The main goals of the surgical treatment of idiopathic sco-liosis are the prevention of deformity progression through arthrodesis and correction of deformity sufficient to balance the spine in both the frontal and sagittal planes. Ideally these goals are accomplished while maintaining as much lumbar spinal flexibility as possible. Under the influence of Dr. John Hall, the group at the Children's Hospital developed such a motion sparing method using anterior spinal fusion and instrumentation.

HISTORY

Anterior spinal instrumentation for the treatment of idio-pathic scoliosis was introduced in the 1960's by Dr. Anthony Dwyer and colleagues 1 . His method offered a number of advantages over existing posterior instrumentation systems, but the construct —composed of screws, staples, and a flexi-ble cable— lacked rigidity and often produced kyphosis. Unacceptable rates of pseudoarthrosis, implant failure, and loss of lumbar lordosis have led to the abandonement of the system when used alone for the surgical treatment of idiopathic scol-iosis 2,3,4,5. It has remained part of the armamentarium for the treatment of neuromuscular curvatures, however, when used in concert with posterior instrumentation. The Zielke system, which uses a flexible solid rod, had similar problems 6-13 .

			Fig. 1. A. This 15-year-old girl is the ideal candidate for short segment anterior instrumentation. She has a progressive 50thoracolumbar curvature and a flexible mild thoracic curve. The apex of the structural curve is the T12-L1 disc; by the selection criteria based on the standing radiographs, the instrumentation would extend from T11- L2. B. The sagittal alignment is within physiologic norms, but note that the segment to be instumented, T11-L2 is kyphotic. C. The left lumbar supine bend shows reversal of the wedge shape between the T10/11 disc and the L2/3 disc; by the bending film selection criteria, the instrumented segment is the same as determined on the standing views. D. The thoracic curvature is flexible, correcting to 20 or less.
A.		B.

C.	D.	E.

E. On the bend view, the vertebral body at the caudal end of the segment to be instrumented comes close to the midsacral line, but in general we have found that once the apex is over- corrected, the fractional curvature centralizes regardless of the alignment on preoperative bending films. F. The spine is well aligned. Overcorrection of the thoracolumbar curvature by 9" results in a straight lumbar spine and spontaneous correction of the thoracic curve to 15. G. The sagittal alignment is normal; the instrumented segments which had 10 kyphosis preoperatively now show a lordosis of 8. The disc spaces appear fused at one year following the procedure.
	F.	G.

Table 1

Standing Posteroanterior Radiograph
    Select most horizontal element at apex
          If vertebral body, instrument 1 body above and 1 below
          (3 bodies, 2 discs)
          If disc, instrument 2 vertebral bodies above, and 2 below
          (4 bodies, 3 discs)
Supine Right and Left Active Bends
     Bend to convexity
          First disc space above and below apex that reverses
          the wedge shape is left unfused
          Bend to concavity
          Vertebral bodies below fusion levels should become
          horizontal with pelvis (a relative requirement)
___________________________________________________
Select method which gives longest construct

PATIENT SELECTION

The ideal candidate for short segment anterior instru-mentation is an adolescent or young adult with a flexible idio-pathic thoracolumbar scoliosis and truncal decompensation. For this type of deformity, traditional posterior instrumenta-tion systems would necessitate fusion to the lower lumbar spine; anterior over-correction, in contrast, may save two lum-bar motion segments. To produce a balanced spine, both the lumbar curvature caudad and thoracic curvature cephalad to the segments to be instrumented must be flexible enough to realign following structural curve correction. The structural deformity must be supple enough to be reversed by 10 degrees; in general this requires a lumbar or thoracolumbar curvature of no greater than 60and a thoracic curvature that corrects to 20 or less on the supine bending film. As anterior instru-mentation has a tendency to produce kyphosis, thoracic hyper-kyphosis is a relative contraindication to this procedure.

SELECTION OF FUSION LEVELS

Selection of proper fusion levels is essential for the suc-cess of the procedure. A set of guidelines based on the radi-ographic scoliosis series has been established specifically for the anterior apex overcorrection technique (Table 1) 15 . The instrumented segments include either three or four vertebra. When the levels of instrumentation as determined by the standing and supine methods differ, the longer of the two con-structs is chosen.

TECHNIQUE

Overcorrection of the structural scoliosis, which is essen-tial to the success of the short segment technique, places sig-nificant stresses on the instrumented portion of the spine. Care is required to minimize these stresses by proper patient positioning, destabilization through soft tissue resection, accurate hardware placement, and controlled reduction maneuvers. The patient is placed in the lateral decubitus posi-tion, the kidney rest is elevated, and the operating table is flexed to improve the exposure. A transpleural, retroperitoneal approach is made to the spine and the proper level is confirmed with intraoperative radiographs. The segmental vessels over the vertebral bodies to be instrumented are ligated and divided, and a subperiosteal exposure is performed. The intervertebral discs are completely excised and the cartilaginous endplates are resected. Currently, ISOLA anterior instrumentation is used. The screw entry site is the mid-lateral aspect of the vertebral body. A staple is placed, taking care to avoid penetration of the end discs, and the screw tract is tapped through the far cortex. The appropriate length screw is placed. It must be oriented par-allel to the end plates with one thread engaging the opposite cortex. Each disc space is distracted with a laminar spreader and a properly sized structural graft, currently a Harms cage, is placed in the anterior aspect of the body slightly toward the concavity. The disc space is the packed with morecelized auto-genous bone graft harvested from the rib. The operating table is then straightened to allow spontaneous correction.

Fig. 2. A. This young woman, who is very active in cheerleading, has a double major curve pattern with the lumbar curve greater than the thoracic, a King 1 pattern. If posterior instrumentation were utilized, both the lumbar and thoracic curvatures would have to be included in the fusion. B. There is normal sagittal alignment, with 5 lordosis through the segment to be instrumented.
	A.	B.

		C. At two years following surgery, the over-correc-tion of 11 is maintained. The lumbar curvature is corrected to 20 and the thoracic corrected to 28. There is wedging of the disc caudal to the fusion, a finding that some feel may cause painful degen-erative changes in the future. To date we have not had this problem. The technique has allowed for spinal flexibili-ty so that she has been able to continue her cheerleading. Should problems arise in the future, the fusion could be extended and she would still have a more flexible spine than if a posterior procedure had been performed. D. The lateral demonstrates physiologic alignment. The lordosis through the instrumented segment has been maintained, and the disc spaces are fused.
C.	D.

A 1/4 inch rigid rod is then cut to the correct length and contoured to provide the appropriate lordosis and scoliosis over-correction, usually 20. The rod is placed in the screws and rotated anteriorly toward a lordotic position, then further rotat-ed until the desired degree of lordosis and reverse scoliosis is obtained. The structural grafts are checked for position, and gentle compression is applied toward the central screws to sta-bilize the construct. Bone graft is added beneath the periosteal sleeve and over the disc spaces. All screws are revisited and tightened. A chest tube is placed and the wound is closed.

A. D.
	B.	C.
	Fig. 3 A. This teenage girl has a 60thoracolumbar curvature and truncal decompensation. B. There is 6kyphosis in the upper lumbar spine. C. Following an apex over-correction of 12, the lumbar spine is straight and the 40thoracic curvature has corrected to 25. There is no significant wedging of the disc inferior to the instrumentation. D. The technique of solid rod and structural anterior graft has allowed correction of the segmental 6 kyphosis to 10of lordosis.

The patients are mobilized as tolerated postoperatively. Patients were instructed to wear a semi-flexible anterior open-ing Boston overlap brace when out of bed for the first three months post-operatively, though this is no longer a routine practice. Physical education is prohibited for six months and contact sports for one year. Follow up radiographs are taken at three, six, 12, and 24 months postoperatively.

PATIENT REVIEW

We have used the technique of short segment anterior instrumentation with Harms cages for over 3 years. Twenty-five patients have been followed for at least two years. All had adolescent idiopathic thoracolumbar or lumbar scoliosis. The average curvatures were approximately 50pre-operatively and were corrected 70%-80% on average. Over-correction of the instrumented segment was maintained at two years. The over-all lumbar lordosis was essentially unaltered and, most impor-tantly, the lordosis through the instrumented segment was maintained or improved. In general, most patients showed no loss of correction over time, though several had a deterioration due to an increase in the compensatory thoracic and subse-quent increased lumbar deformity. To date there have been no pseudoarthroses or other major complications.

DISCUSSION

The fate of the unfused portion of the lumbar spine below an instrumented segment is always a concern. Long-term studies that document increased low back pain following pos-terior instrumentation and fusion have involved the use of a straight Harrington rod with resultant loss of the normal lum-bar lordosis 19 . The newer segmental systems have not been in use long enough for comparison. We believe that increasing the number of spared lumbar motion segments distributes the stress across more levels and minimizes the risk of degenera-tive disc disease. Previous problems of high pseudarthrosis rates, instrumentation failure, and relative kyphosis through the instrumented segments seem to have been solved with the latest generation construct.

While anterior spinal fusion and instrumentation for tho-racolumbar and lumbar scoliosis has been reported in the past, the short segment technique, which requires overcorrection of the structural curvature, is nearly unique to Children's Hospital 15,16,17,20 . This method, we believe, obtains excellent truncal balance while preserving an additional lumbar motion segment. One criticism of the technique has been the wedged disc space that sometimes occurs beneath the instrumented segments 21 . Degenerative changes have been reported to be associated with this finding by Kohler et al 16 . In that series, however, Dwyer instrumentation was used, the short segment overcorrection method was not performed, and the constructs extended into the lower lumbar spine. Furthermore, the exam-ples cited by Kohler et al. showed staple penetration of the disc spaces caudal to the fusion. To date we have not seen wedg-ing associated with discomfort or degenerative changes 15,16 .

D. The lateral radiographs reveal improved alignment, but at the expense of a much longer fusion.
A.	B.
C.	D.

CONCLUSION

We believe that preserving motion segments in the lumbar spine when performing surgical correction of lumbar and tho-racolumbar curvatures is critical, and thus we consider short segment anterior spinal instrumentation to be the treatment of choice for idiopathic scoliosis in properly selected patients. The method requires overcorrection or reversal of the structural curvature. The present technique, which utilizes a rigid rod, segmental screws, and anterior structural support in the form of mesh cages seems to have solved many of the problems that have been previously reported with this technique.