Natural History of Untreated Adolescent and Late-onset Idiopathic Scoliosis

Approximately half a million adults in the United States have spinal curves >30 degrees.4,5 However, the decision to operate on patients with LIS should be based on the premise that the surgery will ease pain and improve function as compared with the natural history of the disease, and will have a relatively low incidence of adverse sequelae. To better understand the treatment of LIS, one must first understand the natural history of untreated idiopathic scoliosis, because many patients with LIS have had undiagnosed or untreated AIS. Another category of LIS comprises patients who have had surgical management of AIS but who have developed symptoms as a consequence of this or have developed degeneration or decompensation in unfused segments of the spine.

Many cases of untreated AIS are asymptomatic or minimally symptomatic throughout adulthood and never require any form of medical management. Several studies have disproved the notion that all types of idiopathic scoliosis inevitably end in disability. In 1969, Collis et al6 reviewed a series of 215 untreated patients with LIS and 100 controls with more than 20 years of follow-up. In 71% of this group, spinal curvatures were >50 degrees. These patients were not treated and were found after more than 20 years of follow-up to have productive lives with minimal disability from their scoliosis.6 Korovessis and co-workers7 followed 91 patients with lumbar curvatures >10 degrees for more than 2 years, and were able to identify multiple risk factors for progression of their curvature. Patients at the greatest risk for progression had curves >30 degrees, >30% apical vertebral rotation, 6 mm or more of lateral listhesis, and degenerative disc disease at the lumbosacral junction.

In a recent landmark prospective study, Weinstein and co-workers3 compared 117 patients who had had untreated AIS with 62 age- and sex-matched volunteers at a 50-year follow-up. The mean age of the patients was 66 years. The main outcomes assessed were mortality, back pain, pulmonary symptoms, general function, depression, and body image. There was no significant difference in survival in the patient and volunteer groups. However, there was an increased likelihood of difficulty in breathing in the scoliosis group, which was statistically significant for patients with a Cobb angle >80 degrees and a thoracic apex of their curvature. Sixty-one percent of the scoliosis patients reported chronic back pain, as compared with 35% of the controls (P = 0.003). There was no significant difference in clinical depression in the two groups, but body satisfaction in the scoliosis patient group was significantly poorer than in the control group.3

Additionally, Weinstein and coworkers’ study showed that coronal curves progress longitudinally and that almost 70% of patients with untreated AIS have progression of their curves after skeletal maturity. Thoracic curves >50 degrees progressed on average by 1 degree per year; however, thoracic curves <30 degrees did not show a propensity to progress over time. Thoracolumbar (TL) curves progressed by ~0.5 degrees per year, whereas purely lumbar curves progressed by 0.24 degrees per year.3

Although many adult patients with LIS experience few symptoms, a subpopulation of patients develops significant chronic symptoms that respond poorly to nonsurgical measures. Dickson et al8 described 81 patients with adult idiopathic scoliosis who underwent surgery, and compared them with 30 patients who refused surgery. The treated patients had “significantly reduced pain (and) fatigue and increased function. ” Eighty percent of these patients reported some pain relief with surgery, whereas only 10% reported some relief of pain without surgery. However, the relief of pain came at the cost of a 43% overall incidence of early and late complications with surgery.

Late-onset Idiopathic versus Degenerative Scoliosis

Adult scoliosis can generally be divided into two major types.5,9–12 LIS develops before skeletal maturity but may become symptomatic in adulthood. Degenerative scoliosis develops after skeletal maturity. The coronal curves in LIS often consist of a main thoracic (MT), proximal thoracic (PT), or TL curve, whereas the curves in degenerative scoliosis are primarily found in the TL or lumbar spine. As patients with LIS age, their pre-existing scoliosis may be complicated by facet-joint pain, disc degeneration, and curve progression.

LIS is an imprecise name for the condition to which it is applied. It can include cases of AIS that were never diagnosed during childhood or adolescence, or that were diagnosed but not treated or were managed with some form of bracing. LIS can also include surgically managed cases of scoliosis that have subsequently developed curve progression above or below the area of fusion, or cases in which symptoms develop as a result of poor sagittal alignment or degenerative changes.

Positive sagittal balance may develop in either LIS or degenerative scoliosis.1,13 It can result from a loss of physiological lordosis in the lumbar spine as a consequence of degenerative changes or of the previous treatment of scoliosis. True flat back syndrome with severe sagittal imbalance traditionally results from the treatment of TL scoliosis, particularly in patients with Harrington distraction instrumentation. Less commonly, thoracic or TL junctional kyphosis may occur with a pseudarthrosis at the site of prior surgery, or proximal junctional kyphosis may occur after spinal instrumentation and fusion.1

Degenerative scoliosis usually presents during the sixth or seventh decades, with an equal incidence in men and women.13,14 The degeneration is often widespread and associated with facet-joint arthropathy, osteophyte formation, and hypertrophy of the ligamentum flavum. The apex of this type of curve is usually between L3 and L4,15 and the curve may have significant rotational translation of the apical vertebra, as well as lateral listhesis of the vertebral bodies at or near the apex of the curve.16

Although it is occasionally difficult to differentiate between LIS with extensive degenerative changes and degenerative scoliosis, and there are similarities in the treatment of both disease processes, this chapter will be restricted to the evaluation and management of LIS.

Adult Scoliosis in Patients with Previous Spine Surgery

As the number of patients treated for AIS increases, it is more common for patients with symptomatic LIS who have undergone previous spine surgery to come to medical attention. A subgroup of patients present with recurrent or progressive scoliosis after prior intervention.17 Some patients develop degeneration at adjacent vertebral levels, whereas others are found to have a symptomatic pseudarthrosis at the site of a prior fusion. Several recent publications have evaluated the long-term results of surgical intervention for AIS, and provide insight into the incidence of symptomatic LIS in patients with previously treated AIS. Danielsson and coworkers followed 283 patients, of whom 156 were treated with Harrington-rod instrumentation and fusion, and the remainder of whom were treated with a brace. The mean follow-up times were 23 years for the surgically treated group and 22 years for the brace-treated group. Surgical complications included pseudarthrosis in three patients and flat back syndrome with positive sagittal imbalance in four patients. Eight of the patients treated with fusion (5.1%) required additional surgery for complications related to fusion. Of the patients who had revision surgery, three had hook displacement caused by fracture of the vertebral arch within the first 2 months after surgery. One patient required surgery for treatment of pseudarthrosis. Two patients who developed flat back syndrome were treated with osteotomies. Two additional patients had a rod removed.18

Several studies have examined the effect of the extent of fusion on the natural history of patients who have undergone fusion for AIS. Poitras and coworkers reported the prevalence, nature, and consequences of back pain in patients who had undergone Harrington-rod instrumentation for AIS. This study sought to determine whether back pain was related to the number of vertebrae fused, the distal level of hook insertion, and the degree of correction. The distal level of fusion was not found to influence the occurrence of back pain during adulthood.17 However, this finding is contradicted by several other studies that have described a relationship between the distal extent of fusion and the occurrence of symptoms later in life. In 1983, Cochran et al reported the results of Harrington-rod instrumentation and fusion in 95 patients. The proportion of patients with back pain increased significantly with lower levels of fusion. Thus, among patients who had fusion at L1, 25% had pain, rising to 30, 39, 62, and 82% of those who had undergone fusion at L2, L3, L4, and L5, respectively.19 The extent of fusion was also found to correlate with back pain in several other series.20,21

It is difficult to set general guidelines for the treatment of LIS in patients who have previously undergone fusion because many factors affect their subsequent condition. Factors including the duration and severity of symptoms, curve progression, degenerative changes at the end of fusion constructs, spinal alignment, and pseudarthrosis can affect the the appropriate treatment, requiring its clinical correlation with specific symptoms. Figure 24.1 shows radiographs of a 47-year-old woman who had undergone treatment of idiopathic scoliosis with a double major curve at the age of 13 years. The patient had a history of back pain with progressive worsening of radicular pain in her lower extremities that was worse on the left than on the right. The patient had undergone extensive trials of conservative therapy, and had a computed tomography (CT) myelogram that showed circumferential narrowing at the L4–L5 and L5–S1, with severe neural foraminal narrowing particularly on the left side at L4–L5.

The patient underwent a two-stage operation, of which the first stage was an anterior retroperitoneal approach to the lumbar spine with radical diskectomies and correction of scoliosis at L4–L5 and L5–S1, and inter-vertebral-body instrumentation at these levels with the use of bone morphogenetic protein (BMP). Stage 2 was a posterior exposure of the thoracic and lumbar spine followed by partial removal of the previously inserted Harrington-rod instrumentation, inspection of the fusion mass and visualization of a known pseudarthrosis at the L3–L4 level, and transpedicular instrumentation of L1 through S1 on the left side and from L3 through S1 on the right side. Connection instrumentation was joined to the Harrington rods to provide instrumentation from T12 through S1 with bilateral iliac screws; bilateral bone grafts harvested from the iliac crest; laminectomies at L3, L4, and L5; and Smith-Petersen osteotomies at L3–L4, L4–L5, and L5–S1. An arthrodesis was done from T12 to S1 with a combination of local bone-graft material, BMP, and allograft. Postoperatively, the patient’s radicular symptoms resolved and her back pain eased considerably.

Clinical Symptoms and Radiographic Associations

Historically, it has been difficult to predict health status on the basis of radiographic measures of deformity. Recent studies of adult scoliosis have attempted to correlate radiographic appearances with clinical symptoms, but this correlation has been inconsistent. The ability to associate clinical symptoms with radiological measures of scoliosis has improved with the use of objective outcome measures.13,22,23 More objective criteria have been used both to assess the risk of progression of symptoms before surgery and the likelihood of a satisfactory result after surgery.

Sagittal balance has been recognized as a critical factor in the assessment of adult patients with spinal deformity. Glassman et al reviewed data from a prospective multicenter study of adult spinal deformity and correlated various radiographic measures of deformity with patient-based outcome measures in adult scoliosis. The radiographic parameters studied were the type, location, and magnitude of spinal curves; coronal balance; sagittal balance; apical rotation; and rotatory subluxation. The study included 172 patients who had not undergone prior surgery and 126 patients who had undergone spinal fusion. Positive sagittal balance was the most reliable predictor of clinical symptoms in both patient groups. TL and lumbar coronal curves correlated with lower outcome scores than did thoracic curves in both patient groups. Coronal imbalance of >4 cm was associated with poorer pain and function scores for patients without prior surgery but not for those who had previously had surgery.23 This study highlighted the strong relationship between sagittal balance and outcomes in adult scoliosis.

Schwab and colleagues were also able to demonstrate a correlation between radiographic parameters and pain in patients with adult scoliosis.13 They prospectively studied 95 patients who completed a clinical questionnaire that included a self-reported visual analogue scale of pain and underwent full-length standing anteroposterior (AP) and lateral plain radiography. Radiographic analysis included measurement of the Cobb angle, the number of vertebrae in each curve, the plumbline offset from T1 to the mid-sacral line, the upper endplate obliquities of L3 and L4, and the maximal lateral olisthesis between two adjacent lumbar vertebrae. Measurements in the sagittal plane included lumbar lordosis, TL kyphosis, and the pelvic tilt index. Lateral vertebral olisthesis, endplate obliquity angles at L3 and L4, lumbar lordosis, and TL kyphosis were significantly correlated with pain. Surprisingly, Schwab and colleagues found that neither the Cobb angle nor age correlated with symptoms.

Presurgical Planning

Most patients with symptomatic adult scoliosis have undergone numerous prior evaluations and treatments before having been referred to a spine surgeon. A careful review of a patient’s past treatments and tests over time can give insight into the progression and response to treatment of the patient’s disease. A correlation of the patient’s symptoms with the clinical and radiological findings provides the clinician with important information toward finding the optimal treatment for the patient.

A thorough clinical and radiographic assessment of spinal deformity should be done in all patients at the time of their initial presentation to the surgeon, especially if any surgical intervention is anticipated. The radiographic analysis begins with full-length, upright 36 x 14-in. posteroanterior (PA) and lateral films that permit measurement of all coronal curves, the thoracic and lumbar curvature in the sagittal plane and the coronal and sagittal balance. This may be complemented with studies done for surgical planning, including supine, flexion, extension, and side-bending radiographs to evaluate curve flexibility. To optimize visualization of the entire spine on the lateral radiograph, the “clavicle position” should be used. In this position the patient fully flexes the elbows, with the hands in a relaxed fist, wrists flexed, and proximal inter-phalangeal joints placed comfortably into the supraclavicular fossa with passive forward flexion of the humerus. This position permits significantly better overall visualization of critical vertebral landmarks.24–27 Ideally, on the lateral radiograph, one should be able to visualize the trunk from C7 to the pelvis, including the hip joints, to assess the global sagittal balance of the spinal column.

Similarly, on the PA view, the margins of the rib cage and the pelvis, along with the femoral heads, should be clearly visualized. Assessment of the hips and an evaluation for possible leg-length discrepancy, arthritis of the hip, and pelvic pathology is essential. Visualization of the ribs helps in determining deformity of the thoracic cage associated with a congenital deformity rather than with untreated AIS. Either a congenital fusion of the ribs or a significant chest wall deformity can be associated with rigid or fused spinal segments. After spinal balance is assessed in the sagittal and coronal planes, Cobb-angle measurements are made on each area of the spine including the cervical, PT, MT, TL, and lumbar areas. The vertebral-body rotation at the apex of the curve in the coronal plane is a factor determining the rigidity of the curve. The greater the vertebral-body rotation the more likely it is that there will be substantial rigidity of the coronal curve. Loss of disc space height, extensive degenerative changes of the facet joints, and bridging osteophytes are also associated with curve rigidity.

Sagittal balance is determined by examining the vertical axis constructed through the middle of the C7 vertebral body and projecting it inferiorly to intersect the horizontal line through the L5–S1 disc space.22 In a balanced spine this line should ideally pass through the posterior third of the L5–S1 disc space. It is probably acceptable to have the C7 plumbline at least pass through or be posterior to the center of the acetabulum. The C7 plumbline may be acceptably located in elderly patients if it is up to 4 cm anterior to the L5–S1 disc space.

CT myelography is a frequently used diagnostic procedure in the evaluation of adults with scoliosis when surgery is planned. The CT myelogram provides intimate details of bone anatomy and is helpful in identifying areas of lateral recess and far-lateral stenosis. The diagnostic value of the CT myelogram is particularly relevant in patients with severe spinal deformity, substantial degenerative changes or a history of previous surgery. An evaluation of bone anatomy with this technique will dictate the specific instrumentation options available for the patient.

Magnetic resonance imaging (MRI) scans of the spine provide additional information about relevant soft tissues such as neural elements and vasculature, and the extent of disc hydration. The degenerative status of the discs in the lower lumbar spine is important in determining the lowest instrumented segment in a corrective construct, a topic discussed later in this chapter.

The determination of whether each sagittal and coronal component of a spinal deformity is fixed or fused and rigid or flexible contributes significantly to the surgical decisionmaking process. The characteristics of each portion of the spinal deformity should be evaluated in determining the overall flexibility of the deformity. Curve magnitudes may vary with the elimination of gravity; supine and standing films should be obtained for every patient who has a significant increase in symptoms when moving from a supine to a standing position. For certain cases in which large rigid scoliotic or kyphotic curves are found, push-prone, traction, or bolster radiographs help in the further assessment of flexibility.22,28,29 A rigid or fixed deformity will dictate whether an anterior release and fusion, a posterior osteotomy, or a vertebral column resection should be done to successfully achieve the desired correction of a deformity and spinal balance. For instance, an isolated 35-degree idiopathic scoliotic lumbar curve in a young adult is usually flexible and is often substantially reduced on side-bending films. Improvements exceeding 70% in postoperatively measured Cobb angles in such patients are common with the use of modern surgical instrumentation and techniques for correcting posterior deformities. However, the treatment of an elderly patient with LIS and a curve of similar magnitude would rarely produce the same amount of correction. Hence, management plans for curves of a particular magnitude may differ according to the inherent flexibility of the curves.

It is also necessary to assess the flexibility of sagittal curves. A bolster placed under the apex of a kyphotic deformity to maximize postural correction is particularly useful. This technique permits a better assessment of sagittal curve flexibility than can be obtained by the patient’s attempting maximal extension.