Staged cervical osteotomy:a new strategy for correcting ankylosing spondylitis thoracolumbar kyphotic deformity with fused cervical spine

A 36-year-old man with thoracolumbar kyphoscoliosis presented to our clinic. He was diagnosed with ankylosing spondylitis at the age of 16 and spine deformity which gradually progressed to a degree where it was impossible for him to stand straight. The preoperative thoracic kyphosis (TK), thoracolumbar kyphosis (TLK), lumbar lordosis (LL), and sagittal vertical axis (SVA) were 93.8°, 30.8°, − 10.3°, and 259 mm, respectively. Although having severe thoracolumbar kyphotic deformity and an unmovable neck, the patient was still able to look horizontally with a CBVA of 21°. His cervical spine was totally fused (Figs. 1 and 2).

A two-stage surgical plan was managed for this patient. For the first stage, the aim was to correct thoracolumbar kyphoscoliosis and restore sagittal balance without consideration of the CBVA. The required correction angle was calculated by Song’s⁷ and Zheng’s⁴ methods. The proper correction was 110° for this patient (Fig. 3). However, if we perform the osteotomy as calculated, the patient will not have a normal visual field and look upward. According to Song’s and Zheng’s research, the osteotomy angle should not be larger than CBVA + (PT − tPT). So the correction angle has to be smaller than 49°. This degree of correction was not able to correct the kyphotic deformity, and the sagittal plane was still in an imbalance condition postoperatively (Fig. 3). So we determined to make two-level osteotomies for correction of 80°. Interrupted two-level osteotomy, using the technique of vertebral column decancellation and asymmetrical osteotomy, was performed at T12 and L2 for 30° and 50°, respectively. The second-stage cervical osteotomy was planned to be performed 3 months later to make the patient have a horizontal visual field.

The main concern about the correction angle of the second cervical surgery was the CBVA. The aim of the second-stage surgery was to make the patient have a normal visual field. In this case, the CBVA before the second surgery was − 21.7°. According to Song’s research, the optimal CBVA should be 10° to 20°. Concerning the high risk of cervical osteotomy, we planned to correct the CBVA to 10° postoperatively. Therefore, a 30° osteotomy angle should be performed at the cervical spine. We measured the antero-posterior diameter of the C7 vertebra. In order to make a 30° correction angle of the wedge osteotomy, the length of anterior edge should be 1.2 cm.

Under general anesthesia and with a halo attachment stabilizing the head, the patient was placed in a supine position. A cervical anterior approach was used for the first step to perform an anterior osteotomy at C7. Considering the scoliosis on the cervicothoracic spine, a coronal angle was preserved during osteotomy. The anteriorly closing wedge osteotomy was performed using the posterior vertebral wall as a hinge (Fig. 4). Gelfoams were filled into the osteotomy space to stop the bleeding, and a gauze was put in front of the vertebrae to prevent esophageal injury.

Then the patient was placed in a prone position. Pedicle screws were placed at C4, C5, C6, T2, T3, and T4. Laminectomies were performed from C6 to T1 (Fig. 4). The pedicles of C7 were resected. Then the C7 and C8 nerve roots were explored insuring the absence of compression. A flexible iron wire was put on the left screws to avoid excessive flexion during reduction. A short titanium rod was put on the right cervical screws for temporary fixation and was held by a clamp for controlling sudden neck movement. The osteotomy was closed mainly by manipulation of the halo. When the dura tends to be extended, reduction was stopped. Posterior stabilization was achieved by a pedicle screw and titanium rod system. Then the nerve root exploration was performed again to ensure they were not compressed.

Finally, the patient was placed in a supine position again. After anterior exposure, the osteotomy site was found to be completely closed. Anterior stabilization was achieved by a cervical spine locking plate. Spinal cord monitoring was used throughout the procedure, and change of monitoring was not detected during operation.

Several osteotomy techniques have been described to restore sagittal balance for AS kyphotic deformity. To treat severe kyphotic deformity with TLK combined with loss of lumbar lordosis, a two-level osteotomy should be applied [4]. In a previous study, Song and Zheng et al. [7] chose hilus pulmonis as the center of gravity for AS TLK and put forward an accurate method for calculating the exact angle required for one-level spinal osteotomy. Zheng et al. [4] provided a further method to calculate the individualized exact angle for two-level osteotomy. In this method, theoretical individual PT is calculated by preoperative PI and is used to achieve a pelvic neutral position, and the center of gravity is designed to be shifted to a pelvic neutral positional plumb line for sagittal balance.

In the reported case, the individualized osteotomy angle was calculated. The preoperative CBVA was 21°, which was almost within the most suitable range. If an osteotomy angle of 110° was performed, the patient will not able to see anything in front of him. If considering the CBVA, the osteotomy angle should not be larger than 49° or severe kyphotic deformity and sagittal imbalance will remain consequently. Therefore, in order to correct sagittal imbalance, a compromised method was adopted. An interrupted two-level osteotomy was performed at T12 for 30° and L2 for 50° to restore sagittal balance. After surgery, SVA was reduced from 25.9 to 13.1 cm. As expected, the postoperative CBVA came to be − 21.7°. The problem of visual field was designed to be settled by the second-stage cervical osteotomy.

Cervical osteotomy was a big challenge because of the difficulty and high risk of neurovacular complications [8, 10]. Performing PSO at the cervicothoracic junction is recommended to correct cervical kyphosis [10‐12]. However, cases of cervical osteotomy correcting hyperlordosis were seldom reported. Sengupta et al. [1] reported a case of flexion osteotomy of the cervical spine. The author performed posterior transverse osteotomy in a right lateral decubitus position initially. And then an anteriorly based wedge was resected from C7 in the same position. An anterior cervical spine locking plate was used for achieving stabilization. Kose et al. [9] reported three cases of anterior closing wedge osteotomy for the treatment of cervical hyperlordosis. The anterior vertebral column was exposed firstly, and C7–T1 closing wedge osteotomy was performed using the disc as a hinge. Then, the posterior fixation and reduction were completed by a posterior approach. Finally, the patient was turned supine again and an anterior plate was applied.

We used classical prone and supine positions and managed anterior-posterior-anterior approach osteotomy for the current case. An anterior wedge osteotomy was performed on the C7 vertebra with the hinge at the posterior vertebral wall. In the patient reported in this study, the cervical spine was fused. Therefore, although the “V”-shaped osteotomy was performed and a part of the anterior and middle columns was resected, the fused posterior column was still able to provide stability for the cervical spine. Additionally, after osteotomy, it was easy to isolate the esophagus and vertebral body, which may decrease the risk of esophageal injury during reduction. Under a cervical post approach, three steps were achieved. Firstly, implanting pedicle screws and rods provides stability after osteotomy. Secondly, decompression of the spinal cord may decrease the risk of neurological deficit. Thirdly, the spinal cord is under direct vision during reduction, which is also a benefit to avoid neurological complications. With the posterior vertebral wall closing, the spinal cord traction was limited. Also, the bone-on-bone contact provides stiffness and higher fusion rates. A stable fixation is also an important issue. We used both anterior and posterior fixation. So the third step of the anterior approach is necessary.

To achieve the goal of reducing cervical lordosis, neither the posterior-only approach nor the anterior-only approach is feasible. The posterior-only approach is not able to shorten the anterior column. It will take risks of spinal cord extension and neurological complications as well. The anterior-only approach makes it easy to perform osteotomy at the anterior and middle columns. But due to the fused posterior column, the process of reduction is hard to perform. Therefore, the anterior and posterior combined approach is required. The combined approach destroys the stability of the cervical spine for correction by osteotomy and decompression.

In most cases, making a smaller correction angle is able to ensure the patient has a horizontal visual field. However, the case we report was special. Before correcting thoracolumbar kyphosis, the patient’s CBVA was normal. Due to the fused cervical spine, once we correct the TLK, no matter how large the correction angle, the patient will look up to the sky. So there was no alternative to make a two-stage surgical planning. The current treatment strategy is suitable for severe cases like thoracolumbar kyphosis combined with cervical ankylosis. In this group of patients, correcting a smaller angle is still not able to guarantee a normal visual field, and a staged osteotomy strategy is more suitable.