Advantages and Disadvantages of Selective Thoracic Fusion

The advantages of selective fusion are maintenance of lumbar motion segments of the spine and correction of the primary deformity in scoliosis, the main thoracic curve. Motion of the spine occurs predominantly at the thoracolumbar junction and in the lumbar segments. It is logical to assume that preservation of these motion segments will provide better long-term health of the spine. Studies have determined that a more proximal fusion level in AIS results in greater motion of the spine.16,17 Wilk et al compared 34 patients who had fusion for AIS, 32 patients who did not have fusion, and 25 control patients, and demonstrated less motion in those patients who had fusion into the lumbar spine than in those who had thoracic fusion only.16 Greater mobility of the lumbar spine appears to be important in the long-term health of the spine and is the primary reason for performing selective thoracic fusion when possible. Clinical studies have substantiated some of the perceived problems with fusion into the lumbar spine. Paonessa and Angler18 reported greater back pain scores, difficulties with normal daily activities, increased need for pain medications, and more episodes of back pain with fusion to L3 or caudally than with fusion to more proximal levels.18 Cochran and coworkers analyzed the long-term functional changes in patients with AIS and demonstrated more low back pain, degenerative facet-joint changes, and disc space narrowing in patients with fusion to L4 or L5.19,20 However, it should be noted that the patients in this study were treated with Harrington implants that flatten the lumbar spine and lead to a flatback deformity and resultant pain and disability. With current segmental fixation and attention to maintaining lumbar lordosis, the long-term results, although yet to be determined, should be improved. Besides permitting greater mobility, a selective thoracic fusion is a shorter and less complicated surgical procedure than is fusion into the lumbar spine.

The disadvantages of selective thoracic fusion include less correction of the coronal-plane deformity in scoliosis, with greater risk of decompensation. For patients who wish to have significant correction of a thoracic deformity, a selective fusion will not allow the complete correction of deformity in the thoracic or lumbar spine. The greatest potential disadvantage of selective fusion is that it may lead to left decompensation, requiring an additional procedure to achieve fusion into the lumbar spine so as to provide coronal balance. This extension of fusion into the lumbar spine may require a more distal level of fusion than would have been needed if the fusion had been done in the primary procedure. It should be remembered that extension of fusion into the lumbar spine because of decompensation following an attempted selective thoracic fusion is a relatively uncommon occurrence.6,7,13,21

Although selective fusion is often the goal in surgery for scoliosis, a variety of opinions exist for when it should be performed and how it should be performed. Newton and coworkers22 analyzed factors involved in the decision to perform a selective fusion for King type II and Lenke type 1B and 1C curves at five different centers. Despite all of the curves being Lenke type 1 curves and therefore requiring side-bending lumbar curve correction of <25 degrees, there was wide variation in the frequency of selective fusion, ranging from 67 to 94%. Newton and colleagues demonstrated that the rate of selective fusion was higher for Lenke 1B than for Lenke 1C curves, at 92% versus 68%. Radiographic factors associated with the fusion of both types of curve included a larger preoperative lumbar curve (42 degrees vs. 37 degrees), greater displacement of the lumbar apical vertebra (3.1 vs. 2.2 mm), and a smaller ratio of thoracic-to-lumbar curve magnitude (1.3 vs. 1.4). However, the most important predictor of fusion into the lumbar spine was the philosophy of each surgeon at each site, with those who felt strongly about preserving motion segments being more likely to perform a selective fusion. If there is a question about the appropriateness of selective thoracic fusion, the matter should be thoroughly discussed with the patient and the patient’s parents, including details of the risks and benefits of selective thoracic fusion.

Anterior versus Posterior Approach for Selective Fusion

Two main approaches are available for fusion of the thoracic spine: anterior and posterior. The anterior approach utilizes instrumentation with vertebral-body screws, which are then connected by single or dual rods to gain correction ( Fig. 12.3 ). The correction maneuvers are typically cantilevering and compression, although rod rotation has also been utilized. The advantage of anterior surgery for a selective thoracic fusion is that, in general, fewer motion segments are fused because this technique allows instrumentation and fusion of the measured Cobb angle, which may not be possible with posterior fusion. Selective thoracic fusion utilizing anterior instrumentation may influence the lumbar curve to a lesser degree than posterior instrumentation if fewer motion segments are included. In addition, the correction mechanics of anterior surgery may not impart a significant rotational force to the lumbar spine, with a correspondingly reduced risk of creating coronal imbalance and decompensation. However, anterior surgery does require entry into the chest, which may injure vital organs and structures and will affect pulmonary function. Yet the use of thoracoscopic techniques limits the negative effect of an anterior approach on pulmonary function.22

The posterior approach to instrumentation and fusion of the thoracic spine is more familiar to most surgeons. It is a straightforward approach that can be performed quickly and most commonly produces outstanding results. Posterior instrumentation and fusion, however, does disrupt the paraspinal musculature, and may therefore have long-term health benefits with respect to back pain. The posterior approach often requires fusion to a more caudal extent than anterior surgery, although the use of thoracic pedicle screws may allow greater ability to preserve motion segments. The use of thoracic pedicle screws readily permits segmental manipulation of the spine to a desired correction through multiple correction strategies including cantilevering, segmental in situ bending, translation, direct vertebral rotation, and incomplete rod rotation ( Fig. 12.4 ). Early reports of posterior techniques with Cotrell-Dubousset and Texas Scottish Rite Hospital instrumentation to accomplish selective fusion included some concern about creating coronal decompensation.7,8–10,23–24 These techniques used all-hook constructs, which may not be comparable to the all-pedicle-screw constructs in use today. Several factors were thought to cause coronal decompensation with these all-hook constructs, including instrumenting into the lumbar curve, overcorrection of the thoracic curve, and continued obliquity of the L4 vertebra relative to the pelvis.

Early comparisons of the anterior and posterior approaches to selective fusion demonstrated superior results with the anterior approach.21,25 Betz and colleagues reviewed their experience with 78 patients who underwent anterior spinal fusion with flexible threaded rods in comparison with 100 patients who underwent posterior spinal fusion with multisegmental hook systems. The anterior approach saved 2.5 motion segments (mean) in that many of these patients did not have fusion into the lumbar spine. Surgeons treating these patients felt that anterior surgery could more often be used to perform a selective thoracic fusion, because posterior instrumentation often led to decompensation.21 To specifically compare the anterior and posterior techniques in a selective-fusion setting, Lenke and colleagues analyzed the cases of 123 patients with selective fusions done either anteriorly or posteriorly. The thoracic-curve correction was superior for the anteriorly treated group (58% vs. 38%) and resulted in greater spontaneous correction of lumbar curves (56% vs. 37%).25 Lenke and colleagues distinguished lumbar curves on the basis of the position of the apical lumbar vertebra relative to the center sacral vertical line (CSVL), as in their subsequent classification of lumbar curves into those with an “A” modifier (CSVL between pedicles), “B” modifier (CSVL touching the vertebra), and “C” modifier (apical vertebra not touching CSVL). The anterior approach produced greater compensatory lumbar correction in all three of these types of lumbar curves, which was most dramatic for type C lumbar curves. These early comparisons of the anterior and posterior approaches were made with posterior techniques in which only hooks were used and the lowest instrumented vertebra (LIV) was selected as the stable vertebra. A more recent study, in which hybrid and all-screw posterior constructs were used, demonstrated no difference between the anterior and posterior approaches for spontaneous lumbar-curve correction in a series of patients matched for LIV, lumbar-curve flexibility, and percent thoracic-curve correction.26

Suk et al were the first to describe thoracic pedicle-screw fixation, demonstrating outstanding overall results with good safety with this technique.27 They reviewed their experience with selective thoracic fusion with segmental pedicle fixation in 203 patients, and showed 69% correction of thoracic curves with a compensatory lumbar-curve correction of 66%. There were no instances of junctional kyphosis; however, coronal decompensation occurred in 10 patients (5%). Also, 17 patients experienced “adding on” to their spinal curves, most likely because of fusion levels that were too short.15 Dobbs and colleagues compared posterior hooks with posterior pedicle screws for the selective fusion of thoracic curves in a series of 66 patients, all of whom had lumbar C modifiers. The thoracic pedicle-screw group had greater thoracic correction (53% vs. 34%) and a greater lumbar-curve response (38% vs. 30%) with fewer instances of decompensation than did the posterior-hook group at 2-year follow-up. There were no complications or reoperations in either group. Pedicle-screw instrumentation allows surgeons to keep the lowest instrumented vertebra appropriately tilted, and provides greater ability to “dial-in” the amount of thoracic correction desired on the basis of preoperative thoracic- and lumbar-curve flexibility.28 These studies support the concept that pedicle screws offer the ability to improve the thoracic curve in scoliosis without decompensation.

In analyzing only patients with a lumbar type C curve, and who were treated with mixed instrumentation consisting of all-hook, all-screw, and hybrid constructs, Edwards and coworkers, at an average 5-year follow-up, demonstrated a 36% overall correction in the patients’ thoracic curvature, accompanied by a 34% lumbar-curve correction, which occurred primarily in the more cephalad segments of the lumbar curve.29 The majority of patients did well with respect to spontaneous correction of their lumbar curvature, although preoperative coronal imbalance was a predictive factor for postoperative coronal balance. Those patients who had a coronal imbalance postoperatively had poorer functional outcome scores, as measured with the Scoliosis Research Society (SRS)-24 instrument. This series had mixed instrumentation, all hooks, all screws, and hybrids.