Cervical pedicle screw (CPS) placement was introduced in the early 1990s. It offers greater fixation strength than lateral mass screw placement, as it allows fixation from the posterior column to the anterior vertebral body. CPS fixation is sometimes used in the treatment of various degenerative, traumatic and developmental spinal conditions for potentially excellent biomechanical fixation and deformity correction [1]. However, CPS fixation has been critiqued for its potential risks of serious injury to neurovascular structures such as the spinal cord, nerve root and vertebral artery. To avoid such serious risks, we have performed CPS fixation combined with posterior decompression under computed tomography (CT)-based navigation system [2,3,4]. There are many reports on short-term surgical results of CPS fixation, with only a few studies on mid-term results.

Chronic neck pain due to decreased lordotic angle of the cervical spine and adjacent segment degeneration (ASD) has been reported after cervical spine fusion in long-term follow-up [5]. The reports regarding ASD involved anterior cervical decompression and fusion (ACDF), while ASD after CPS fixation was not reported. Furthermore, there were few reports concerning axial pain after CPS fixation. This study aimed to evaluate midterm clinical and radiological results of computer-assisted CPS fixation combined with laminoplasty for cervical instability.

With approval of the investigational review board of our hospital (No. 2188), records of nineteen patients (13 males, 6 female; mean age, 58.3±13.0 years) who underwent cervical and/or upper thoracic (C2-T1) pedicle screw fixation, combined with posterior decompression for cervical instability, from September 1997 to November 2007, were analyzed. Consecutive patients with follow-up duration of five-year or more were included in this study. This surgical treatment was indicated for patients with destructive disease such as rheumatoid arthritis (RA) and destructive spondyloarthropathy (DSA) or cervical myelopathy caused by spinal cord compression and cervical instability.

Diseases included DSA in four patients, spine tumor in four patients, cervical spondylotic myelopathy with instability in three patients, cervical spondylotic myelopathy associated with athetoid cerebral palsy in three, RA in three and cervical trauma in two. The fusion levels were C2-C3 in one patient, C2-C5 in two, C2-C6 in two, C2-C7 in one, C3-C4 in two, C3-C5 in two, C3-C7 in four, C3-T1 in one, C4-C7 in one, C5-C6 in one, C5-C7 in one and C6-C7 in one. The mean number for the fused level was 2.8±1.4 intervertebral levels (range, 1 to 5 levels) (Table 1).

The follow-up period ranged from 60 to 156 months (mean, 90.2±34.5 months). None of the nineteen patients had died at the final follow-up. The average surgical time was 316±137 minutes (range, 122-600 minutes). The average blood loss was 721±1,080 mL (range, 100-4,600 mL). In total, complications were reported in four cases: C5 palsy, which was observed in two patients (10.5%); implant failure of screw pull-out, which was observed in one patient (5.3%); surgical site infection, which was observed in one patient (5.3%); and severe blood loss (>2,000 mL), which was observed in one patient (5.3%). None of these patients required a revision surgery, as the complications were not severe and there were no complaints.

The mean JOA scores (full score, 17 points) before and five-year after surgery were 9.0±4.6 (range, 0-15.5) and 12.8±4.1 points (range, 4.5-17 points), respectively. The mean JOA score at the five-year follow-up was significantly improved, compared to the mean preoperative score (p=0.001) (Fig. 1A). The improvement rate of JOA scores five-year after surgery was 51.8%±32.1% (range, 0-100%). Mean VAS for neck pain was 49.4±28.0 (range, 0-80) at five-year after surgery.

A total of one hundred screws were inserted transpedicularly from C2 to T1, with assistance of a CT-based navigation system. The rate of grade 3 pedicle screw perforations was 5.0% (5/100), whereas the combined rate of grade 2 and 3 perforations was 14.0% (14/100). There were no clinically significant complications such as vertebral artery injury, spinal cord injury or nerve root injury caused by any screw perforation. The evaluation was performed separately in C2, C3 to C6, and C7 to T1. The perforation rate of grade 3 was 18.2% (2/11) in C2. The perforation rate of grade 3 was 2.8% (2/72) in C3 to C6. The perforation rate of grades 2 and 3 was 13.9% (10/72) in C3 to C6. The perforation rate of grade 3 was 5.9% (1/17) in C7 to T1. The perforation rate of grades 2 and 3 was 11.8% (2/17) in C7 to T1. The mean C2-7 lordotic angle of the neutral position improved from 6.4°±14.7° (range, -11° to 39°) to 10.6°±13.4° (range, -7° to 41°) immediately after surgery and 7.8°±15.7° (range, -13° to 39°) at five-year after surgery, but these were not significant (p=0.144, 0.338) (Fig. 1B). We evaluated separately in categories of short fusion (<3 intervertebral) and long fusion. C2-7 lordotic angle in short fusion was improved from 2.6°±11.6° to 7.0°±13.9°(p=0.128) immediately after surgery and 4.9°±12.4° (p=0.176) at five-year after surgery. C2-7 lordotic angle in long fusion was improved from 9.1°±16.5° to 13.2°±13.0° (p=0.476) immediately after surgery and 10.0°±18.0° (p=0.656) at five-year after surgery. Bony union was noted in 94.7% (18/19) of patients. The rate of ASD in the follow-up period was 21.1% (4/19), but there was no patient who needed revision surgery for ASD.

Cervical spinal fixation using CPS for cervical instability caused by cervical trauma was first described Abumi et al. [9] and Jeanneret et al. [10] in 1994. The indications for CPS have since been extended and now include conditions such as RA, DSA and spinal tumor [11,12,13]. CPS can achieve better rigid fixation than other cervical fixation methods and can be combined with posterior spinal cord decompression. However, CPS insertion is technically demanding because of the narrow pedicle diameter, and there are risks of serious neurovascular complications including vertebral artery tear, spinal cord injury and nerve root injury [14]. To avoid these serious neurovascular complications, we used a CT-based navigation system [2,3,4]. CPS perforation rate in our study was as low as in previous reports (Table 2) [15,16,17,18,19]. In the patients selected for our study, the mean JOA score was significantly improved at five-year after cervical reconstruction by CPS fixation without serious neurovascular complications. In some cases, blood loss may be extensive and surgical time long. Blood loss and surgical times were greater during the introduction period of the CT-based navigation system. There was a learning curve for this procedure, and blood loss and surgical times were reduced with experience. The patient in whom 4,600 mL of blood loss was reported had multiple myeloma in the C4 vertebra. We treated this patient with posterior CPS fixation between C2 and C6. In particular, subtotal spondylectomy and vertebral body replacement were performed. As the blood loss during tumor resection was high, we did not consider that the blood loss resulted from vertebral artery injury. Another reason for the high blood loss could be that the surgeon was not experienced with CPS fixation, as this surgery was performed at the time of introduction of the procedure.

In previous reports, VAS for neck pain after cervical spinal surgery was 16 to 41 in patients with ACDF [20,21] and 9 to 49.7 in patients with laminoplasty [22,23,24]. VAS for neck pain was a little higher than previous reports in our study. We regarded the neck pain after surgery as a weak point of CPS fixation due to the necessity to release the lateral side of lateral mass and the longer skin incision, compared to laminoplasty. Andaluz et al. [5] reported a 30.8% incidence of chronic neck pain during long-term follow-up, 75% of which developed in patients with postoperative kyphosis. A significant relationship between postoperative C2-7 regional kyphosis and chronic neck pain was also noted. McAviney et al. [25] reported a normal range for the C2-7 lordotic angle to be 31° to 40°; and they found a statistically significant association between cervical pain and lordosis <20°. In our study, the C2-7 lordotic angle was improved after CPS fixation, but it was not a significant improvement. We think that cervical alignment correction by CPS fixation may not be enough, and it may be better to preoperatively correct the lordotic angle by Mayfield head holders with radiographic image.

ASD is a common complication after spine surgery in long-term follow up. ASD is defined by asymptomatic new radiographic evidence. Baba et al. [26] followed 106 patients after ACDF and found that more than 25% developed ASD within 8.5 years of follow-up. Gore and Sepic [27] reported that of 125 patients of anterior cervical fusion with five-year follow-up, ASD developed in 25% and preoperative ASD worsened in 25%. Although ACDF is often beneficial at the fused segment, it may initiate or accelerate ASD [26]. The risks of developing new symptoms secondary to adjacent segment pathology causing radiculopathy and/or myelopathy after cervical fusion surgery range from a cumulative rate of 1.6% to 4.2% per year. In our study, 21.1% of patients had ASD, as observed in the follow-up period. The rate of ASD in our study was lower than in previous reports regarding ACDF. No patient complained of symptoms due to ASD, and no revision surgery was necessary.

C5 palsy is another well-known potential complication of cervical spine surgery, with reported rates ranging from 0% to 30% [28,29]. The palsy is typically unilateral, but the condition is bilateral in 5% to 7% [30]. It is manifested by paresis of the deltoid and/or biceps brachii muscle, sensory deficits and/or intractable pain in the shoulders. In the present study, two patients (10.5%) had C5 palsy that eventually was improved in both patients.

The limitations of this study included a lack of control group, small sample size, various diseases, various fusion levels and retrospective design. On the basis of the analysis of mid-term outcomes in nineteen patients, we concluded that CPS fixation with a CT-based navigation system was effective in the treatment of cervical instability.

In our study, the mean JOA score significantly improved at five-year after cervical pedicle screw fixation without serious complications. Mean visual analog scale for neck pain was 49.4±28.0 at five-year after surgery. The rate of ASD in the follow-up period was 21.1% (4/19). Our midterm results showed that cervical pedicle screw fixation was useful for treating cervical instability. Severe complications could be prevented by using a CT-based navigation system.