History and physical examination
A 15-year-old female diagnosed with AIS presented with a progressive curve classified as Lenke type 1A. The right thoracic Cobb angle was 56° (T6-T12), with a nonstructural left lumbar curve of 23° before her initial surgery (Figure 1A). She experienced menarche at age 12. She had an unremarkable past medical and family history and on physical examination, the forward bending test revealed a right thoracic hump. Her neurologic examination was normal. Magnetic resonance imaging and computed tomography (CT) myelogram showed no abnormal findings in her spine.
The patient underwent posterior spinal fusion (T6-L2) and three-level thoracoplasty (9-11th ribs) with 6.3 mm titanium rod and screw and hook instrumentation, sublaminar wiring, and autogenous rib and local bone grafting. The surgery substantially corrected her deformity, leaving her with a right thoracic scoliosis of 24°. She did well after the surgery until she started experiencing pain secondary to prominent implants in the upper thoracic spine one year after surgery. Three years after the initial surgery, at age 18, her back pain had worsened due to metallic prominence, and she wanted to remove her implants. We agreed to remove her implants, because our group had concluded at that time that the spinal implants should be removed after bony fusion to avoid corrosion of the spinal implants . At the time of examination just prior to the planned implant removal, her Cobb angle was noted to have increased 5, to 29° (T6-T12) (Figure 1B).
Implant removal was performed, but part of the sublaminar wiring could not be removed due to solid bony union surrounding the wires. Intraoperative exploration of the fusion mass revealed no evidence of pseudarthrosis. One month after removal, her right thoracic curve was 33°, representing an additional 4° loss of correction (Figure 2A). The patient failed to follow up, and we had no further contact with her until she returned for care 18 months after implant removal after noting a recurrence of her deformity.
Comparison of radiographs at that time with radiographs prior to implant removal demonstrated significant progression of her thoracic curve, from 29° to 57°. The kyphosis between T5 and T12 also had increased, from 45° to 62° (Figure 2B). She had no neurologic deficits since the time of the initial instrumentation. On examination, the thoracic curve demonstrated minimal flexibility on preoperative bending films (Figure 3). Preoperative three-dimensional CT scanning revealed partial clefts between T11 and T12, but the CT images were inconclusive as to whether the fusion mass had a complete lack in continuity or non-union (Figure 4). She was assumed to have solid fusions at this time.
During the third operation, the fusion mass was completely exposed and explored (Figure 5). Although no pseudarthrosis had been identified initially, more meticulous exploration using a nibbler and rongeur discovered left partial clefts of the fusion mass at T11–T12 above her previous posterior fusion, though no motion within the fused segments was detected. No other site of possible pseudarthrosis was found. Because of its rigidity, the deformity required Smith-Peterson osteotomy of the fusion mass and revision posterior fusion with instrumentation using 6.3 mm titanium rods at T5-L2. A CT-based navigation system was used to perform the osteotomy and to insert pedicle screws due to the lack of normal anatomical structures. The surgery reduced her scoliosis to 18° (Figure 6).