History

Anterior approaches to the thoracolumbar and lumbar spine were used for decades before their application to the treatment of idiopathic scoliosis. In the mid-1960s, Dwyer and colleagues introduced their system for the anterior correction and stabilization of scoliosis. Large screws were placed across the vertebral bodies and were connected by a flexible titanium cable on the convex side of the curve. Compressive maneuvers were then used to correct scoliosis by shortening the longer, convex side of the curve. Disadvantages of this system were numerous. The instrumentation could not be adjusted after the screw-cable interface had been crimped, nor did it provide rigid internal fixation, which resulted in the poor long-term maintenance of correction and necessitated postoperative immobilization of the spine. The kyphosing effects on the spine of the Dwyer technique, its inability to produce vertebral derotation, and the high pseudarthrosis rates with this technique have lead to its disuse.1–5

In 1976, Zielke shared his experience with a modification of the Dwyer system that utilized a solid 3.2-mm threaded rod instead of a flexible cable. This provided more rigid fixation that was able to resist the kyphosing effects of the anterior approach and provided for vertebral derotation. However, postoperative immobilization with a brace was still necessary, and unacceptably high rates of pseudarthrosis were reported in early series using Zielke instrumentation.68

In the 1980s and1990s, systems utilizing one or two solid rods provided more rigid fixation that obviated postoperative external immobilization, decreased pseudarthrosis rates, and maintained the correction of curvature obtained in both the sagittal and coronal planes.9–12

Correction of scoliosis in the coronal plane and hypolor-dosis in the sagittal plane is achieved by 90-degree rotation of the rod used in the construct for treating scoliosis, with mild compression applied on the convex side of the instrumented curve. Lumbar lordosis is maintained by the stiffness of the treatment construct and the placement of cages around adjacent vertebral bodies at the appropriate levels, or by structural bone grafts. Postoperative immobilization is usually not required.

Patient Selection

Those patients best suited for anterior instrumentation and fusion of their TL/L curve fit the criteria established by Lenke and colleagues in their description of the surgical treatment of type 5 curves:13 The proximal thoracic (PT) and main thoracic (MT) curves are nonstructural, whereas the TL/L curve is structural (comprising the major curve in the type 5 category, with the largest Cobb angle). The apex of the thoracolumbar curve is at T12-L1, and that of the lumbar curve is between the L1-L2 disc and L4. All Lenke type 5 curves have a ?C? lumbar modifier by definition. A patient who has undergone a prior abdominal operation may not be a candidate for the anterior approach. A dilemma may exist if the last vertebral body to be instrumented is that of L4 or L5, because of the overlying great vessel (aorta) impeding exposure of the anterior lower lumbar spine. A large body habitus should not be a contraindication to anterior instrumentation of the spine.

Sanders and colleagues attempted to set forth criteria for including the thoracic curve in fusion.14 The preoperative thoracic-curve magnitudes in their study ranged from 30 to 55 degrees. They determined that a patient with a TL/L-to-thoracic-curve Cobb-angle ratio of ≥1.25 and a thoracic curve that bent out to ≤20 degrees, along with closed trira-diate cartilages, had the best chance of having a satisfactory result of surgery. A satisfactory result was defined as a thoracic curve ≤40 degrees, ?reasonable? balance and sagittal alignment, and no need for additional procedures. The need to treat the thoracic curve is a relative contraindication to anterior fusion, because both the TL/L and thoracic curves cannot be instrumentated through a single thoracoabdominal incision.

Guille and colleagues15 reviewed the Harms Study Group (HSG) database and found 109 patients with Lenke type 5 curves for whom data had been prospectively collected at the time of fusion. Eighty-four patients (77%) had had a selective fusion of the TL/L curve only, whereas 25 patients (23%) had had fusion of the thoracic curve as well. Preoper-atively, the patients’ data were statistically similar in terms of sagittal parameters, skeletal maturity, and age at surgery. Surgeons in the HSG broke the rules of the Lenke classification and fused both the TL/L and thoracic curves in patients with larger thoracic curves, larger thoracic rib humps, and greater thoracic apical translations. Scores on the Scoliosis Research Society (SRS)-24 questionnaire were statistically similar in the two study groups at the time of a follow-up review, except that the group with fusion of their thoracic curves reported less function.

Preoperative Planning

Routine full-length posteroanterior (PA) and lateral radiographs should be made of the patient in the standing position, as well as lateral side-bending films of the patient in the supine position. As outlined earlier in this book, the standard measurements of spinal curvature and shoulder height should be made. Usually, the levels to be fused and instrumented are those within the TL/L Cobb angle. The uppermost level to be instrumented is the proximal vertebra in the measured Cobb angle. Care should be taken if this vertebra is the T12 level. The lower end-vertebra of the measured curve, determined from the standing PA view, is usually the best choice as the last instrumented level. This vertebra usually has minimal rotation, but may have a marked tilt. It is important to view the patient’s side-bending films to ensure that this vertebra can be made horizontal (after correction of the scoliosis). The long-term results of residual obliquity at this end-vertebral level, or those of facet incongruity, are unknown.

The vertebral bodies should be evaluated for their ability to accommodate two screws if a dual-rod construct is being considered. The radiographs and operative plan are ideally discussed preoperatively with the access surgeon. The thoracolumbar operative approach will not be discussed in this chapter.

Operative Technique

The spine should be exposed to reveal all of the vertebrae to be included in the fusion and instrumentation of a Lenke type 5 curve. Postural reduction of the magnitude of the curve is usually obtained by positioning the patient on the operating table, and is even greater after the discectomies are done. Care should be taken to avoid violating disc spaces not to be included in the fusion. Complete discectomies, down to bleeding bone, are done at all intended levels of the fusion procedure. Usually, there is no need to remove the annulus on the concave side of the patient’s curve or the posterior longitudinal ligament, but these steps may allow greater correction in larger curves. Soft tissue is removed from the vertebral bodies to allow the accurate placement of instrumentation. It is important to identify the posterior border of the vertebral body to be instrumented so that one of the screws to be inserted in it can be placed as posteriorly as possible, which is especially important with dual-rod constructs. Staples help demarcate screw-insertion sites and aid in avoiding the plowing of screws during corrective maneuvers. Particular care must be taken to accurately measure vertebral screw length and avoid excessive prominence of screws into the distal cortex. With open exposure of the lumbar spine, the surgeon can place an index finger on the contralateral side of the vertebral body during screw placement. Following screw placement, attention is given to the placement of inter-vertebral-body cages or grafts. The cage or graft should be placed on the concave sides of the vertebral bodies to aid in correction of the scoliosis, and in a slightly anterior position to create and maintain lumbar lordosis. The rods in a dual-rod construct are each usually 4.5 mm in diameter, whereas single rods are generally ≥5.0 mm. A mild curve is bent into the rod, which is placed with the apex of the bend facing upward, which aids in insertion of the rod. A 90-degree rod rotation is then done so that the apex of the bend in the rod faces anteriorly, thus creating a lordosis. If a second rod is used, it is inserted in situ, because the correction has already been achieved by the first rod. Mild compression is then done on the convex side of the patient’s curve for further correction. Radiographs are then made to ensure that the end instrumented vertebra is horizontal.

Dual versus Single Rods and Inter-vertebral-body Implants

Discussion of the biomechanical merits of single- versus dual-rod instrumentation abounds in the literature, with proponents of both types of construct.16–20 There is also debate about optimal rod diameter and the benefits of dual rods.21,22 Fricka et al have shown in a bovine model that dual-rod constructs are stiffer in torsion and flexion-extension loading.17 They found that structural support with dual rods helped with lumbar lordosis but did not increase construct stiffness. With single-rod constructs, the addition of structural support added stiffness in flexion. Oda and coworkers demonstrated in a calf-spine model that increased rod diameter did not improve construct stiffness or affect rod-screw strain.18 Dual-rod constructs had greater construct stiffness and less rod-screw strain. Chang et al showed no statistical difference in range of motion or load sharing when dual rods, each of 6.35 mm or 5.5 mm diameter, were used.19 Zhang and coworkers described a novel implant for anterior instrumentation.20 This rod-plate construct was significantly stiffer and provided a more stable bone-screw interface than did a single-rod-with-cages construct, but was comparable in stiffness and stability to dual-rod constructs. If the size of the vertebral body permits, a dual-rod construct should be used, and obviates the need for a brace. Frequently, a brace will be needed when only a single-rod construct is used.

Most studies support anterior inter-vertebral-body grafts and cages as aiding in fusion, sagittal-plane correction, and the maintenance of sagittal-plane correction. Lenke and Bridwell reviewed the use of titanium mesh cages for anterior spine surgery.21 Described benefits included sagittal-plane correction and the maintenance of lumbar lordosis, as well as a decreased frequency of pseudarthrosis. Lowe et al showed in human cadaver and bovine models that dual rods with structural inter-vertebral-body support were the best combination for increasing stiffness, and that the addition of cross-links did not add to stiffness but to torsional strength.16 Ouellet and Johnston showed that rib-strut grafting with single-rod constructs decreased the rate of pseudarthrosis, but this particular form of grafting did not affect the maintenance of correction or sagittal alignment.11,22 They suggested that mesh cages or femoral ring allografts may be better for these functions. The placement of inter-vertebral-body cages or grafts increases anterior column height and stability, but longer follow-up will be needed to ascertain whether they maintain sagittal balance.

Clinical Results

Guille and colleagues reviewed the HSG database and found 100 patients with Lenke type 5 curves treated with the anterior approach.23 Thirty-nine of the patients had received single-rod constructs ( Fig. 19.1 ) and 61 had received dual-rod constructs ( Fig. 19.2 ). Preoperatively, the two groups were statistically similar in their radiographic measurements. Intraoperatively, patients with the single-rod constructs had shorter operative times and less blood loss. Average coronal-curve correction was 83% for the single-rod group and 72% for the dual-rod group. Measured values of sagittal alignment were improved and maintained in both groups. At a minimun of 2 years follow-up, the patients treated with single rods had better maintenance of coronal correction, whereas four of the patients treated with dual rods had pseudarthroses.

Turi and coworkers reported their experience with single-rod Texas Scottish Rite Hospital instrumentation in 14 patients.9 Average coronal-curve correction was 76%, with no pseudarthrosis. There was an average 5-degree loss of correction at follow-up. Sweet and colleagues reviewed 47 patients who had anterior fusions instrumented with 5.0-mm or 5.5-mm single-rods with cages for their TL/L curves. The average correction was 70%. Two patients developed a pseudarthrosis. Kaneda et al reported their experience with 20 patients who had idiopathic TL/L scoliosis treated with dual-rod anterior instrumentation.12 They showed an 83% coronal-curve correction, maintenance of sagittal alignment, and no pseudarthrosis. Hurford et al reported their experience with 42 patients who underwent dual-rod anterior instrumentation and fusion for idiopathic scoliosis.24 At follow-up, they found a 67% average correction and no pseudarthrosis. This percent correction compares favorably with the 70% correction seen in a group of patients treated at the same institution with single-rod constructs,25 also without any cases of pseudarthrosis.