Minimally Invasive versus Conventional Open Surgery for Fixation of Spinal Fracture in Ankylosed Spine

Introduction: This was a retrospective study aimed to investigate the perioperative outcomes of long construct minimally invasive spinal stabilisation (MISt) using percutaneous pedicle screws (PPS) versus conventional open spinal surgery in the treatment of spinal fracture in ankylosing spondylitis (AS) and diffuse idiopathic skeletal hyperostosis (DISH). Material and Methods: Twenty-one patients with AS and DISH who were surgically treated between 2009 and 2017 were recruited. Outcomes of interest included operative time, intra-operative blood loss, complications, duration of hospital stay and fracture union rate. Results: Mean age was 69.2 ± 9.9 years. Seven patients had AS and 14 patients had DISH. 17 patients sustained AO type B3 fracture and 4 patients had type B1 fracture. Spinal trauma among these patients mostly involved thoracic spine (61.9%), followed by lumbar (28.6%) and cervical spine (9.5%). MISt using PPS was performed in 14 patients (66.7%) whereas open surgery in 7 patients (33.3%). Mean number of instrumentation level was 7.9 ± 1.6. Mean operative time in MISt and open group was 179.3 ± 42.3 minutes and 253.6 ± 98.7 minutes, respectively (p=0.028). Mean intra-operative blood loss in MISt and open group was 185.7 ± 86.4ml and 885.7 ± 338.8ml, respectively (p<0.001). Complications and union rate were comparable between both groups. Conclusion: MISt using PPS lowers the operative time and reduces intra-operative blood loss in vertebral fractures in ankylosed disorders. However, it does not reduce the perioperative complication rate due to the premorbid status of the patients. There was no significant difference in the union rate between MISt and open surgery.


INTRODUCTION
Spinal fractures in ankylosing spondylitis (AS) patients are four times more common than the general population with lifetime incidence ranging from 5% to 15% [1][2][3][4] . Seventy percent of vertebral fractures in AS were associated with spinal cord injury at initial presentation following a trauma5. Mortality rates ranged from 7 to 32% 1,5,6 . Patients with diffuse idiopathic skeletal hyperostosis (DISH) had up to 1.5 times higher prevalence of vertebral fractures 7 . Conservative treatment plays little role in these fractures due to the long lever arm forces exerted at the fracture site. Nonsurgical treatment was associated with higher risks of complications such as pulmonary complications, thromboembolism and decubitus ulcers 8 . Therefore, many authors had described various surgical techniques such as posterior approach or combined anterior-posterior approach. However, posterior spinal stabilisation remains the most favored option 5,9,10,11 . Although combined anterior-posterior technique is advantageous from a biomechanical standpoint, it poses higher surgical risks 12 . In contrast, open surgery utilising posterior approach requires polysegmental long-construct instrumentation to counteract the forces acting at the fracture site. The role of minimally invasive stabilisation (MISt) using percutaneous pedicle screws (PPS) for vertebral fractures in ankylosed spine has not been elucidated. Previous reports on the role of MISt using PPS showed lower surgical risks i.e. reduced blood loss and operation time, but, in their report, only short segment stabilisation was investigated 13,14 . Recently, the use and advantages of long construct MISt using PPS in traumatic fractures and spinal metastasis have been investigated [15][16][17] . In this study, we would like to report the outcome of long construct MISt using PPS in the treatment of hyperextension spinal fractures in the ankylosed spine.

MATERIALS AND METHODS
We retrospectively reviewed patients with AS and DISH who were treated for spinal fractures in a single tertiary institution from 2009 to 2017. Ethical approval was obtained. Inclusion criteria were patients who had underlying AS or DISH, who presented with vertebral fractures and treated surgically with a long construct spinal fixation (stabilisation of at least three levels above and three levels below the fractured vertebra), either by open surgery or MISt using PPS (Fig. 1). All patients underwent a computed tomography (CT) scan of the spine prior to surgery. A total of 21 patients were included.
All patients were positioned prone on a four-post frame on a Jackson table to allow good visualisation of the spinal radiographic anatomy on anteroposterior (AP) and lateral fluoroscopic views. One of the most important surgical pitfalls in AS and DISH surgery is sagittal malalignment and neurological injury during positioning and surgery. Careful, gentle positioning of a patient with an ankylosed spine is extremely important as excessive movement over the fracture site can lead to sagittal malalignment and neurological injury. In patients with hyperkyphosis, a Wilson frame might be useful to accommodate the kyphotic alignment of the spine. The height of the Wilson frame can be adjusted to allow in-situ fusion while preventing sagittal malalignment and fracture displacement which may cause neurological injury. PPS were performed simultaneously on both sides by two surgeons.
A true AP view of the corresponding vertebra was obtained, in which both superior and inferior endplates were parallel and both pedicles were equidistant from the spinous process (Fig. 2a). A skin incision of 2cm was made just lateral to the lateral edge of the pedicle for the thoracic spine and 1-2cm lateral to the lateral border of the pedicle in the lumbar spine. The fascia was incised, and the muscles were split parallel to its fibers. Two 11G trocars were positioned at the lateral edge of the pedicle (right: 3 o'clock, left: 9 o'clock) (Fig. 2b). However, different starting point were chosen for the upper thoracic level, T1-T6 vertebrae (right: 2 o'clock, left: 10 o'clock) as described by Kwan et al 18 . The trocar was then advanced until the tip of the trocar approached the medial wall of the pedicle on AP view (Fig. 2c). A lateral view was obtained. On the lateral view, the tip of the trocar should be at or slightly deeper than the posterior vertebral border (Fig.  2d). The trocar was then advanced until the middle of the vertebral body (Fig. 2e). A guide wire was inserted. The screw was then inserted along the direction of the guide wire, while avoiding inadvertent guide wire advancement (Fig.  2f). Once the screw position was confirmed with the lateral fluoroscopic image, the guide wire was removed. Similar steps were repeated for the rest of the planned instrumentation vertebrae (Fig. 3a).
For open surgery, the pedicle screws were inserted using 'freehand technique' in the thoracic and lumbar spine. In the cervical region, lateral mass screws were inserted.
Rods were contoured to allow in situ fixation of the fractured vertebra without any correction of preexisting deformity. To allow in situ fixation, we have developed an extension of the screw sleeve that would mimic the final position of the rod when seated in the screw head. The rod was then contoured with all screw sleeves and extensions positioned in a parallel alignment. At the proximal thoracic junction, rods were inserted from a caudad to cephalad direction whereas for the thoracolumbar or lumbosacral junction, rods were inserted from a cephalad to caudad direction (Fig. 3b). Nuts were inserted. Final tightening of the whole construct was performed (Fig. 3c) and lastly, deep fascia and skin were closed.
Post-operatively, patients with intact neurology were allowed to sit up and to ambulate to washroom with an external brace once the pain was tolerable. Continuous bladder drainage tube was removed when patients start to ambulate. Post-operative analgesia comprised of patientcontrolled analgesia with morphine, oral celecoxib and acetaminophen. On day 2 or 3 post-operatively, surgical wounds were inspected. All patients were required to wear an external brace for three months. In the presence of neurological deficit, the post-operative rehabilitation protocol was decided by the spine rehabilitation team. AP and lateral standing radiographs were taken before discharge (Fig. 4).
All patients underwent CT scans between four to six months post-operatively. Fracture union was assessed based on the sagittal, coronal as well as axial images. A fracture union was defined as presence of bridging trabeculae across the fracture site within the vertebral body or formation of marginal or non-marginal syndesmophytes across two vertebral levels 19 (Fig. 5).
Data collected included age, gender, diagnosis, level of injury, fracture type according to AO classification, type of surgery (open surgery or MISt), level of instrumentation, number of instrumentation levels, pre-operative and postoperative neurological function according to Frankel grade, American Society of Anesthesiologist Physical Status Classification (ASA) and Charlson comorbidity index (CCI). The perioperative outcomes that were recorded included operative time, intra-operative blood loss, complications, duration of hospital stay and union rate.
Student's t-test was used for comparison of continuous variables while chi-squared tests were used for comparison of categorical variables between open surgery and MISt. Statistical analysis was performed using IBM SPSS Statistics for Windows, version 24.0 (IBM Corp., Armonk, N.Y., USA) with statistical significance, p value <0.05.

RESULTS
The mean age for this cohort was 69.2 ± 9.9 years.

DISCUSSION
Both AS and DISH cause similar clinical manifestations; including spinal stiffness, increased bone fragility, and difficulty with balance and gaze resulting in a higher risk of vertebral fractures 1,14 . Owing to the inherent instability of the spine resulting from ossification of surrounding soft tissues, vertebral fractures in the ankylosed spine are prone to neurological deficits 1,5,20 . Conservative treatment plays little role and is associated with poorer outcomes with spinal malalignment and persistent instability resulting in neurological impairment 21,22 .
Various surgical strategies have been described (Table IV) The findings in our study were comparable to previous studies with shorter operative time, lower intra-operative blood loss and shorter hospital stay in the MISt group. No neurological deterioration was encountered. Complications and union rate were comparable to previous reports. These findings could be attributed to the utilisation of multiple "keyhole" incisions and percutaneous pedicle screws insertion which required less muscle dissection, reduced bleeding rate, reduced time for hemostasis and wound closure, increased screw placement accuracy and faster recovery. In our study, it is important to note that there was one mortality in the open surgery group compared to none in the MISt group. However, we did not find any significant difference in the perioperative complication rate between both MISt and open surgery groups. This could be probably due to a small sample size.
The number of instrumentation level was not fully investigated. Most studies only included short construct fixation (mean number of instrumentation level, Sapkas et al 23  There were some limitations in this study. The sample size was small because the incidence of vertebral fractures in ankylosing disorders was low. Secondly, post-operative functional outcomes of patients were not reported. Further studies should be organised prospectively to evaluate the long-term functional outcomes in this group of patients. Although there were no significant differences in patients' demographics between patients who had MISt versus open surgery, this was a non-randomised sample. Further prospective randomised study might be more useful to investigate the true differences between MISt and open surgery in treating spinal fractures among patients with ankylosed spine.

CONCLUSION
MISt using PPS lowered the operative time and reduced intra-operative blood loss in fixation of vertebral fractures in AS and DISH. However, it did not reduce the perioperative complication rate because of the patients' premorbid status. There was no significant difference in the union rate between MISt and open surgery.