Definition and Incidence of Double Thoracic Curves

The reported incidence of the DT curve pattern depends upon the classification scheme used to categorize curve types for AIS. Moe and Kettleson, in their early series of 166 patients treated with braces, found that the DT pattern was the third most common curve pattern treated, occurring in 14% of patients.4 Recognition of AIS patterns that contained more than one major curve was facilitated by Moe’s routine use of 36-inch-long films.

King et al described the DT curve pattern that the authors defined as the King V pattern.2 Left PT curves that had an elevated left first rib and a positive T1 tilt (defined as the left upper corner of T1 being higher than the right upper corner generating a tilt of T1 into the concavity of the PT curve) were thought to be complete curves with structural characteristics that needed to be included in fusion procedures ( Fig. 18.1 ). PT curves that did not have a positive T1 tilt were believed to be fractional curves and did not require inclusion in fusion procedures. The King V pattern was the fourth most common of five patterns of curvature in AIS, and occurred in 11.6% of patients.

King and colleagues recognized the importance of curve flexibility as determined from side-bending radiographs. However, they applied an analysis of curve flexibility and a flexibility index only to the MT and thoracolumbar/lumbar (TL/L) curves to help differentiate between true double major and false double major patterns. Therefore, curve flexibility did not play a role in the criteria used to define the King V DT pattern. Although the King classification calls attention to the PT curve, it provides the physician with a somewhat vague guide for assessing the PT curve. This, together with poor inter- and intraobserver validity, reliability, and reproducibility in defining curves according to the King classification system, motivated Lenke and coworkers to develop a more comprehensive classification scheme for AIS that utilized curve magnitude, curve flexibility, and sagittal profile to differentiate structural from compensatory curves.5–6 The Lenke classification defines the major curve as the largest measured curve on standing 36-inch posteroanterior (PA) radiographs. The smaller minor curves are defined as structural if they remain at ≥25 degrees on side-bending radiographs or if they have a hyperkyphotic sagittal profile with ≥20 degrees of focal kyphosis. For DT curves in the Lenke classification, the MT curve is always larger than the PT curve, and is the major curve. If the PT curve remains at ≥25 degrees on side-bending radiographs or if the T2-to-T5 focal kyphosis on lateral radiographs is ≥20 degrees, the PT curve is considered structural. This, in combination with a nonstructural thoracolumbar or lumbar curve, denotes a DT or Lenke 2 curve pattern. Using the Lenke classification with more refined criteria to delineate structural curves, Lenke and colleagues reported a 20% prevalence of DT curve patterns in a multicenter retrospective review of 606 operatively treated AIS patients.3

Looking to the future, further attention has been directed to establishing more reliable measures for differentiating structural from nonstructural minor curves. Criticisms of using side-bending radiographs to define a curve include variable patient effort during side-bending radiographs and inconsistent technique. This motivated Cheh et al to compare the reliability of supine 36-inch anteroposterior (AP) radiographs with that of side-bending films when predicting curve flexibility.8 Cheh and coworkers reported that PT curves that remained at ≥30 degrees on supine radiographs were highly likely to be structural curves, and concluded that supine radiographs are a reproducible means for defining curve structurality and may eventually replace side-bending radiographs.

Clinical Evaluation

The clinical evaluation of a patient with scoliosis, regardless of age or etiology, should include a frontal view of the shoulders and posterior view of the entire spine, unobstructed by hair and clothing, to permit assessment of shoulder balance and trapezial fullness. Left trapezial fullness and an elevated left shoulder in the presence of a right MT rotational prominence should alert the physician to the possibility of a structural left PT curve and DT curve pattern.9,10 Patients with PT kyphosis may have a midline upper thoracic prominence, a forward-protruding cervical spine, or both. This may also indicate a structural PT curve, secondary to kyphosis, as part of a DT curve pattern.

Radiographic Evaluation

The radiographic evaluation of a patient with scoliosis should include upright 36-inch PA and lateral radiographs that allow the surgeon to visualize the spine at least from the C7 vertebra to the pelvis, including the femoral heads. The size of all curves should be routinely measured on the PA film so as not to ignore a potentially structural minor curve in the presence of a large major curve. PT focal kyphosis is measured from T2 to T5. As previously indicated, ≥20 degrees is indicative of a structural PT curve and is consistent with a DT curve pattern in the presence of a structural MT curve and nonstructural TL/L curve.6 Side-bending and supine radiographs are obtained to evaluate curve flexibility and provide further criteria for curve structurality. PT curves of ≥25 degrees on side-bending radiographs or of ≥30 degrees on supine radiographs are structural. The angle of T1 tilt is measured from the intersection of a line drawn along the T1 cephalad endplate and a line parallel to the horizontal (or perpendicular to the vertical edge of the radiograph) ( Fig. 18.2 ).11 A positive T1 tilt (elevation of the left upper corner) and angulation of the T1 endplate into the concavity of the PT curve on standing PA radiographs may be considered indicative of a complete PT curve, whereas a neutral or negative T1 tilt may be indicative of a fractional PT curve. However, some studies have supported curve flexibility and shoulder measurements as more accurate than T1 tilt in predicting the structural behavior of a PT curve.2,12–14

Measures of shoulder balance include clavicle angle, radiographic shoulder height, coracoid height difference, and first rib-clavicle height difference. These measures allow preoperative assessment of shoulder balance and can be used to predict postoperative shoulder balance.13 The clavicle angle is formed by the intersection of a tangential line touching the two highest points of the clavicle with another line perpendicular to the horizontal ( Fig. 18.3 ).14 Radi-ographically determined shoulder height is the difference in the soft-tissue shadow directly superior to each acromioclavicular joint as measured on a standing PA radiograph ( Fig. 18.4 ).14 The difference in coracoid height is measured as the difference between the horizontal lines traced along the superior edge of each of the two coracoid processes. A positive value is denoted when the left coracoid process is higher than the right.15 The first rib-clavicle height is defined as the vertical distance from the apex of the first rib to the immediately overlying superior clavicle.13

Several features of a DT curve are associated with central nervous system (CNS) abnormalities. Chiari malformation, hydrosyringomyelia, and other CNS abnormalities are uncommon in AIS, with a reported incidence ranging from 2 to 4%.16–20 Focal and global hyperkyphosis has been associated with an increased incidence of CNS abnormalities found on magnetic resonance imaging (MRI) in patients with presumed AIS.16,19 Davids et al reported a 2% incidence of CNS abnormalities among 274 patients with presumed AIS who had an MRI scan.16 Absence of thoracic apical lordosis was the most valuable indicator of a potential CNS abnormality on MRI scanning. Speigel and colleagues reported that in addition to an atypical curve pattern (e.g., left MT curve), the DT curve pattern was independently associated with an increased incidence of CNS abnormality on MRI scanning.19 In light of these reports, physicians evaluating a patient with a DT curve pattern and associated focal PT kyphosis should consider obtaining a screening MRI of the neural axis to rule out an underlying CNS abnormality that may require neurosurgical intervention before the definitive treatment of scoliosis.

Nonoperative Treatment

As in all patterns of AIS, the primary treatment goal is to prevent curve progression. It has been established that the risk of curve progression depends on the magnitude of the curve, the amount of growth that the patient has remaining, and where the patient’s remaining growth will occur in relation to the rapid phase of the adolescent growth spurt.21 The decision to observe, brace, or surgically treat the condition is based on these factors. The specifics of bracing can be found in Chapters 7 and 8 of this book. With regard to the DT curve pattern, the main challenge of bracing is controlling the PT curve. A thoraco-lumbo-sacral orthosis (TLSO) brace is effective only for curves with an apex at T7 or caudally, because the TLSO brace fits beneath the arms and cannot generate the appropriate reduction forces for controlling high thoracic curves. Consequently, a Milwaukee-type cervical-thoracic-lumbar-sacral (CTLSO) brace with an attached chin piece is recommended for treating the PT curve. Poor compliance with Milwaukee-type brace-wearing and poor efficacy of bracing in controlling PT curves have been documented.22 –25 Lonstein and Winter reported that 26% of 218 patients with a DT curve pattern treated with a brace went on to surgical treatment, as compared with an average of 22% progression to surgery for patients with all other curve patterns.26 They acknowledged that the PT curve is difficult to control with external bracing. Lonstein and Winter also indicated that the PT curve is rarely progressive, and that this portion of the DT curve pattern plays less of a role in progression to surgery than does the MT curve. They therefore believed that the DT pattern is amendable to bracing with a Milwaukee brace. Documentation of the PT curve size, flexibility, and T1 tilt angle before bracing is essential because worsening of the PT curve can be a perceived complication of bracing when in fact a structural PT curve was present before bracing.