Scoliosis represents a complex three-dimensional deformity of the spine and thorax.^1,² Scoliosis is defined as a coronal deformity of the spine with a Cobb angle measuring greater than 10 degrees (Fig. 31.1).³ Small-magnitude curves may progress during periods of skeletal growth.⁴ Larger curves may continue to increase, even in adulthood.^5,⁶ Thus, a pediatric patient with scoliosis should be followed until skeletal maturity, and young adults with large untreated curves should be monitored periodically for curve progression.

FIGURE 31.1 Measurement of the Cobb angle to determine the coronal magnitude of the curve. The two most tilted vertebrae are identified, and a line is drawn marking the angle of the end plate. The angle between the end plates is then measured. To facilitate measurement of the angle, it is also possible to drop a perpendicular line from the end plate and measure the angle subtended by the perpendiculars.

Congenital, neuromuscular, and idiopathic curves are commonly diagnosed in childhood. Congenital scoliosis results from malformation of the vertebral elements, either due to failure of formation or failure of segmentation. A hemivertebra with a contralateral bar results in the greatest risk of curve progression.⁷ Particularly in small children, it is difficult to predict which congenital curves will be progressive and will subsequently require surgical intervention. Because of a high associated risk, all children with congenital scoliosis require evaluation for concomitant renal, cardiac, and intrathecal abnormalities. Owing to the structural abnormalities, bracing in congenital scoliosis is typically not indicated. Neuromuscular curves are those associated with an underlying neurological difference, either structural (Chiari malformation, syrinx, tethered cord, diastematomyelia) or systemic (e.g., cerebral palsy, Charcot-Marie-Tooth disease, muscular dystrophy, neurofibromatosis). As idiopathic scoliosis is a diagnosis of exclusion, neurological causes for the deformity must be ruled out by medical history, clinical examination, radiographs, and in selected cases, axial imaging.

Idiopathic scoliosis is the most common type of scoliosis in children. Based on age at diagnosis, idiopathic scoliosis is traditionally classified into infantile (age less than 3 years), juvenile (age 3 to 10 years), and adolescent (age greater than 10 years). More recently, the term early-onset scoliosis has been used for spinal deformities in children under the age of 5 years, including both idiopathic, congenital, and neuromuscular curves, in order to acknowledge the special challenges of treating very young children with scoliosis.⁸ Infantile neuromuscular scoliosis is frequently progressive, but if the curve is flexible it may be effectively managed with bracing and casting (Fig. 31.2).⁹ Idiopathic infantile scoliosis is more common in males than females, and patients may present with a left-sided curve. Factors associated with progressive infantile idiopathic scoliosis include age at treatment, rib vertebral angle greater than 20 degrees, a phase II rib, or a curve greater than 60 degrees.⁹^–¹¹

FIGURE 31.2 A, A 14-month-old child presenting with 70 degrees right thoracic curvature. B, Significant improvement in standing film in Mehta cast. Long-term results of this technique in large magnitude curves are unpredictable, although this serves as an effective delay tactic.

Although the cause of idiopathic scoliosis is unknown, family members of affected patients have an increased risk of developing scoliosis, reflecting a genetic component to the disease.¹²^–¹⁴ Several genes have been associated with familial scoliosis.^15,¹⁶ The incidence of adolescent idiopathic scoliosis (AIS) with a curvature greater than 30 degrees ranges from 1 to 3 in 1000 children.¹⁷^–²³ Small curves are common in both males and females, although progression is more common in female children.

Natural history studies of scoliosis reveal the greatest morbidity from early-onset scoliosis, with higher mortality rates and decreased pulmonary function for thoracic curves.²⁴ One series of patients with untreated juvenile and adolescent scoliosis revealed only one death from cor pulmonale, and overall mortality rates similar to those for the average population.⁶ Back pain could not be correlated with the severity of the deformity, although patients with lumbar curves had more back pain.⁶ Patients with thoracic curves have been noted to have poor pulmonary function because of decreased lung volumes from the coronal plane deformity.^6,²⁵ Thoracic curves measuring 70 degrees may result in noticeable changes on pulmonary function testing, and curves greater than 100 degrees may result in clinical symptoms.²⁴

Curves are at risk for progression during periods of rapid skeletal growth, specifically the first 2 years of life and the adolescent growth spurt.⁴ There is increased risk of curve progression for large curves and for skeletally immature patients (Table 31.1).⁴ Bone age, peak height velocity, Risser sign, and elbow and hand maturity indices may aid in determining risk of curve progression.²⁶^–³⁰ Active investigation is under way to identify genetic markers that would assist in predicting curve progression.³¹ For adolescent idiopathic scoliosis, thoracic and double curve patterns are at higher risk of progression than thoracolumbar curves.⁴ Curves less than 30 degrees at skeletal maturity are thought to be mostly stable and unlikely to progress over time.^5,³² Curves greater than 50 degrees at skeletal maturity commonly progress at a rate of up to 1 degree per year for thoracic curves.^5,³² Curves between 35 and 50 degrees at skeletal maturity present a diagnostic dilemma due to insufficient clinical data and should be followed radiographically into adulthood.

TABLE 31.1 Adolescent Idiopathic Scoliosis Treatment Algorithm

Growing Child

Open triradiate cartilage, Risser 0, 1, peak height velocity of 8-10 cm/year growth⁴²

Curves 5-19 degrees have a 22% chance of progression.⁴

Curves 20-29 degrees have a 68% chance of progression.⁴

Curve Size^∗	Treatment	Follow-up Interval
≤25 degrees	Observation	4-6 months
25-45 degrees	Bracing (may consider in Risser 0 at 20-25 degrees with documented progression)	4 months
>45 degrees	Surgery—consider anterior release with open triradiate cartilage

Adolescent Nearing Maturity

Closed triradiate, Risser 3-4, postmenarchal females

Curves 5-19 degrees have a 2% chance of progression.

Curves 20-29 degrees have a 23% chance of progression.

Curve Size^∗	Treatment	Follow-up Interval
≤30 degrees	Observation	12 months to document no progression
35-45 degrees	Consider bracing, efficacy not determined	4 months if braced 6-12 months otherwise
>50 degrees	Surgery

Skeletally Mature Child

Risser 5, menarche + 2 years in females, cessation of growth^5,⁶

Large thoracic curves >50 degrees may progress up to 0.5-1 degree/year.^5,⁶

Lumbar curves >40 degrees may progress.⁵

Curves <30 degrees are unlikely to progress.⁵

Curve Size^∗	Treatment	Follow-up Interval
≤50 degrees	Observation	1-2 years
>50 degrees	Surgery

Note: Superscript numbers refer to references for this chapter, listed online.

∗ Measurement variability for Cobb angle is reported at ±5-6 degrees.^75,⁷⁶

Indications

Bracing and surgery are the mainstays of treatment for scoliosis. Treatment must be tailored to the patient and family. Bracing is typically indicated in growing children with progressive curves measuring 20 to 45 degrees.³³ In attempts to prevent curve progression, bracing is most effective in skeletally immature patients (Risser 0 or 1, 2) and in females who are less than 12 months past menarche.³³ Numerous types of braces have been described,³⁴^–³⁸ but a TLSO (thoracic lumbar spine orthosis) is commonly used.³³ Brace wear has been shown to be efficacious if the brace is worn.^39,⁴⁰ The brace is discontinued at the completion of growth. Casting is well tolerated in the young child and may be a preferable technique over bracing owing to derotational molding of the deformity and improved compliance.^9,¹⁰ It has been shown to be curative in some cases of infantile scoliosis, although in some instances these curves are self-resolving.⁹

When the curve magnitude exceeds 50 degrees in AIS, surgical treatment is offered, typically with posterior spinal fusion. Fusion of smaller curves can be considered in skeletally immature patients who have been documented to have significant progression or in patients who find the appearance and trunk shift unacceptable. Pain is not typically an indication for surgery. Patients with severe deformity or who are skeletally immature (open triradiate cartilages) may also be treated with anterior release or posterior fusion to improve correction and prevent subsequent rotational and coronal or “crankshaft” deformity.⁴¹^–⁴⁴ In the era of pedicle screw fixation, the necessity of anterior procedures is debated in the literature and may be a matter of surgeon preference.⁴⁵ Anterior release, fusion, and instrumentation in the thoracic spine may be performed through a thoracotomy or through a thoracoscopic approach.^46,⁴⁷ Anterior procedures, particularly open surgery, may result in decreased pulmonary function.⁴⁶^–⁴⁸ Large curves and complex deformities can also be addressed by osteotomies through a posterior approach to achieve satisfactory correction. These procedures may include, in order of increased correction and neurological risk, Ponte/Smith-Peterson osteotomies, pedicle subtraction osteotomies, and vertebral column resections.⁴⁹^–⁵¹

Neuromuscular scoliosis is guided by similar treatment principles. Progressive imbalance from muscle weakness or spasticity may result in a sweeping thoracolumbar or C-type curvature and subsequent pelvic obliquity, truncal shift, and difficulty with wheelchair seating and caregiving.^52,⁵³ The role of bracing in the neuromuscular population is not well defined, but may be a helpful delay tactic. Custom wheelchair seating systems can also serve as an external brace in many instances. Surgical management is indicated when bracing or wheelchair positioning can no longer accommodate the patient’s deformity and the curve progresses to greater than 50 degrees (Fig. 31.3).⁵² Smaller curves may continue to progress in adult patients with neuromuscular scoliosis, although curves greater than 50 degrees increase at a faster rate.⁵⁴ Patients with underlying cardiac or pulmonary deficits as classically seen in Duchenne muscular dystrophy may benefit from early surgery performed for smaller magnitude curves rather than waiting for the curve to progress while the patient’s medical status deteriorates.⁵⁵ New corticosteroid treatment protocols for Duchenne muscular dystrophy may redefine these indications.⁵⁶ Surgical treatment of nonambulatory patients typically involves fusion from the upper thoracic spine to the pelvis, although recent literature has shown improved mobility in paraplegic children fused to L5. Long-term results from this technique are not established. In ambulatory patients with an underlying neuromuscular condition, longer fusions are typically indicated to avoid subsequent decompensation.

FIGURE 31.3 Patient with neuromuscular scoliosis preoperatively and following posterior spinal fusion from T3 to the sacrum with pedicle screw instrumentation.

Early-onset scoliosis has unique considerations. Thoracic fusion in a small child is unacceptable and may later lead to thoracic insufficiency syndrome and significant shortening of the trunk. Delay tactics include casting, bracing, and halo gravity traction.⁹ When curves do not respond to these measures, surgical intervention may be undertaken to prevent curve progression and allow for spinal growth.^57,⁵⁸ Several growing spine systems are in use, including growing spine-based and rib-based devices.^57,⁵⁹^–⁶² There is limited consensus as to the indications and applications of these devices. These are seen as a last resort because the child is then committed to twice yearly lengthening procedures and a high rate of complications, including infection, wound problems, early unintended fusion, implant failure, and rarely neurological complications.⁵⁷ A planned definitive fusion is performed at the completion of the lengthening procedures. For growing rods, a dual-rod construct has been shown to have fewer complications and improved correction, although a single rod can be used in conjunction with a brace.^60,⁶³ The implants are frequently quite prominent. Prior to initiating surgical treatment, nutritional and medical status must be optimized. Large curves result in increased work of breathing and energy expenditure. Thus, children with significant deformity are often malnourished and require dietary supplementation and in many instances gastrostomy tube placement to improve soft tissue coverage and reduce wound complications.

Diagnosis

Scoliosis is typically painless in the pediatric population, and children for the most part are asymptomatic. Mild back pain, however, is not uncommon. Screening examinations by pediatricians or school nurses frequently result in referrals. Parents may notice shoulder, scapula, or waist asymmetry, particularly if they are attuned to the diagnosis by a family history of scoliosis.

All patients with presumed idiopathic scoliosis should be evaluated with a medical history, complete neurological examination, and radiographs. A family history should be taken for scoliosis or a history of underlying neurological abnormalities. Females should be asked about age at menarche, because this landmark is an excellent predictor of skeletal maturity. On physical examination, the patient’s back should be carefully examined for asymmetry, trunk shift, and sagittal plane alignment. AIS is balanced in the coronal plane and typically hypokyphotic, so the presence of kyphosis or excessive truncal shift is an indication for axial imaging ⁶⁴ (Fig. 31.4). Asymmetrical abdominal reflexes may be associated with a neural axis abnormality.⁶⁵ The Adams’ forward bend test may reveal a thoracic or lumbar rib prominence resulting from rotational deformity. The pelvis should be assessed for an occult leg-length discrepancy, which may give the appearance of scoliosis. The skin should be examined for masses, café au lait patches, and axillary or inguinal freckling indicative of neurofibromatosis. A hairy patch, dimple, or pigmentation over the lumbosacral region suggests spinal dysraphism. The patient should be examined for excess ligamentous laxity or pectus deformity, raising the possibility of an underlying syndrome.

FIGURE 31.4 A, Posteroanterior (PA) standing scoliosis film in a patient with adolescent idiopathic scoliosis. B, Lateral spine film reveals hypokyphosis over the region of the curvature. C, Postoperative PA and, D, lateral views following pedicle screw fixation.

Neural axis abnormalities are found in 13% to 26% of patients with juvenile scoliosis and 2.6% to 14% of patients with adolescent-onset scoliosis.⁶⁵^–⁶⁷ Up to 50% of patients with a left thoracic rather than right thoracic curve may have an underlying neural axis abnormality.⁶⁷^–⁶⁹ As many as 20% to 50% of children with infantile scoliosis may have an underlying neural axis abnormality.^67,⁷⁰ No difference in treatment outcome has been reported based on the presence of a neurological abnormality, although a large syrinx, tethered cord, or Chiari malformation may require surgical intervention (Fig. 31.5).⁶⁹ In some instances, scoliosis may improve after treatment of the neural axis abnormality, but this is not predictable.

FIGURE 31.5 Scoliosis with a syrinx. A, Posteroanterior (PA) scoliosis film of a female patient presenting with deformity, back pain, and headaches. B, Magnetic resonance (MR) image of the spine reveals a syrinx and Chiari malformation.

Standing posteroanterior (PA) and lateral spine images should include the neck, clavicles, and pelvis. A lateral image should be taken at presentation in order to document the sagittal plane alignment and rule out congenital deformity. Subsequently, the curve can be followed by standing PA imaging unless there is a concern for kyphoscoliosis.

The idiopathic scoliosis curve types have been classified by King and associates ⁷¹ and more recently by Lenke and colleagues (Fig. 31.6).⁷² These classifications assist in choosing fusion levels and help identify idiopathic curve types for research purposes. With the Lenke classification, premeasured radiographs have high inter- and intrarater reliability,^72,⁷³ although there is only poor to good interobserver reliability on unmeasured radiographs.^73,⁷⁴ Cobb angle measurement error among different observers is 5 to 6 degrees; thus, more than 5 degrees increase in curve size is necessary in order to document curve progression.^75,⁷⁶ Digital imaging systems may improve measurement accuracy.^75,⁷⁶ Children who may be candidates for treatment should be followed with radiographs several times annually during periods of rapid growth.

FIGURE 31.6 Lumbar modifier for the Lenke classification. The lumbar modifier is determined by the relationship of the pedicles in the lumbar curve to the central sacral vertebral line (CSVL).

Evidence-Based Medicine

Treatment for scoliosis includes observation, bracing or casting, and surgery. There are no studies that support the use of activity modification, manipulation, exercises, physical therapy, or electrical stimulation for the correction of scoliosis.⁷⁷^–⁸⁰ Casting has been shown to be efficacious for infantile curves at high risk of progression. Braces, if worn consistently, may prevent curve progression.^39,⁴⁰

Surgical treatment is undertaken to prevent further curve progression, but has the added benefit of deformity correction. The importance of specific correction objectives are debated among surgeons and include coronal plane correction, restoration of thoracic kyphosis and lumbar lordosis, sagittal balance, and vertebral derotation.⁸¹ Beginning with Harrington rods, a variety of implants have been used including hooks, wires, and screws.⁸² Pedicle screws are now the gold standard due to no violation of the canal, superior rotational and curve correction, resulting improvements in pulmonary function tests (PFTs).⁸²^–⁸⁵ Pedicle screws can be placed using anatomical, fluoroscopic, or computed tomography (CT)-guided navigation.⁸⁶ The role of intraoperative CT is under development, but may be of great assistance in complex deformity and congenital cases when the anatomical landmarks are unreliable. Posterior spinal fusion for scoliosis results in improved patient perception of appearance and improved functional outcome scores. Long fusions limit spine motion and have the theoretical risk of increased stress and degenerative changes at adjacent motion segments.⁸⁷

Selection of fusion level and density of implants is subject to significant variability in practice and is beyond the scope of this text.^71,^88,⁸⁹ Definitions of the end, neutral, and stable vertebrae may aid in selection of fusion levels (Fig. 31.7). Occasionally, levels can be preserved in thoracolumbar curves by using an anterior versus posterior approach. Bracing or casting is not routinely recommended postoperatively because new fixation techniques are quite sturdy and reliable.

FIGURE 31.7 End, neutral, and stable vertebrae. The end vertebrae (EV) are the most tilted vertebrae, which are selected for measurement of the Cobb angle. The neutral vertebra is the most cephalad vertebral body below the end vertebra with no rotation (pedicles appear symmetrical). The stable vertebra (SV) is the most cephalad vertebral body below the end vertebra which is bisected by the central sacral vertebral line (CSVL).

Complications can be limited by diligent preoperative planning and clear communication with the medical and anesthesiology teams. The complication rates for posterior spinal fusion is reported at 5% to 15% for idiopathic scoliosis patients ⁹⁰^–⁹² and 28% to 33% for neuromuscular scoliosis patients.^93,⁹⁴ Higher complication rates are associated with prolonged surgical and anesthetic time, increased blood loss, hybrid constructs, and renal disease.^90,⁹² Growing spine patients face a 48% complication rate.⁶⁰ All patients should have blood products available. Tranexamic acid may help reduce blood loss.⁹⁵ Thoracic surgery should be performed with intraoperative somatosensory evoked potentials (SSEPs) and motor evoked potentials (MEPs) in collaboration with a neuromonitoring team and neurologist.⁹⁶ These tests require special anesthetic techniques such as total intravenous anesthesia. The Stagnara wakeup test can be used in instances when signals cannot be obtained in the patient and serves as a secondary method to assess neurological function intraoperatively. From an anesthesia standpoint, significant lead time may be required. The airway should be carefully protected, and the wound should be filled with saline to prevent pulmonary air embolus from deep inspiration at the time of the wakeup test. Forceful injection of thrombin products into the pedicle tract has been reported to cause neurological injury and should be avoided.⁹⁷ Neurological deficit can present in a delayed fashion postoperatively, likely due to vascular effects on the spinal cord.⁹⁸ Cases with osteotomies, high blood loss, and preoperative kyphosis or myelopathy are at greater risk of neurological complication. Thus, for high-risk procedures, careful postoperative monitoring, maintenance of blood pressure with fluid or in some instances with pressors, and adequate maintenance of hemoglobin should be ensured.

Conclusion

Scoliosis is the most common spinal deformity in the pediatric population. Advances in instrumentation and monitoring have rendered posterior spinal fusion a safe, reliable treatment to provide deformity correction and prevent curve progression in adulthood. Further work is ongoing about the role of intraoperative imaging. Early-onset scoliosis is a field with a wide variety of treatment techniques and limited consensus on the optimal surgical indications or implants. Although significant progress has been made in the treatment of scoliosis, much work remains to be done, both to better understand the cause of scoliosis and to improve the surgical treatment of the disorder.

Part II. Adult Spinal Deformity: Considerations in Evaluation and Treatment

Spinal deformity in the adult is increasingly recognized as a significant health care issue. With an aging population there are also increasing functional expectations of an older population in Western societies. The costs of caring for the elderly who lose autonomy due to spine deformity are substantial. Pain and disability are closely associated with spinal deformity in the adult population. One prevalence study ⁹⁹ reports a 60% rate of spine deformity in a population over 60 years of age.

The treatment of adult spinal deformity (ASD) has received rather sparse attention compared to deformity in the pediatric population. Despite the large number of adults affected by spine deformities and the high potential for marked loss of function,¹⁰⁰ adult deformity remains poorly understood. Progress in research is hampered by the diversity of disorders associated with spinal deformity in the adult, the lack of a coherent system for classifying patients, and until recently, poorly understood correlations between deformity parameters and patient-reported disability.

This section of the chapter will summarize the efforts to date that have determined important correlations between health-related quality of life (HRQOL) measures and radiographic parameters as well as outline classification approaches and alignment guidelines related to ASD. The current state of the art will be presented as well as ongoing work that is certain to impact our management of the often complex situation in adult patients affected by a spinal deformity.

Key Radiographic Parameters and Health-Related Quality of Life

Although ASD has a wide range of causes, most cases appear to relate to the degenerative process of aging (de novo degenerative spinal deformity, or DDS) or adolescent scoliosis in an adult (ASA).¹⁰¹ Other causes of deformity include osteoporosis, trauma, infection, and iatrogenic factors. For DDS and ASA, the two largest categories, a common pathway appears to lead to pain and loss of function.

Many clinicians have investigated the regional and global alignment in the normal (asymptomatic) adult population ¹⁰²^–¹⁰⁸ (Fig. 31.8). These data are helpful in understanding normative ranges but offer limited value in understanding disability related to ASD. However, a study by the Spinal Deformity Study Group (SDSG),¹⁰⁹ including 947 adults with spinal deformity, set out to answer this key question. Data from their multicenter prospective database were analyzed for HRQOL/radiographic correlation. All patients had radiographic analysis: frontal Cobb angle, deformity apex, lumbar lordosis, intervertebral subluxation. Health assessment included the Oswestry Disability Index (ODI) and Scoliosis Research Society instrument (SRS). Patients were classified by deformity apex, lordosis (T12-S1), and intervertebral subluxation. The correlation analysis revealed that HRQOL was driven by several key parameters (in the following order): global sagittal alignment (sagittal vertical axis, SVA), loss of lumbar lordosis (LL), intervertebral subluxation. This data set did not include pelvic parameters, which subsequently emerged as being critical to understanding ASD. In a recent investigation by Lafage and co-workers¹¹⁰ radiographic pelvic parameters were established as an integral component in patient-reported function in the setting of spinal deformity.

FIGURE 31.8 Sagittal spinal radiological parameters. Thoracic kyphosis is measured from the superior end plate of T4 to the inferior end plate of T12. Lumbar lordosis is measured from the superior end plate of L1 to the superior end plate of S1. Sagittal vertical axis (SVA) is defined as the horizontal offset from the posterosuperior corner of S1 to the vertebral body of C7. T1 spinal-pelvic inclination and T9 spinal-pelvic inclination are defined as the angle between the vertical plumbline and the line drawn from the vertebral body center of T1 or T9 and the center of the bicoxofemoral axis.

Several authors have described the pelvis as a regulator of the sagittal plane alignment, and extensive studies have been conducted to understand the relationship between pelvic parameters and spinal alignment. Radiographic parameters of pelvic alignment (Fig. 31.9) have now emerged as essential components of the global standing alignment,^108,¹¹¹^–¹¹⁴ and their clinical relevance is briefly summarized here:

Pelvic tilt (PT) is defined by the angle between the vertical and the line through the midpoint of the sacral plate to the femoral head axis; it is commonly reported as a compensatory mechanism: when the spine tilts forward (age-related change, sagittal imbalance, loss of lordosis, increase of kyphosis) the subject will try his or her best to maintain an economic posture and to keep the spine as vertical as possible (i.e., bring the spine over the pelvis). One way to maintain this spinal-pelvic alignment is to retrovert the pelvis (increase of pelvic tilt), which may be seen as a backward rotation of the pelvis around the hips.

Pelvic incidence (PI) is defined as the angle between the perpendicular line to the sacral plate at its midpoint and the line connecting this point to the femoral head axis. This is a morphological parameter of primary importance because of its correlation with ideal lumbar lordosis. In a simplified formula lumbar lordosis should be within 9 degrees of PI (LL = PI ± 9) degrees.

Sacral slope (SS) is defined as the angle between the horizontal and the sacral plate; this last parameter completes the geometric relationship where PI = PT + SS. Normative values and age-related changes of the pelvic parameters are provided (Table 31.2).

FIGURE 31.9 Pelvic parameters. Pelvic tilt is defined by the angle between the vertical and the line through the midpoint of the sacral plate to the femoral head axis (retroversion is then measured as a pelvic tilt increase, anteversion as a pelvic tilt decrease). Sacral slope is defined as the angle between the horizontal and the sacral plate. Pelvic incidence is defined as the angle between the perpendicular to the sacral plate at its midpoint and the line connecting this point to the femoral head axis. Pelvic incidence = Pelvic tilt + Sacral slope.

TABLE 31.2 Normative Values for Radiographic Sagittal Key Parameters

Classification

To date, plain x-ray films remain the basis for diagnostic approaches to many orthopedic disorders. Although newer imaging modalities have emerged, substantial variation in interpretation of information related to spinal diseases exist. In terms of establishing a classification of adult deformity it may be ideal to base this principally on x-ray image information to remain simple in form and reliable in application.

Given that treatment for adults with deformity is driven by the patient who wishes to regain function without pain, a classification of adults must take this as a basis upon which to build categories. It is not reasonable to transpose classification efforts applied to AIS because the treatment concerns are of a very different nature. Pain and disability are rarely noted in AIS.

A significant advance in the classification of scoliotic deformity is attributed to Lenke and colleagues.⁷² In a 2001 publication, this classification offered an updated approach to AIS which was more comprehensive than the King classification⁷¹ and offered sagittal plane consideration. Furthermore, the new AIS classification offered guidelines on surgical fusion levels using segmental fixation.

The classification of adult scoliosis ¹¹⁵ is clearly more complex than approaches to AIS. Adult deformity covers a broad range of segmental, regional, and global expression (Fig. 31.10). The treatment is driven by disability and pain and not simply by radiographic appearance. The latter has thus made it difficult to arrive at a clinically relevant x-ray classification. One of the first studies aimed at isolating the pain generator in adult scoliosis was published in 2002.¹¹⁶ This work by Schwab and associates launched an ongoing effort at reconciling HRQOL measures and radiographic presentation of adult deformity. The approach has led to the Schwab-SDSG classification of which the basic structure hinges on the following:

Basic coronal curve locations and basic patterns (single vs. multiple)

A principally sagittal plane deformity category

Regional and global modifiers with three grades tied to HRQOL

FIGURE 31.10 Adult scoliosis. Preoperative PA (A, B) and lateral views (C, D) show both coronal and sagittal plane deformity.

All aspects of the classification found a basis in HRQOL analysis. The cutoffs between groups within each modifier were determined by the HRQOL measures, splitting the population into discrete groups by clinical impact of each modifier parameter. Description of the classification is summarized in Table 31.3.

TABLE 31.3 Schwab-SDSG Adult Spinal Deformity Classification

Type	Location of Deformity^∗
Type I	Thoracic-only scoliosis (no thoracolumbar or lumbar component)
Type II	Upper thoracic major, apex T4-8 (with thoracolumbar or lumbar curve)
Type III	Lower thoracic major, apex T9-T10 (with thoracolumbar/lumbar curve)
Type IV	Thoracolumbar major curve, apex T11-L1 (with any other minor curve)
Type V	Lumbar major curve, apex L2-L4 (with any other minor curve)
Type K	Deformity in the sagittal plane only
Lordosis Modifier	Sagittal Cobb Angle: T12 to S1
A	Marked lordosis >40 degrees
B	Moderate lordosis 0-40 degrees
C	No lordosis present <0 degrees
Subluxation Modifier	Maximum Value: Frontal or Sagittal Plane (Anterior or Posterior)
0	No subluxation
+	Subluxation 1-6 mm
++	Subluxation >7 mm
Global Balance Modifier^†	C7 Offset From Posterior Superior Corner of S1: Sagittal Plane
N	0-4 cm
P	4-9.5 cm
VP	>9.5 cm

SDSG, Spinal Deformity Study Group.

∗ Apical level of the major curve or sagittal plane only.

† N, normal; P, positive; VP, very positive.

Alignment Goals When Treating Adult Spinal Deformity

Spinal-pelvic realignment is a complex undertaking. Other than decompression and stabilization of spinal segments (which is beyond the scope of this chapter) realignment of the sagittal spinal-pelvic axis is of primary importance during surgical treatment of symptomatic ASD. As outlined previously, normative data from an asymptomatic population may provide insight into ranges for regional and global alignment during standing posture. However, for optimal clinical outcomes, treatment should be adapted to a given individual based upon specific realignment needs.

Tailoring patient-specific treatment involves the crucial PI-LL relationship, wherein ideal lumbar lordosis is proportional to pelvic morphology quantified as pelvic incidence (PI). Additionally, alignment parameters driving pain and disability, namely, sagittal vertical axis (SVA) and PT, need to be addressed. The surgical method of realignment (i.e., pedicle subtraction osteotomy versus a Smith Peterson osteotomy, intervertebral spacer, rod contouring) remains of secondary importance to the primary goal: obtaining surgical realignment objectives. As a general concept and pragmatic tool for clinical application, spinal-pelvic realignment objectives involve attention to the following three key parameters ¹¹⁷ (Fig. 31.11):

SVA: Global spinal realignment should attempt to obtain a postoperative SVA less than 50 mm. Restoration of SVA facilitates level gaze and achievement of a physiological standing posture. An SVA less than 50 mm brings the plumbline behind the femoral heads to relieve the complaint of “falling forward.” Clinically, this threshold has been met with better HRQOL scores. Similarly, the reference T1 − SPI < 0 degrees may be used. Both parameters outline the same principle.

PT: Pelvic realignment should attempt to obtain a postoperative PT less than 25 degrees. Attention to PT, as outlined by clinical data, is necessary to obtain optimal outcomes. Additionally, PT realignment restores appropriate femoral-pelvic-spinal alignment required during efficient ambulation (need an extension reserve to clear the step). Realignment of the SVA less than 50 mm in the setting of an elevated PT means the ASD patient is still compensating for residual structural spinal deformity. This parameter independently has been shown to correlate to impairment in walking tolerance and therefore should be realigned appropriately.

LL = PI ± 9 degrees: Finally, to achieve patient-specific alignment treatment LL = PI ± 9 degrees may pragmatically be used. Increasing the angulation of the hypolordotic spine to match the patient’s spinal-pelvic morphological type (i.e., PI) assures appropriate lordotic alignment. This chain of correlation has been extensively studied by many authors.^108,^111,^118,¹¹⁹ as it relates to the subject’s morphological type.¹¹³

FIGURE 31.11 Realignment objectives in the sagittal plane. SVA <50 mm, PT <25 degrees, and LL = PI ± 9 degrees sets the stage for achievement of a successful harmonious spinal-pelvic realignment. LL, lumbar lordosis; PI, pelvic incidence; PT, pelvic tilt; SVA, sagittal vertical axis.

Predicting Outcome Related to Surgery for Adult Spinal Deformity

In building predictive models of outcome from surgery it is key to identify what combination of patient characteristics (including Schwab classification modifiers) and treatment options can be used to predict which patients will meet an outcome threshold for success.¹²⁰ Previous work by the SDSG on minimal clinically important difference (MCID) has identified improvement thresholds for discriminating between patients who report being satisfied with surgical results and those who report being less than satisfied. Based upon those identified thresholds, binary logistic regression was used to examine how adult classification types and surgical strategy combine and interact to predict successful surgical outcomes. A second approach was to employ multiple linear regressions. For both binary and logistic regression, predictor variables included all classification types, all treatment options within surgery, patient age, gender, BMI (body mass index), previous surgery, and baseline health status. Backward and stepwise techniques were used to eliminate redundant predictor variables.

An analysis of patients most and least likely to reach an MCID can be drawn from the predictive models. It was found that in addition to the classification modifiers, baseline HRQOL was associated with higher chance of a poor outcome (Tables 31.4 and 31.5).

TABLE 31.4 Strength of Predictive Models ¹ Meeting MCID Threshold² ^∗

Outcome Score¹	% Correct Classification by Model	Area Under ROC Curve
SRS pain	81.10%	0.864
SRS appearance	75.40%	0.838
SRS pain and appearance	78.10%	0.845
SF-12v2 PCS	77.90%	0.862

MCID, minimal clinically important difference; SF-12v2 PCS, Medical Outcomes Study, short-form, version 2, physical component summary; ROC, receiver operating characteristic; SRS, Scoliosis Research Society.^1,²

∗ 0.80 and above is considered good discrimination.

TABLE 31.5 Summary of Post–Scoliosis Surgery Treatment Groups Likely to Reach MCID (Success) or Unlikely to Do So (Poor Outcome)

	Groups Less Likely to Reach Threshold Improvement	Groups With Higher Chance of Reaching Threshold Improvement
SF-12v2 PCS MCID summary	Apical level III Marked lordosis No subluxation Negative sagittal balance Baseline PCS ≥35	Apical level IV Subluxation + or ++ Positive sagittal balance Surgery involved osteotomy Surgery involved fixation to sacrum Baseline PCS <35
SRS combined pain/appearance MCID summary	Apical level III Marked lordosis No subluxation Negative sagittal balance	No lordosis Subluxation ++ Circumferential surgery Surgery involved osteotomy
SRS pain MCID summary	Apical level III Marked lordosis No subluxation Negative sagittal balance Posterior-only surgical approach No fixation to the sacrum	Apical level IV Subluxation ++ Circumferential surgery Surgery involved osteotomy Baseline PCS <35
SRS appearance MCID summary	Apical level V Surgery involved fixation to sacrum	No previous surgery Apical level IV Baseline PCS ≥35

MCID, minimal clinically important difference; SF-12v2 PCS, Medical Outcomes Study, short-form, version 2, physical component summary; SRS, Scoliosis Research Society.