation of growth. Using two different animal models (rats and calves), thye showed that the force exerted by external ring fixators was capable of producing vertebral segment wedging akin to that seen in human idiopathic scoliosis.[37, 38] Correlation of this laboratory information with the clinical setting has drawn attention to the fact that wedging occurs both from the vertebral bodies themselves and from the disk spaces, with more thoracic wedging coming from the vertebral bodies.[39] The asymmetric mechanical forces have also been associated with elevated synthetic activity in the convex side of scoliotic curves.[40]

Bylski-Austrow and Wall led a group of Cincinnati Children's Hospital researchers who further analyzed the mechanical modulation of spinal growth. Using a porcine model, they successfully induced growth changes by means of an endoscopically implanted spinal staple.[41]  Within the context of 8 weeks' follow-up, they were able to create 35-40° of scoliotic curvature in growing pigs. Histologic analysis of vertebral specimens revealed increased paraphyseal density and disorganized chondrocyte development in the region of the staple blades.

Genetic roots of the disease referred to as idiopathic scoliosis have been strongly suggested by several avenues of research. An X-linked inheritance pattern (with variable penetrance and heterogeneity) was suggested by several authors.[42] Studies of twins with scoliosis pointed in a similar direction.[43, 44]  More than 90% of monozygotic twins and more than 60% of dizygotic twins demonstrate concordance regarding their idiopathic scoliosis.[43] Some evidence has also directed attention to portions of chromosomes 6, 10, and 18 as possible scoliosis-related loci.[45]

Scoliosis is almost always discussed in terms of its prevalence (ie, the total number of existing cases within a defined population at risk). Rates may vary quite significantly according to what particular definition of scoliosis is used and what patient population is being studied. Several important studies are included below.

Stirling et al studied almost 16,000 patients aged 6-14 years in England and found the point prevalence of idiopathic scoliosis (Cobb angle >10°) to be 0.5% (76/15,799).[46] The prevalence of scoliosis was highest (1.2%) in patients aged 12-14 years.[46] Data such as these have helped reinforce the idea that the focus of screening efforts should be on children in this age group. When smaller Cobb angle measurements (eg, ≥6°) have been accepted, a significantly higher scoliotic rate may be identified, such as the 4.5% rate reported by Rogala et al.[47] Other studies using the 10° definition of scoliosis have placed the overall prevalence in the 1.9-3.0% range.[48]

Henderson et al suggested that scoliosis may develop more frequently in children born to mothers who are aged 27 years or older.[49] It could be hypothesized that gene fragility might be involved (eg, higher rate of infants with Down syndrome born to older mothers). The precise explanation as to why this might be the case has not been elucidated. In addition to this, no other authors have duplicated these results.

As mentioned previously, most patients with idiopathic scoliosis are female, and the vast majority of research has focused on females. One of the only articles written on idiopathic scoliosis in males is that by Karol et al, from the Texas Scottish Rite Hospital. These authors showed that boys with scoliosis are at risk for curve progression for a longer period than girls are. They also suggested that efforts to screen for boys with scoliosis should be performed a little later than similar screenings for girls.[50]

Clinical outcomes following treatment of idiopathic scoliosis are strongly linked to curve magnitude.[51] Unrealistic presurgical expectations have been shown to correlate with a decreased likelihood of postoperative satisfaction.[52] More long-term follow-up studies of surgically treated patients with scoliosis are becoming available. This section outlines some of these data.

One study reported that conservative treatment may result in decreased self-concept in adolescent patients with mild-to-moderate scoliosis, particularly in patients with Cobb angles of 40-50°. Comparatively, the study reported that surgically treating these patients resulted in a significant increase in self-concept.[53]

A large cohort (nearly 2000 subjects) of patients with idiopathic scoliosis in Montreal, Canada, referred to as the St Justine Cohort Study, was monitored for 10-20 years. These patients were compared to a population-based control group drawn from the general Quebec population. Compared to the general population and regardless of whether their scoliosis was treated surgically or nonsurgically, patients with scoliosis were found to have a higher self-reported rate of arthritis and poorer perceptions of their overall health, body image, and ability to participate in vigorous activities.[54, 55]

A subset of the cohort (700-1500 patients) was analyzed further with respect to low back pain.[56, 57] The researchers found a higher overall rate of significant back pain reported within the last year (75% of patients with scoliosis vs 56% of control subjects).[56] Patients with scoliosis who were treated surgically also reported a high rate (73%) of back pain within the last year, but it did not correlate with the distal extent of the spinal fusion. The St Justine authors went on to state that their study "does not provide any evidence that extending the level of fusion down even as far as L4 will increase the prevalence of back pain in adulthood."[57]

Asher et al performed a retrospective study to determine implant/fusion survivorship without reoperation and the risk factors influencing such survival. Of the 207 patients followed, 19 (9.2%) required reoperation, with 16 of those being for indications related to posterior spine instrumentation. Survival of the implant/fusion without reoperation for spine instrumentation-related indications was 96% at 5 years, 91.6% at 10 years, 87.1% at 15 years, and 73.7% at 16 years. The need for reoperation was significantly influenced by two implant variables: transverse connector design and the lower instrumented vertebra anchors used.[58]

Luhman et al reviewed the prevalence of and indications for reoperations in patients undergoing spinal fusion for idiopathic scoliosis. Of the 1057 fusions, 41 (3.9%) required reoperation: 11 anterior, 25 posterior, and five circumferential. In addition, 47 other procedures were needed: 20 revision spinal fusions (for pseudarthroses, uninstrumented curve progression, or junctional kyphosis); 16 because of infections (five acute, 11 chronic); seven for implant removal because of pain and/or prominence (four complete, three partial); two (4%) revisions for loosened implants; and two elective thoracoplasties.[59]

Yaszay et al measured the effects of different surgical approaches for adolescent idiopathic scoliosis on pulmonary function over a 2-year period in 61 patients. Patients were evaluated for vital capacity (VC) and peak flow (PF) before surgery and after surgery at 1, 3, 6, 12, and 24 months. Scoliosis approaches that penetrated the chest wall were found to result in a significant decline in postoperative pulmonary function. Return of pulmonary function did not occur until 3 months after posterior fusion with thoracoplasty; until 3 months after open anterior fusion; and until 1 year after video-assisted thoracoscopic surgery (VATS).[60]

After a 10-year follow-up, the data from another study noted that patients who experienced intraoperative chest wall violation during their spinal fusion demonstrated a significant decrease in percent-predicted forced VC and forced expiratory volume in 1 second (FEV1) values. However, those who underwent posterior-only procedures showed significant improvements in forced VC and FEV1 absolute values without any change in percent-predicted values; no changes were noted in percent-predicted values at 5 and 10 years in either group. These results suggest that procedures sparing the chest wall may result in better long-term pulmonary function.[61]

Regarding possible prognostication related to curve progression, Wei-Jun et al suggest that body weight in adolescent males may be an important parameter. Abnormal pubertal growth was noted in idiopathic scoliosis patients compared with healthy controls, with longitudinal growth being similar but body weight being significantly lower in the male adolescent scoliosis subjects.[62]

The use of 3D computer modeling combined with machine learning demonstrates promise for reasonably predicting curve progression.[63]   However, the 3D classification system and machine learning techniques are not yet widely available. Therefore, these techniques are still under study and are not standard practice at present. 

The vast majority of patients with idiopathic scoliosis initially present because of a perceived deformity. This may be patient or family perception of asymmetry about the shoulders, waist, or rib cage. The asymmetry of the rib cage can cause patients to present with concerns for rib-cage deformity or breast asymmetry before they are aware of having scoliosis. A primary care physician or school-screening nurse may perceive some of these findings as well.

The Adams forward-bending test (in conjunction with the use of a scoliometer) has been found to be an effective screening tool. Traditionally, a value of 7º of angulation on the scoliometer has been used to determine whether the child or adolescent should undergo radiography or be referred to an orthopedic surgeon.

Highlights of the patient's history include obtaining information relative to other family members with spinal deformity, performing an assessment of physiologic maturity (eg, menarche), and ascertaining the presence or absence of pain.

Traditionally, scoliosis has been described as a nonpainful condition, and aggressive workup has been recommended for patients in whom this rule is violated.[64] Ramirez et al from the Texas Scottish Rite Hospital studied more than 2400 patients with scoliosis and found that a full 23% (560 of 2442 patients) had back pain at the time of presentation.[65] An underlying pathologic condition was identified in 9% (48 of 560) of the patients with back pain; in most cases, the underlying condition was spondylolysis or spondylolisthesis, but one instance of an intraspinal tumor was also noted. 

Physical examination should include a baseline assessment of posture and body contour. Shoulder unleveling and protruding scapulae are common. In the most common curve pattern (right thoracic), the right shoulder is consistently rotated forward, and the medial border of the right scapula protrudes posteriorly. Truncal shift is present when the shoulders, neck, and head are translated laterally instead of centered over the pelvis.

A basic neurologic evaluation should be performed, including assessment of lower-extremity (and often upper-extremity) reflexes, abdominal reflex, Babinski sign, and clonus. The presence or absence of hamstring tightness should be investigated, and screening should be performed for ataxia and/or poor balance or proprioception (ie, Romberg test).

Recognizing a difference in limb lengths will prove valuable, in that a significant percentage of patients presenting with scoliosis have several centimeters of limb-length discrepancy; this creates an oblique takeoff of the spine, which, in turn, causes the compensatory curve back to the midline. This pattern of scoliosis is not truly idiopathic. This scoliosis is secondary to a limb-length discrepancy and, if still flexible, will be corrected if the radiographs are taken with a equalizing lift under the shorter limb.

Laboratory workup for patients with scoliosis consists primarily of preoperative testing. Most, if not all, patients undergo preoperative assessment of hemoglobin and hematocrit levels. A type and screen is obtained, if not a complete type and crossmatch for packed red blood cells.

Autologous blood can be drawn preoperatively to avoid the need for allogenic blood transfusion. However, there are concerns related to obtaining autologous blood, such as possible wastage of blood products if there is no significant intraoperative blood loss and possible inappropriate transfusion out of a felt need to use the autologous blood already obtained.[66]  If there is a small interval of time between when the autologous blood is obtained and when the surgical procedure is performed, using this blood can result in lower hemoglobin and hematocrit levels at the start of the surgical case.

Radiography

Multiple authors have cited the value of bending radiographs, including those over a fulcrum.[67] Klepps and Lenke et al found that thoracic fulcrum bending radiographs worked best for them when dealing with isolated main thoracic curves.[68]

After the publication of the King classification in the early 1980s, the thoracic curve patterns found in adolescent idiopathic scoliosis were commonly classified according to this system.[69] Subsequently, significant questions were raised regarding its reliability and reproducibility.[70, 71] In addition, it was noted that the King classification alone (in its original form) does not allow comprehensive curve classification (eg, lumbar and thoracolumbar curve patterns).[72]  Many now regard the King classification as primarily of historical interest. The Lenke classification (see below) has become more commonly used for adolescent idiopathic scoliosis.[73]

Multiple authors have analyzed the ability of orthopedic surgeons to reliably measure scoliosis radiographs. Morrissy et al used 50 radiographs and four examiners (two experienced orthopedic surgeons, one fellow, and one senior resident) to study their ability to make Cobb angle measurements. With the examiners choosing end vertebrae and measuring scoliotic curves accordingly, intraobserver variability was 4.9°.[74]

Carman et al used eight scoliosis radiographs measured by five examiners (four orthopedic surgeons, one physical therapist) to evaluate interobserver and intraobserver variation. They found that a 10° measurement difference is necessary before there is a 95% confidence level that one Cobb angle measurement is truly different from another.[75]

Lenke classification

Currently, the Lenke classification system is commonly used to categorize adolescent idiopathic scoliosis. This system, first published in 2001, includes the following three components[73] :

• Curve type (1, 2, 3, 4, 5, or 6)
• Lumbar spine modifier (A, B, or C)
• Sagittal thoracic modifier (–, N, or +)

On coronal and sagittal radiographs, the six types specified by Lenke et al have specific characteristics that distinguish structural and nonstructural curves in the proximal thoracic (PT), main thoracic (MT), thoracolumbar (TL), and lumbar (L) regions.[73]  Regional curves are measured, the major curve is identified, and a determination is made as to whether the minor curve is structural. The curve is then assigned to the appropriate numeric type (1 through 6).

The lumbar spine modifier is based on the relation of the center sacral vertical line (CSVL) to the apex of the curve. If the CSVL passes between pedicles of apical lumbar vertebrae, the modifier A is assigned; if it touches a pedicle, the modifier B is assigned; and if it does not touch apical lumbar vertebrae, the modifier C is assigned.

The sagittal thoracic modifier is based on the sagittal Cobb angle from T5 to T12. If the angle is less than 10º (hypokyphotic), the modifier – is assigned; if it is 10-40º (normal), the modifier N is assigned; and if it exceeds 40º (hyperkyphotic), the modifier + is assigned.

Further application of this classification system to the surgical instrumentation and fusion planning are discussed in the "Surgical Therapy" section below.

Magnetic resonance imaging

It has been suggested that magnetic resonance imaging (MRI) is indicated primarily in patients who have idiopathic scoliosis with unusual complaints (eg, severe unexplained headaches), as well as in those with atypical clinical findings (eg, ataxia, cavus foot deformity, severe kyphosis, or syndromic features).[76]  In cases of infantile and juvenile idiopathic scoliosis, it is more common to obtain a total spine MRI to look for intraspinal pathology (eg, tethered cords, syrinx, and Chiari malformations), especially when the curve exceeds 20º.

Routine MRI evaluation of all patients with adolescent idiopathic scoliosis is not recommended but can be warranted in the setting of a rapidly progressing curve, associated kyphosis, or other atypical features (eg, a left-side thoracic curve).[77]

Bone age radiographs

Left-hand radiographs are used to determine skeletal maturity in adolescents with scoliosis. The Sanders classification is used to score the hand radiographs and is more accurate than the Risser classification in prediciting the stage of adolescent growth velocity.[78]

Pulmonary function testing

Pulmonary function studies have been used extensively in the evaluation of patients with idiopathic scoliosis.[7, 79, 80, 81] In general, patients whose scoliosis surgery does not involve disruption of the chest wall can be expected to experience improved postoperative pulmonary function.[82, 60] Other authors have suggested that an impairment in respiratory mechanics may persist after successful scoliosis surgery.[83] Preoperative pulmonary function testing is of questionable value in patients with moderate deformity (average Cobb angle, 48°), in that most of these patients can be expected to have normal or only mildly abnormal results.[84]

Scoliosis is clearly a disease that is strongly influenced by, if not completely rooted in, spinal growth. It has even been referred to by some as "an unsynchronized growth."[85]

Hsu et al studied muscle biopsies from 27 patients with idiopathic scoliosis who were undergoing posterior spinal fusion. Specimens were obtained from the paraspinal musculature of both the convex and the concave side. All patients had thoracic curves in the range of 37-81°.[86]  In all, 68% of the patients demonstrated abnormalities in muscle fiber distribution. The abnormalities were similar on the convex and concave sides, the most notable being a reversal of the normal type 2 fiber ratio, so that type 2A fibers predominated over type 2B fibers. These changes are similar to those seen in endurance training and might be due to the extra work of trying to maintain posture in the setting of scoliosis.

An extensive yet incomplete understanding of the natural history of idiopathic scoliosis remains a reality. Thus, selection of recommended treatments for idiopathic scoliosis continues to be associated with more than a modicum of uncertainty. The main treatment options for idiopathic scoliosis may be summarized as the three Os, as follows:

• Observation
• Orthosis
• Operative intervention

When to choose each of these treatments is a complicated matter.

The risk of curve progression varies according to the idiopathic scoliosis group to which a patient belongs (ie, infantile, juvenile, or adolescent).

The future of the understanding of idiopathic scoliosis will clearly be guided by human genome analysis.[87] The characterization of the structure and function of specific gene loci and the eventual ability to regulate their expression will undoubtedly form the basis of scoliosis treatments of the future. Someday, clinicians may look back upon present mechanically based treatments of scoliosis and wonder how patients ever benefited.

There continue to be controversies regarding several surgical tactics that may be used to treat similar curve types. Examples of this include anterior fusion and instrumentation versus posterior fusion and instrumentation for isolated thoracic curves. Both validated methods of curve classification and prospective randomized controlled studies comparing the surgical methods will be necessary before definitive answers can be embraced.

Future potential also exists in strategies for modulating spinal growth as a means of treating idiopathic scoliosis. This modulation may be genetic or mechanical in nature.

Infantile idiopathic scoliosis

Although defined by a seemingly arbitrary age limit (< 3 y at the time of diagnosis), infantile idiopathic scoliosis demonstrates marked differences that distinguish it from the other two categories of idiopathic scoliosis.

Infantile idiopathic scoliosis is the only type of idiopathic scoliosis whose most common curve pattern is left thoracic. It is the only type of scoliosis that is more common in boys. It is more common in European patients or those of immediate European descent. In the past, infantile idiopathic scoliosis may have constituted as many as 41% of all idiopathic scoliosis cases in parts of Europe, but the current rate would appear to be closer to 4%. This is still dramatically higher than the estimated 0.5% rate in North America.[88]

Infantile idiopathic scoliosis is also the only type of idiopathic scoliosis with any significant reputation for spontaneous resolution. Reported spontaneous resolution rates are in the range of 20-92%.[8, 89] Ceballos et al, in a study that included 113 Spanish patients with infantile idiopathic scoliosis, found a 92% rate of associated plagiocephaly and an almost 25% rate of congenital hip dysplasia.[90] In addition, they found that nearly 74% of their patients' curves were of the resolving variety (mainly left thoracic curves) and the other 26% were progressive curves (mainly double primary–type curves).

Prediction of curve progression in infantile idiopathic scoliosis has been tied to assessment of the rib vertebral angle difference (RVAD) originally described by Mehta in 1972.[91] As described by Mehta, this measurement is carried out at the apical vertebra of the curve. In instances in which the curves resolved spontaneously, the RVAD was less than 20° in about 80% of cases, and in those instances in which the curves were progressive, the RVAD exceeded 20° in about 80% of cases. Other authors have confirmed the prognostic value of the RVAD, as well as its reliable application.[90, 92]

Treatment indications

Nonoperative treatment of progressive infantile idiopathic scoliosis predominates and may involve the use of conventional thoracolumbosacral orthosis (TLSO)-type braces, Milwaukee-type braces, and even intermittent Risser casting. Some have questioned the value of bracing in infantile idiopathic scoliosis and have stated that "a curve that resolves in a brace would probably have resolved without treatment."[88]

If surgical treatment becomes necessary, anterior release and fusion followed by posterior spinal fusion with instrumentation is considered to be the functional treatment. Every effort should be made to delay such surgical intervention as long as possible so as to optimize spinal growth, but relentless curve progression should not be accepted or tolerated while some arbitrary chronologic age is awaited.

Although convex spinal epiphysiodesis (which has been shown to be quite effective in the management of congenital scoliosis) is intuitively attractive, it has not been shown to be as reliable in the setting of infantile idiopathic scoliosis.[93] Addition of some type of posterior instrumentation may improve the results of epiphysiodesis.[94]

A treatment outline for infantile idiopathic scoliosis may be as follows:

• Curves less than 25° with an RVAD less than 20° are preferentially observed and monitored with spinal radiographs at regular intervals
• Curves exceeding these parameters are either braced or will begin elongation derotation flexion (EDF) casting (also known as Mehta casting) ^[95]
• Surgery is considered for curves not adequately controlled with nonoperative measures

Juvenile idiopathic scoliosis

Juvenile idiopathic scoliosis most closely mimics the epidemiology and demographics of the adolescent version of the disease. It is more common in females, and its most common curve pattern is a right thoracic curve.[9] In fact, given its demographic similarities, high rate of progression, and need for surgery, juvenile idiopathic scoliosis might be considered to be a malignant subtype of adolescent idiopathic scoliosis.

Robinson and McMaster studied 109 patients with juvenile idiopathic scoliosis in Scotland and found that 95% (104/109) demonstrated curve progression and 64% (70/109) progressed to require a spinal fusion.[9] This spinal fusion rate is similar to that reported by James 15 years earlier.[96]

A study from Washington University found a 50% rate of neural axis abnormalities in young children (< 10 y) with idiopathic scoliosis.[97] These findings included Chiari type I malformations and dural ectasia. At least one case report also exists in which a spinal intraosseous arteriovenous malformation was found in association with juvenile scoliosis.[98]

Treatment indications

One potential treatment algorithm for juvenile idiopathic scoliosis is as follows:

• Observation for curves less than 25° with follow-up radiographs at regular intervals
• Bracing for curves that range from 25º to 40° and at least consideration of bracing (based on curve flexibility) for curves from 40º to 50°
• Bracing for smaller curves that demonstrate rapid progression to the 20-25° range
• Surgical intervention for inflexible curves that exceed 40° or virtually any curve that exceeds 50°.

Bracing and casting may be used outside the above-mentioned parameters in an effort to help control a large curve in a young child for whom the surgeon is attempting to optimize spinal growth. Similar recommendations exist regarding the value of MRI in juvenile idiopathic scoliosis due to a significant rate of neural axis abnormalities.[97]

Adolescent idiopathic scoliosis

Adolescent idiopathic scoliosis is the most common type of idiopathic scoliosis and the most common type of scoliosis overall. Progressive curvature may be predicted by a combination of physiologic and skeletal maturity factors and curve magnitude. Small curves in more mature patients have a substantially lower risk of progression (~2%) than larger curves in more immature patients, in whom the risk is much higher (approaching or exceeding 70%).[99]

Treatment indications

Treatment recommendations for adolescent idiopathic scoliosis are driven almost totally by curve magnitude (the only caveat being that brace treatment is thought to be effective only in patients who are still growing). It is thus somewhat ironic to note that stated recommendations urge observation for curves less than 30°, bracing of curves that reach the 30-40° range, and consideration of surgery for curves that exceed 40°. This amounts to a 10° window between observation and major spinal surgery. It is even more ironic to note that 10° is a commonly discussed margin of error for measuring such scoliotic curves.

Additional patient factors may also influence some orthopedic surgeons to brace patients with curves measuring less than 30° or in excess of 40°. For instance, a rapidly progressive curve in a 12-year-old child that suddenly goes from 16º to 26° may easily prompt bracing.

When it comes to surgical considerations, patients with adolescent idiopathic scoliosis may be functionally subdivided into those patients in whom significant anterior spinal growth is a concern and those in whom it is not. This amounts to a quantification of risk of development of the complication known as crankshaft phenomenon.[100] This can have a major impact on the surgical treatment plan, in that a child at significant risk for crankshaft phenomenon will require an anterior spinal fusion procedure.

Much effort has been devoted to predicting which patients may suffer from this continued anterior spinal growth that results in progressive angulation and rotation of the spine.[100, 101, 102, 103, 104] In fact, a hierarchy of risk can be constructed in which progressively more precise estimates can be made. In this hierarchy, the presence of a radiographic Risser sign and the attainment of menarche are somewhat predictive but less so than closure of the triradiate cartilage; reaching one's peak height velocity is perhaps the most powerful predictor of being at rather low risk for the crankshaft phenomenon.

Nonoperative management consists of either simple observation or orthosis use. Observation is watchful waiting with appropriate intermittent radiographs to check for the presence or absence of curve progression. If the patient is in a rapid phase of growth like the toddler and adolescent growth spurt, radiographs are usually checked about every 6 months. During less rapid growth periods (age 5-10 y), radiographs typically can be obtained annually. Orthosis use for scoliosis is discussed extensively below. Some studies have found physiotherapeutic scoliosis-specific exercises (PSSEs) to be beneficial in adolescents with idiopathic scoliosis.[105]   

No other treatments, including electrical muscle stimulation, usual physical therapy, spinal manipulation, and nutritional therapies, have been shown to be effective for managing the spinal deformity associated with idiopathic scoliosis. The lack of demonstrated effectiveness in this context means either that scientifically valid studies have been done that do not support the treatment or that no such studies have yet been published that would allow an evidence-based evaluation.

The International Scientific Society on Scoliosis Orthopaedic and Rehabilitation Treatment (SOSORT) has published guidelines for the use of conservative treatment approaches to idiopathic scoliosis.[106]  (See Guidelines.)

The first widely used scoliosis brace with proven effectiveness was the Milwaukee brace. This brace was developed by Walter Blount and Albert Schmitt and introduced at a meeting of the American Academy of Orthopaedic Surgeons (AAOS) in 1946.[107] It originally was designed for use as a component of the surgical treatment of scoliosis and only later evolved into a standalone nonoperative treatment.

In a study of 1020 patients with adolescent idiopathic scoliosis treated with the Milwaukee brace, Lonstein and Winter reported that this orthosis was effective in preventing significant curve progression in patients with 20-39° curves.[108] These same authors recommended that adolescents with a curve of 25° and a Risser sign of 0 be braced immediately and not wait for evidence of curve progression. Other authors have shown that an average curve correction of 20% in the brace (Milwaukee brace) is associated with bracing success.[109, 110]

Rowe et al performed a meta-analysis aimed at evaluating the efficacy of nonoperative treatments for idiopathic scoliosis.[111] They calculated the weighted mean proportion of success for three nonoperative treatments: observation, electrical stimulation, and bracing. They were able to successfully combine data on 1910 patients from 20 different studies for the purposes of meta-analysis and reported the following main results:

• Observation, 49% success rate
• Electrical stimulation, 39% success rate
• Bracing 8 hr/day, 60% success rate
• Bracing 16 hr/day, 62% success rate
• Bracing 23 hr/day, 93% success rate

In a prospective multicenter study from the Scoliosis Research Society, Nachemson et al found brace treatment (an underarm plastic brace worn for at least 16 hr/day) to be successful 74% of the time.[112] Some authors have not been able to identify a major difference between full-time bracing (23 hr/day) and part-time bracing (12-16 hr/day).[113]

The psychological stress associated with scoliosis has been documented,[114] and this does not improve compliance with brace wear. MacLean et al from Vanderbilt studied 31 adolescent and preadolescent females who were undergoing part-time brace treatment for their idiopathic scoliosis.[115] Part-time bracing was defined as 13-16 hr/day. The investigators noted that 84% of patients described the initial period of bracing in "stressful terms" and experienced lower levels of self-esteem. A reassuring finding was that no overt psychopathology was identified in this study.

Compliance with prescribed brace-wear regimens has been shown to be poor. DiRaimondo and Green found that on average, patients wore their braces only 65% of the prescribed amount of time.[116] Patients prescribed part-time (16 hr/day) bracing actually demonstrated worse compliance (58%) than those prescribed full-time (24 hr/day) bracing (71%). Overall, only 15% of patients demonstrated a highly compliant (≥90%) brace-wear routine.

Questions have also been raised regarding the consistency of strap tension in TLSO bracing. Using an instrumented load cell to measure strap tension, Aubin et al studied 34 of their patients with braces in Quebec.[117] They found marked variability in tension, with the greatest change occurring while patients were recumbent.

In part because of the aforementioned psychological and brace-wear compliance issues, new approaches to bracing are being developed.[118, 119] One such approach, developed by Coillard and Rivard of the St Justine Hospital in Montreal, is a dynamic bracing method known as the SpineCor Brace or the St Justine Brace.[120] It involves elastic straps that are anchored on a pelvic corset, and on the basis of curve morphology, these straps are tensioned to exert corrective forces. This brace is a radical departure from traditional plastic and metal orthoses.

Early results with the St Justine Brace were encouraging, with success rates comparable to those of traditional bracing. Continued follow-up of their growing international cohort of patients is necessary. A study by Gutman et al found the SpineCor brace to be less effective than the Boston brace for treatment of adolescent idiopathic scoliosis.[121]

The Providence Brace is designed to be worn only at nighttime and in the supine position. It is worn for 8-10 hours at a time. The in-brace curve correction is more aggressive than is the case with the full-time braces, and follow-up studies show that it does improve the coronal curvature as the patient continues growth.[122]  This brace, like all braces, requires compliance in order to achieve good results. Some surgeons will select this brace when they do not believe their patients will be compliant with the full-time or other part-time braces discussed above.  

Contraindications

Few, if any, absolute contraindications exist regarding scoliosis care, just as few, if any, absolute indications for intervention exist. Accepted contraindications for bracing include skeletal maturity and excessive curve magnitude. Thoracic lordosis and certain curve patterns (eg, double thoracic curves) also have been offered as at least relative contraindications to bracing.

The main contraindication for posterior scoliosis surgery would be medical instability and inability to survive surgery. Anterior scoliosis surgery would also be contraindicated in these patients, as well as in those with a precarious pulmonary status.

Even in the setting of adequate correction and solid fusion, as many as 38% of patients still have occasional back pain.[71, 123]  Therefore, the primary aims of scoliosis surgery are to halt curve progression, to decrease the spinal deformity, and to minimize complications in the process. The gold standard of scoliosis treatment remains spinal fusion. The surgical techniques used to achieve such an arthrodesis are vastly more important than the instrumentation system that the surgeon needs to use, if any.[6, 124]

Modern instrumentation systems allow adequate curve correction but possess little or no ability to diminish associated rib humps.[125] With the use of direct vertebral-body derotation techniques. about 50% of the rib hump can be improved; otherwise, thoracoplasty is required to resolve the rib deformity.[126]  It is possible that derotation of the instrumented curve may occur at the expense of creation of new rotation in uninstrumented portions of the spine; accordingly, counterderotation of the lowest instrumented vertebra is important to prevent this occurrence.[127]

Preoperative considerations

Preoperative evaluation includes a number of factors: curve location, curve magnitude, curve flexibility, curve rotation, trunk shift, and shoulder height. These parameters are considered in conjunction with patient maturity factors to determine the optimal treatment choice. Surgical treatment focuses on halting the curve progression, improving the coronal-sagittal balance, leaving as many mobile segments free as possible, and avoiding complications.

Surgeons may choose from a diverse array of anchors to secure large-diameter rods (usually in the 0.25-in. [6.35-mm] range) to the spine. These anchors include sublaminar bands, pedicle hooks, transverse process hooks, sublaminar wires (Luque wires), spinous process wires (Drummond wires), and pedicle screws.

When there is a large, stiff thoracic curve (usually not bending less than 50°), additional tactics are considered, such as Ponte osteotomies or an anterior release. 

Levels of fusion

The Lenke classification provides surgeons with a common language for discussion, and a number of studies have examined the outcomes of surgical treatment of the various Lenke curves. These studies give surgeons a foundation to start with in determining which levels to instrument and fuse. 

Lenke 1

In determining the upper instrumented vertebra (UIV), preoperative shoulder symmetry is the important factor. For right thoracic curves with a high right shoulder, the UIV typically can stop at T4 or T5 to level the shoulder postoperatively. However, some surgeons have expressed concern that there is an increased risk of proximal junctional kyphosis (PJK) when the UIV is stopped below T4. When the shoulders are level preoperatively, T3  is chosen as the UIV. When the left shoulder is higher, T2 is recommended as the UIV.

For the lower instrumented vertebra (LIV), historically, the stable vertebra (the vertebra most closely bisected by the central sacral vertebral line [CSVL]) was suggested during the use of Harrington rod fixation.[128]

Suk et al argued that with the introduction of pedicle screw fixation and derotation techniques, it is critical to understand the rotation of the compensatory lumbar curve.[129] They reported that results are satisfactory with fusion down to the neutral rotated vertebra (NV) if the NV is within one segment of the end vertebra (of the Cobb angle), as well as with fusion down to one vertebra short of the neutral vertebra (NV-1) if the NV is two or more segments distal to the end vertebra. Although this logic is now applied to type Lenke 1 curves, it should be noted that this study analyzed single thoracic curves based on type 3 and 4 curves of the King classification system.

One should note the tilt of the L4 vertebra when treating a Lenke 1A curve. Cho et al reviewed the risk of adding on for Lenke 1AR curves (L4 tilted to the right) versus Lenke 1AL curves (L4 tilted to the left) and found that 1AR curves were 2.2 times more likely to experience adding on than 1AL curves were.[130] They recommended that for 1AR curves, the LIV should be closer to the NV and the last substantially touched vertebra of the CSVL.

Treatment of type C curves is more nuanced because of concerns about lumbar curve decompensation. Finding the apical vertebra translation (AVT) length for the thoracic curve and the thoracolumbar/lumbar curve and calculating their ratio (thoracic AVT to thoracolumbar/lumbar AVT) is a reliable method to help determine the potential for selective thoracic fusion. In addition, the apical vertebral rotation (AVR) and the relative difference in the Cobb angles are also considered.

It is expected that AVT, AVR, and Cobb ratios greater than or equal to 1.2 will do well, but values of 1.0 or less are associated with decompensation.[131]  However, in the application of this method, it is recommended to refrain from using derotation techniques and to undercorrect the thoracic curve. Other authors have challenged this recommendationand found good results when using apical thoracic derotation, in addition to counterderotation below at the end vertebra to initiate detorque for the lumbar curve.[132]

Lenke 2

When there are two structural thoracic curves, both are included in the fusion. Therefore, when the proximal curve is included, the UIV generally is not below T3. If the left shoulder is higher, using a UIV of T2 is generally sufficient to level the shoulders.

For the LIV, the rules are very similar to those for Lenke 1 curves. Selective thoracic fusion is used for Lenke 2A and 2B curves. When the C modifier applies, an AVT ratio of 1.2 for the main thoracic curve to the thoracolumbar/lumbar curve is again used to determine the adequacy of selective thoracic fusion.

Lenke 3

For Lenke 3 curves, it is generally accepted that both structural curves should undergo instrumentation and fusion. Because the proximal thoracic curve is nonstructural, the UIV typically ends at either T3 or T4, but again, some surgeons will even stop at T5. The LIV is generally at L3 or L4.  One wants to avoid ending the instrumentation at the apex of the lumbar curve.

In the effort to leave as many mobile disk segments as possible, there are some helpful radiographic indicators for determining whether L3 is an adequate LIV. If side bending radiographs show a minimally rotated L3 vertebra that levels with the pelvis or tilts the opposite direction, or if the Cobb angle ends at L3 and the last touch vertebra is L4 or above, then L3 will be an adequate LIV.[133]

The indications for selective thoracic fusion have been expanding over the years to include Lenke 3C curves with a thoracic-to-thoracolumbar/lumbar AVT ratio of 1.2 or greater. Also, adequate outcomes were found using the direct vertebral rotation and counterrotation technique previously described under the Lenke 1 discussion.

Lenke 4

When the proximal thoracic, main thoracic, and lumbar curves are all structural, typically all of the curves are instrumented and fused. The rules for Lenke 2 curves should be followed to address the proximal curve, and the rules for addressing Lenke 3 curves should be followed to determine the LIV.

Lenke 5

Lenke revealed that using a ratio of 1.25 for the AVT, AVR, and Cobb angle of the thoracolumbar/lumbar curve to the thoracic curve indicates a good long-term result when selective thoracolumbar/lumbar fusion is performed. In addition, Lenke gave one relative contraindication for performing this procedure: He warned that if the left shoulder was depressed preoperatively, it would be further lowered postoperatively. Finally, Lenke advised caution in the use of selective thoracic fusion in very immature (Risser 0) patients out of concern regarding progression of the thoracic curve.

Lenke 6

Lenke 6 curves are treated similar to Lenke 3 curves but since the thoracolumbar/lumbar curve is the major curve, it will always be instrumented and fused unlike in some cases of the Lenke 3C curves.  The decision to use and LIV of L3 vs L4 is determined again using side-bending radiographs and the last touched vertebra of CSVL described under the treatment of Lenke 3 curves.

Posterior approach

The major superficial muscles of the back are not often directly visualized during posterior surgical approaches for scoliosis, but they must not be forgotten. These muscles include the trapezius, the rhomboid major, the rhomboid minor, and the latissimus dorsi. Using an animal model, Kawaguchi et al showed that significant posterior muscle injury can be induced by the pressure exerted by surgical retractors.[134] This certainly makes a case for intermittent removal and replacement of such retractors during the course of posterior spinal surgery.

The route for exposure of the posterior spinal elements passes through the cartilaginous apophyses of the spinous processes. These structures, often referred to as the cartilaginous caps, are systematically split in the midline to allow sequential subperiosteal dissection of the spinous processes, laminae, facet joints, and transverse processes.

The laminae of the thoracic vertebrae spread out from the midline like wings and flow upwards (cranially) in the direction of the transverse processes. The facet joints of the thoracic spine are shingled in a coronal plane in such a way that the inferior facet contributing to each joint is located posteriorly and the superior facet is located anteriorly. The thickness of the interior and superior facets of the thoracic spine is in the range of 3-5 mm.[135] The thoracic facet joints are located a mere 7-11 mm from the midline of the posterior spine.

In progressing from the thoracic to the lumbar spine, important differences are noted. The V-shaped laminae of the thoracic spine give way to the butterfly-shaped laminae of the lumbar spine. This orientation change is important for the surgeon to remember when exposing these bony elements. The facet joints of the thoracic spine, which are oriented in more of a coronal plane, transition into the more sagittally oriented facet joints of the lumbar spine. The transverse processes of the thoracic spine, which seem to flow directly up and away from the laminae, change significantly in the lumbar spine so that they are no longer in close proximity to the laminae and are located anterior and inferior to the lumbar facet joints.

The ribs are also obviously absent in the lumbar vertebrae. What some consider a rib remnant does persist and is referred to as a mammillary body or mammillary process. It is most pronounced near the thoracolumbar junction but may be identified on nearly all of the lumbar segments. In the sagittal plane, one must also appreciate that the normal gentle kyphosis of the thoracic spine reaches its apex at about the T7-9 region. Below this, a rather definite transition to lumbar lordosis occurs, with an apex around the L3 level.

Thoracic kyphosis is typically in the range of 20-40° (Cobb measurements usually taken from the top of T3 to the bottom of T12). Some authors have stated that up to 50° of thoracic kyphosis should be considered normal.[136] Normal lumbar lordosis is considered by some to range from 35º to 55° (Cobb measurements usually taken from the top of L1 to the top of L5).

Anterior approach

Anterior scoliosis surgery involves three main strategies, as follows:

• Anterior lumbar or thoracolumbar surgery through a retroperitoneal approach that may or may not involve a diaphragmatic incision
• Anterior thoracic surgery via traditional open thoracotomy
• Anterior thoracic sugery via video-assisted thoracoscopic surgery (VATS)

Various factors relative to skeletal maturity, curve location, and curve flexibility help determine which (if any) of these anterior surgeries may be appropriate.

The most common reason to use the retroperitoneal approach is for an instrumented anterior thoracolumbar spinal fusion. The most common curve pattern in that particular type of scoliosis is an apex left curve pattern; accordingly, the patient is usually positioned lying on the right side. This position is advantageous in that it provides the best access to the scoliotic spine and also places the thick-walled aorta closer to the surgical field (as opposed to the thin-walled inferior vena cava).

After superficial muscle dissection, the surgeon approach proceeds through the bed of the rib that corresponds with the highest vertebra in which instrumentation is planned. This is often either the ninth or tenth rib, with the rib itself being harvested for later use as a bone graft.

Careful dissection is then performed to mobilize the peritoneum (with its contents) in an anterior direction; it is peeled off of the undersurface of the diaphragm. Posterior division of the diaphragm (leaving about a 2-cm cuff for repair) helps to avoid damage to the phrenic nerve. Diaphragmatic division begins with splitting of the costal cartilage and proceeds in a posterior direction with intermittently placed tagging sutures to aid in closure.

The remainder of the retroperitoneal approach to the thoracolumbar spine requires careful superior retraction of the lung, anterior retraction of the peritoneum (with associated aorta and ureter), and posterior retraction of the iliopsoas musculature. Careful identification and division of the segmental vessels (overlying the vertebral bodies) is carried out with either electrocautery or ligatures.

Small sympathetic nerve branches in this same area are sacrificed during this stage of the exposure. This results in at least a transient period in which the left foot (for a left-side approach) will be both pinker and warmer than the contralateral foot. At times, this may result in nursing personnel notifying the surgeon that the contralateral foot is pale and cold, but in reality, it is the foot ipsilateral to the exposure that has changed.

Open thoracotomy might be performed either for anterior thoracic spine release followed by posterior fusion or for anterior thoracic spine fusion with instrumentation. The most common curve pattern to address with this approach would be a right thoracic curve; accordingly, the patient would be positioned with the right side upward.

A similar rib selection and resection technique may be used if desired. From the interior of the chest, the intercostal musculature (located between each of the ribs) can be seen. Identification of the azygos vein (anteriorly oriented along the vertebral bodies) is necessary. Further medial (ie, central) and running parallel to the azygos vein is the thoracic duct. Several portions of the sympathetic chain may be sacrificed as the segmental vessels overlying the thoracic vertebral bodies are divided and mobilized anteriorly and posteriorly. Blood flow changes similar to those noted in the retroperitoneal approach may be noted in the right foot (for a right thoracotomy).

In addition to this, thoracic surgical dissection carries with it the possibility of sacrificing branches to the greater splanchnic nerve, which would theoretically decrease the visceral referred pain that one might feel from an inflamed gallbladder or similar condition.

Thoracoscopy or miniopen thoracotomy techniques are being used to perform vertebral body tethering. Just as arthroscopic knee surgeons enjoyed an expansion in visualized anatomy in comparison with that visible in knee arthrotomies, the endoscopic spine surgeon benefits from much greater intrathoracic latitude. Most VATS procedures involve the right thoracic cavity, and this discussion focuses on that particular side.

Proper rib counting and visualization of the superior intercostal vein (formed by the confluence of the second, third, and fourth intercostal veins) as it empties into the azygos vein are necessary steps to orient the surgeon. Beyond this, one also notes the mounds and valleys of the thoracic spine, with the mounds being the disks and the valleys being the vertebral bodies with the segmental vessels that overly them.[137]

The same anatomy outlined in the thoracotomy discussion still clearly applies, but further endoscopic fine points are needed. Specifically, the thoracic spine may be considered to be composed of three separate fields, with important anatomic nuances.[138] The upper field may be considered to be T2-5, the middle field may be considered to be T6-9, and the lower field may be considered to be T10-L1.

The upper field is dominated by the superior intercostal vein, and it is characterized by the fact that the rib heads tend to span their respective disk spaces completely and to articulate with two vertebral bodies. This results in a rib such as the third rib coming directly into the region of the T2-3 disk space so that it will articulate with both the T2 and T3 vertebral bodies.

In the middle field, the rib head once again comes directly in toward the disk space, but now it firmly attaches itself only to the disk space proper.

In the lower field, the rib head articulates directly with its corresponding vertebral body. Thus, in the lower field, the 11th rib is traced to its corresponding vertebral body and then moves directly cephalad to reach the T10-11 disk or directly caudad to reach the T11-12 disk. Once the vertebral bodies have been exposed in a skeletally immature patient, the growth cartilage of the vertebral endplate can be visualized. It has an odd tendency to appear green in color (a quirk of endoscopic optics) and is colloquially referred to as a Wolf line, in honor of Randall K Wolf.

Fusionless procedures

Fusionless techniques such as vertebral body tethering (VBT) are being evaluated in hopes of maintaining spinal flexibility and preserving mobile disk segments.  VBT corrects the curvature of the spine acutely by means of compression using a thick band called a tether.

With time, VBT, if successful, results in long-term curve correction, provided that adequate growth modulation of the instrumented vertebrae occurs.  Timing is critical to the success of this procedure: If it is done too early, it can result in overcorrection, and if it is done too late, the lack of sufficient growth modulation can lead to curve recurence if the tether eventually breaks. The US Food and Drug Administration (FDA) device exemption criteria for VBT currently include ages 8-16 years, Sanders stage 4 or less, primary thoracic curves of 35-60°, and lumbar curve of less than 35°.[139]

In addition to VBT, the use of a posterior dynamic deformity correction device (PDDD) has received a Humanitarian Device Exemption from the FDA for patients who have adolescent idiopathic scoliosis with single curves classified as Lenke 1 or Lenke 5, a Cobb angle of 35-60º that reduces to 30º or less on lateral side bending radiographs, and thoracic kyphosis less than 55º as measured from T5 to T12.[140]  Unlike VBT, this fusionless technique is performed via a posterior approach to the spine. Also, instead of compressing the curvature to obtain curve correction and growth modulations, the PDDD uses distraction across the concave side of the apex to offload the vertebral body growth plates.

Many questions remain as to the long-term results of PDDD use. European studies reported high failure rates for the first-generation device.[141] However, the device currently used in the United States was revised to address these failure concerns. Another concern is that exposing the spine in the growing child can cause a phenomenon called autofusion. To date, however, autofusion has not been reported with this device.  Growth modulation with this techique has been documented, and as with VBT, the amount of modulation is dependent on the patient's skeletal maturity at the time of device implantation.[142]

Postoperative patient management after spinal fusion includes an inpatient hospital stay, in which the goals are to work on early mobilization (out of bed on postoperative day 1), pain management, progression to normal diet, and need for transfusion monitoring. Bracing is not routinely prescribed unless there are concerns about poor bone quality that may result in the event of hardware failure.

With a good postoperative pain pathway and an uncomplicated postoperative course, most patients can be safely discharged home by postoperative day 3. Historically, bowel movements were part of the discharge critieria, but many surgeons have moved away from this and instead make sure that the patients are on a postoperative bowel regimen to prevent constipation.

After discharge, the course may vary widely from surgeon to surgeon. The important thing, however, is to prepare the patient with some general expectations. Return to school can be as early as 3 weeks for some patients but as late as 6 weeks for others. Recommendations for return to sports activity also vary widely among surgeons, with some allowing sports around 3 months and others waiting until 6 months. In addition, depending on the extent of the fusion, some surgeons may permanently restrict patients from returning to high contact/collision sports such as football.

Postoperative physical therapy is helpful and can generally start around 2-4 weeks after discharge, with the focus on increasing the patient's post-operative stamina.

Complications are of great concern to parents, patients, and surgeons. Thankfully, complications are rare with modern scoliosis surgery, despite the magnitude of these spinal deformity procedures.[59, 143] Several important intraoperative, early postoperative, and late postoperative complications are discussed here.

McKie and Herzenberg described coagulopathy as a complication of intraoperative blood salvage during scoliosis surgery.[144] These authors suggested that thrombin and Gelfoam that may have been aspirated along with salvaged blood may have contributed to the disseminated intravascular coagulation experienced by their 17-year-old patient. This effect of the thrombin and Gelfoam would have been in addition to that of hemodilution (hemodilution-induced platelet and leukocyte activation syndrome).

The importance of appropriate intraoperative spinal cord monitoring during scoliosis surgery is hardly debatable. Such monitoring can allow early recognition and treatment of spinal cord dysfunction.[145] Somatosensory and motor evoked potentials are commonly used to monitor spinal cord function. A Stagnara wake-up test may also still be employed if the surgeon desires. Current efforts at monitoring have helped achieve and maintain a very low rate of spinal cord injury (less than one half of one percent).

Some concern exists regarding postoperative activity level and the possible hazards of trauma. Neyt and Weinstein reported a case of lumbar spine fracture dislocation in a teenage boy 3 years after successful scoliosis surgery.[146] The boy's fusion extended from the second thoracic vertebra to the first lumbar vertebra, and his subsequent fracture dislocation occurred at the L2-3 level.

Delayed infections following posterior spinal fusion with Texas Scottish Rite Hospital instrumentation have been reported. Richards described 10 such patients who presented with infections at an average of about 2 years after successful spinal fusion.[147] Low-virulence organisms such as Propionibacterium acnes were the main cause, and instrumentation removal was successful in eradicating the infections. Richards hypothesized that the infections might be related to the amount of hardware (eg, hooks, rods) used and suggested that efforts at minimizing such hardware might help prevent such infections.

Much has been written regarding a particular complication called crankshaft phenomenon, which may occur after posterior spinal fusion of idiopathic scoliosis in patients who have significant anterior spinal growth remaining. Sanders et al reported that the risk of the crankshaft phenomenon was highest in patients younger than 10 years and in patients with a Risser sign of 0 with an open triradiate cartilage.[148]   However, since the implementation of selective segmental instrumentation with pedicle screws, this phenomenon has seemed to occur less often.

Significant concern exists regarding the inferior (caudad) extent of a patient's spinal fusion and its potential relation with future low back pain.[149] Connolly led a group of researchers at the Toronto Hospital for Sick Children who studied this question in 83 patients fused with Harrington instrumentation to the second, third, fourth, or fifth lumbar vertebra. At an average of 12 years (range, 10-16 y) after surgical treatment, these patients were found to have a statistically higher rate (76%) of low back pain than a control group (50%).

Some complications have been associated with anterior surgical approaches to scoliosis. For instance, chylothorax and tension pneumothorax have both been reported in association with VATS procedures.[150, 151]

Pseudarthrosis is a complication that represents a basic failure of the central intention of scoliosis surgery: bone fusion. Luckily, pseudarthrosis is very rare in modern scoliosis surgery. This is in small part due to excellent stable internal fixation (scoliosis instrumentation systems) and in large part due to proper attention to fusion technique.

Pseudarthrosis may be suggested by persistent pain, progressive deformity, or broken hardware. Previously, tomographic plain x-rays (tomograms) were commonly used to image suspected pseudarthrosis. This is no longer the case; such tomography equipment is on the endangered species list of imaging devices. Computed tomography (CT) may be helpful, but clinical suspicion and fusion mass exploration (a rare case for modern-day exploratory surgery) remain the cornerstones of pseudarthrosis diagnosis and treatment.

In 2018, the International Scientific Society on Scoliosis Orthopaedic and Rehabilitation Treatment (SOSORT) published guidelines on the use of conservative treatment approaches to idiopathic scoliosis.[106] Recommendations included the following.

Bracing

Bracing is recommended to treat adolescent idiopathic scoliosis. 

Bracing is recommended to treat juvenile and infantile idiopathic scoliosis as the first step in an attempt to avoid surgery or at least postpone it to a more appropriate age.

Bracing is recommended in patients with evolutive idiopathic scoliosis above 25° during growth; in these cases, physiotherapeutic scoliosis-specific exercises (PSSEs) alone (without bracing) should not be performed unless prescribed by a physicans expert in scoliosis.

Casting (or rigid bracing) is recommended to treat infantile idiopathic scoliosis to try stabilizing the curve. 

It is recommended not to apply bracing to treat patients with curves below 15° ± 5° Cobb, unless this is otherwise justified in the opinion of a clinician specializing in conservative treatment of spinal deformities.   

Bracing is recommended to treat patients with curves above 20° ± 5° Cobb, still growing (Risser 0 to 3), and with demonstrated progression of deformity or elevated risk of worsening, unless otherwise justified in the opinion of a clinician specializing in conservative treatment of spinal deformities. 

Very hard rigid bracing (casting) is recommended to treat patients with curves between 45° and 60° in an effort to avoid surgery.

It is recommended that each treating team provide the brace that they know best and have more experience with. No particular brace that can be recommended over the others.

It is recommended that braces be worn full-time or no less than 18 hours daily at the beginning of treatment, unless otherwise justified in the opinion of a clinician specializing in conservative treatment of spinal deformities.

Give that there is a dose response to treatment, it is recommended that the hours of bracing per day be proportionate to the severity of the deformity, the age of the patient, the stage, the aim and overall results of treatment, and the achievable compliance. 

It is recommended that daily brace wear be proportionate to the deformity severity, the age of the patient, the scoliosis stage, the aim and overall results of treatment, and the expected compliance. 

It is recommended that braces be worn until the end of vertebral bone growth and that the wearing time then be gradually reduced, unless otherwise justified in the opinion of a clinician specializing in conservative treatment of spinal deformities. 

It is recommended that the wearing time of the brace be gradually reduced, while stabilizing exercises are performed, to allow adaptation of the postural system and maintain results. 

It is recommended that any means be used to encourage compliance, including a careful adherence to the recommendations defined in the SOSORT Guidelines for Bracing Management.

It is recommended that compliance to bracing be regularly checked through compliance-monitoring devices.

It is recommended that brace quality be checked by means of an in-brace x-ray. 

It is recommended that the prescribing physician and the constructing orthotist be experts according to the criteria defined in the SOSORT Guidelines. 

It is recommended that bracing be applied by a well-trained therapeutic team that includes a physician, an orthotist, and a therapist, according to the criteria defined in the SOSORT Guidelines. 

It is recommended that all phases of brace construction (prescription, construction, check, correction, follow-up) be carefully followed for each single brace according to the criteria defined in the SOSORT Guidelines.

It is recommended that the brace be specifically designed for the type of the curve to be treated. 

It is recommended that the brace proposed for treating a scoliotic deformity on the frontal and horizontal planes take into account the sagittal plane as much as possible. 

It is recommended to use the least invasive brace suitable for the clinical situation, provided that equivalent effectiveness is retained, so as to reduce the psychological impact and ensure better patient compliance.

It is recommended that braces not restrict thorax excursion in a way that reduces respiratory function.

It is recommended that braces be prescribed, constructed, and fitted in an outpatient setting.

It is recommended that braces be regularly changed according to growth and/or specific pathologic needs as judged by a physician expert on scoliosis.

It is recommended that out-of-brace x-rays be regularly performed to check the effectiveness of bracing treatment: the number of hours out of brace before x-ray should correspond to the daily weaning time.

Physiotherapeutic scoliosis-specific exercises

Prevention of scoliosis progression during growth

PSSEs are recommended as the first step in treating idiopathic scoliosis to prevent or limit progression of the deformity and bracing. 

It is recommended that PSSEs follow the SOSORT Consensus and be based on autocorrection in three dimensions, training in activities of daily living (ADLs), stabilization of the corrected posture, and patient education.

It is recommended that PSSEs follow one of the schools whose approach has been shown to be effective in scientific studies.

It is recommended that PSSE programs be designed by therapists specifically trained in the approach they use.

It is recommended that PSSEs be proposed by therapists included in scoliosis treatment teams, with close cooperation among all members.

It is recommended that PSSEs be individualized according to patient needs, curve pattern, and treatment phase. 

It is recommended that PSSEs always be individualized, even if performed in small groups.

It is recommended that PSSEs be performed regularly throughout treatment to achieve best results. 

It is recommended that therapists implement a compliance system for exercise tracking. 

It is recommended that therapists regularly assess the quality of PSSEs performed by patients.  

It is recommended that PSSE difficulty be progressively increased according to patient ability. 

It is recommended that PSSEs be taught individually in a one-to-one relationship to ensure individualized care; regular performance could also be at home or in small groups.

Use during brace treatment and surgical therapy

1. It is recommended that PSSEs are performed during brace treatment.

2. It is recommended that, while treating with PSSEs, therapists work to increase compliance of the patient to brace treatment.

3. It is recommended that spinal mobilization PSSEs are used in preparation to bracing.

4. It is recommended that stabilization PSSEs in autocorrection are used during brace weaning period.

5. It is recommended that PSSEs in painful operated patients are used to reduce pain and increase function.

6. It is recommended that aerobic physiotherapy training be used prior to surgery.

Other conservative treatment

It is recommended that manual therapy (gentle short-term mobilization or soft-tissue-releasing techniques) be proposed only if it is associated with stabilization PSSEs, unless otherwise justified in the opinion of a clinician specializing in conservative treatment of spinal deformities.

It is recommended that correction of real leg-length discrepancy, if needed, be decided on by a clinician specializing in conservative treatment of spinal deformities.