Evaluation of the long-term results for single fully segmented hemivertebrae were subjected to single-stage excision via posterior approach alone.
To describe the long-term results of this procedure.
In the case of congenital spinal deformity caused by a single, full hemivertebra, excision of the hemivertebra is ideal for obtaining a good correction percentage even in short segments. Recently, single-stage excision of a hemivertebra using a combined anterior and posterior approach has been reported.
Five patients with a hemivertebra underwent surgery. The hemivertebra involved the thoracolumbar region in three cases and the lumbosacral region in two cases. After removal of a lamina of the hemivertebra, the body of the hemivertebra was visualized easily because the spinal cord had deviated to the concave side of the curve. The vertebral body, along with its cranial and caudal discs, was curetted with this approach. Thereafter, bone chips were grafted into the defect created by vertebrectomy. The results of this surgical procedure, especially those observed during long-term follow-up evaluation, were investigated.
For patients with a thoracolumbar hemivertebra, scoliosis improved from 49° ± 6° to 22.3° ± 3.5°, for a 54.3% correction. The correction ratio for kyphosis was 67.4%. Over an average 12.8-year follow up period, loss of scoliotic curvature correction was only 3.7°. In contrast, the hemivertebral correction ratio for patients with a lumbosacral hemivertebra remained 32.5% because of difficulty using internal fixation associated with patient age. At the most recent follow-up assessment, one patient exhibited deterioration of coronal spinal balance.
The described procedure was less invasive because it avoided an anterior approach, yet it yielded satisfactory long-term results for thoracolumbar hemivertebrae.
Eight patients with spinal deformity involving a hemivertebra underwent surgery using the aforementioned procedure and were followed up for more than 10 years. One of these patients required an additional operation after loosening of internal fixation. The deformity of another patient involved a wedged vertebra rather than a hemivertebra. Still another patient could not be followed because of her move to a new address. After the exclusion of these three patients, the study included five patients.
Of the five patients in the study, four were boys and one was a girl. Their mean age at the time of surgery was 10 years (range, 3.6–13.7 years). The mean follow-up period was 12.8 years (range, 11.5–14.9 years). The hemivertebra involved the thoracolumbar region in three cases and the lumbosacral region in two cases (Table 1). All of the patients with a thoracolumbar hemivertebra exhibited regional kyphosis. To determine operative invasiveness, volume of blood loss was reviewed from the clinical records along with transfusion and operative time. The correction ratios of both the main curve in the standing anteroposterior film and the kyphosis in the lateral standing film were evaluated. The correction ratio of the compensatory curve created above or below the main structural curve also was investigated. Loss of correction for both the main curve and the kyphotic angle in the sagittal plane were reviewed. For cases of lumbosacral hemivertebra, a perpendicular line was drawn from the center of the C7 vertebral body, and the distance from this line to the center of the sacrum was measured to examine pre- and postoperative spinal balance.
After induction of general anesthesia, the patient was placed in the prone position with the abdomen relieved of all pressure on rolls. The back was prepared and draped in routine fashion. A longitudinal skin incision was made on the back at the center of the hemivertebra. Paravertebral muscle was retracted laterally, and the lamina was explored. The lamina of the hemivertebra was identified and removed with its attached transverse process. Epidural bleeding was controlled with thrombin-soaked Gelfoam (Pharmacia Corp., Peapack, NJ). The dural sac usually had deviated to the concave side of the curve. Generally, the hemivertebrae in the thoracolumbar region also had deviated dorsolaterally because of the kyphotic deformity. The vertebral body of the hemivertebra could therefore be identified posteriorly. Because the pedicle of the hemivertebra was thicker than usual, cancellous bone in the vertebral body was curetted easily via the pedicle. Thereafter, the body of the hemivertebra was resected completely with a high-speed drill. In the pediatric spine, these procedures can be performed easily because the vertebral body is encapsulated by cartilage and periosteum, which is thicker than in the adult spine.
After the vertebral disc and endplate in both the cranial and caudal adjacent segments had been curetted, bone chips obtained during curetting of the vertebral body were inserted in the defect. For thoracolumbar hemivertebrae a Harrington compression rod then was applied to the convex side of the curve and a Harrington distraction rod to the concave side (Figure 1). The facet and the lamina to which internal fixation had been applied were decorticated on both the concave and convex sides of the curve. All the bones removed during the laminectomy were used as graft material throughout the area along with the bones grafted from the iliac crest.
Internal fixation was not performed in one 3-year-old patient with a lumbosacral hemivertebra. In this case, after curettage of the vertebral body without removal of disc tissue either above or below the vertebral body, a cast was applied to correct the deformity without any bone grafting. In another patient with a lumbosacral hemivertebra, the discs above and below the vertebral body were curetted, and interbody bone graft was performed with application of a compression device to the convex side of the curve alone.
Blood loss in this procedure ranged from 110 to 1360 mL (mean, 660 mL), and the volume of transfusion averaged 200 mL. The operative time ranged from 225 to 425 minutes (mean, 350 minutes). In the patients with a thoracolumbar hemivertebra, the main structural curve was corrected from 49° to 22.3° on the average, and the correction ratio ranged from 46.9% to 60% (mean, 54.3%). For compensatory curves, 31.4% correction was obtained in the upper curve and 61.3% in the lower curve (Table 2). For sagittal curvature, thoracolumbar regional kyphosis was corrected from 48° to 15° on the average, and the correction ratio was 67.4% (range, 58–77.6%). At the most recent follow-up visit, correction loss averaged 3.7° (Table 3).
In the patients with a lumbosacral hemivertebra, scoliosis was corrected from 34.5° to 23.5°, and the correction ratio averaged 32.5%. The upper compensatory curve exhibited 41.4% correction (Table 2). In this type of deformity, decompensation of spinal balance becomes a problem. Table 4 shows the changes in coronal spinal balance from before to after surgery as well as the change observed at the most recent follow-up visit. In one patient decompensation of spinal balance had completely recurred from regrowth of the hemivertebral body.
From the *Department of Orthopaedic Surgery, Osaka City University Medical School, and the
†Department of Orthopaedic Surgery, Osaka City General Hospital, Osaka, Japan.
Acknowledgment date: November 13, 2000.
First revision date: March 15, 2001.
Second revision date: May 29, 2001
Acceptance date: June 4, 2001.
Device status category: 1.
Conflict of interest category: 12.
Address reprint requests to
Hiroaki Nakamura, MD
Department of Orthopaedic Surgery
Osaka City University Medical School
1-4-3 Asahi-machi Abeno-Ku