We analyzed the medical records of 325 consecutive patients who underwent subaxial cervical spine surgery (subaxial group) and who underwent myelography for thoracolumbar disorders (thoracolumbar group) at our institute from December 2012 to June 2020. The study’s protocol was approved by the institutional review board and informed consent was obtained from all participants. The subaxial group included those with their main lesion from C3 to C7, and the thoracolumbar group included those with their main lesion from T1 to L5. In both groups, those with rheumatoid arthritis and congenital skeletal anatomy were excluded because they had been already known as risk factors of a HRVA. In addition, those with pyogenic spondylitis, spinal tumor, spinal injury, and patients with previous cervical spine surgery were excluded (39 patients in the subaxial group and 71 patients in the thoracolumbar group, Figure 1).
Ultimately, the subaxial group included 94 patients (67 men, 27 women, mean age 61.9 years) and the thoracolumbar group included 121 patients (71 men, 50 women, mean age 67.1 years).
The subaxial group included 48 patients with ossification of the posterior longitudinal ligament (OPLL), 38 with cervical spondylosis, 3 with cervical spondylotic radiculopathy, 3 with cervical spine kyphosis, and 2 with cervical spondylotic amyotrophy. The surgical method in the subaxial group was posterior fixation including C2 (46 patients), posterior fixation above C3 (25 patients), anterior fusion (16 patients), and posterior laminoplasty without fixation (7 patients). In 46 patients who underwent posterior fixation including C2, the instruments used for C2 were as follows: PS was applied bilaterally to 20 patients, unilaterally to 2 patients, PS and laminar screw on each side to 21 patients, pars screw and laminar screw on each side to 2 patients, and laminar screw bilaterally to 1 patient.
The thoracolumbar group included 82 patients with lumbar spinal canal stenosis, 17 with degenerative lumbar kyphoscoliosis, 15 with thoracic ossification of the yellow ligament (OYL), and 7 with thoracic spondylotic myelopathy.
We performed preoperative 3-dimensional computed tomographic angiography (3D CTA) for all the patients in the subaxial group and CT myelography for all the patients in the thoracolumbar group. For all the patients in each group, we searched for HRVAs in axial slices (1 mm of thickness) that were along the orthogonal horizontal plane (Figure 2). We defined a HRVA as a maximum C2 pedicle diameter of <3.5 mm on the axial image, because the minimal diameter of the commonly used screw is 3.5 mm.
In addition, we evaluated osteoarthritis of the atlantoaxial (C1/2) facet joints in coronal plane of CT of all the patients in each group. Assessment of facet joint osteoarthritis was carried out with a grading scale as described previously [8,9]. Grade 0 indicates a normal facet joint, grade 1 shows joint space narrowing, grade 2 shows narrowing and sclerosis of facet joint, and grade 3 shows narrowing, sclerosis and osteophytes (Figure 3).
The sex, age of patients, body mass index (BMI), osteoarthritis grade of C1/2 facet, and prevalence of a HRVA in the 2 groups were compared using an unpaired Student t test, a Pearson χ2 test and a Mann-Whitney U test as appropriate. Logistic regression was used to identify the risk factor for a HRVA. Age, sex, BMI, and osteoarthritis grade of C1/2 facet joints were added as independent variables for multiple logistic regression analysis. Statistical significance was considered established at p < 0.05 and all the analyses were performed using IBM SPSS Statistics for Windows (version 26.0).
The instruments used for C2 pedicles with a HRVA in the subaxial group were analyzed. For all the patients in the subaxial group, we analyzed neurovascular injury as a complication of surgery derived from screw insertion.