Unilateral Versus Bilateral Pedicle Screw Fixation With Oblique Lateral Interbody Fusion in the Treatment of Multi-level Degenerative Lumbar Disease: a Prospective Study

Background: The purpose of this study was to compare the clinical and radiological outcomes of oblique lateral interbody fusion (OLIF) combined with unilateral (UPSF) and bilateral pedicle screw xation (BPSF) in the treatment of multi-level degenerative lumbar disease (LDD). Methods: We conducted a prospective study from May 2016 to June 2018. A total of 81 LDD patients were randomized into groups. The rst group involving 39 patients received OLIF combined with USPF (UPS group), and the second group of 41 patients received OLIF combined with BPSF (BPS group). The perioperative conditions included intraoperative bleeding, operating time, average hospital stay, and hospitalization expenses. The clinical outcomes included Visual Analogue Scale (VAS), Oswestry Disability Index (ODI), and complications. Moreover, the radiological outcomes included disc height, fusion rate, and cage subsidence rate. These perioperative conditions, clinical outcomes, and radiological conditions of the two study groups were evaluated and compared. All the patients were followed up for 24 to 28 months, with an average follow-up time of 26 months. Results: The intraoperative bleeding, operating time, and hospitalization expenses of the UPS group were signicantly less compared with those of the BPS group. Compared with preoperatively, the VAS and ODI scores of the two groups were improved signicantly after the operation. At 6 months after operation, the VAS score for back pain and ODI of the UPS group were better than in the BPS group (P < 0.05). Both groups were able to maintain the intervertebral space height of the fusion segment, and there was no signicant difference in the fusion rate at the 24 months. There were no signicant differences in complications. At the 24 months, there was no signicant difference in the cage subsidence of the fusion segment between the two groups. Finally, there was no loose screw and screw fracture in both groups.

long-term follow-up studies have found that BPSF is prone to complications, such as reduced bone mass in the fused segment, and adjacent segment disease [4]. Kabins et al. reported for the rst time that UPSF can achieve similar clinical e cacy and fusion rate as BPSF [5]. Yücesoy et al. found that BPSF signi cantly reduces lumbar range of motion (ROM) compared with UPSF through biomechanical studies [6]. Gu et al. conducted a prospective follow-up of 74 patients and reported that in MIS-TLIF, UPSF has the advantages of short operative time and causes less intraoperative blood loss, and is equivalent to BPSF in clinical e cacy and radiological outcomes [7].
Currently, reports on UPSF and BPSF are mostly focused on TLIF and improved TLIF, while OLIF studies are rare [8]. Wen et al. through a retrospective study documented that OLIF combined with UPSF is similar to BPSF in clinical e cacy. Unfortunately their research focused on single-segment surgery of the lumbar spine [9]. At present, there is no research on multi-level LDD with OLIF combined with different posterior xation. The previous nite element study of our research group found that the ROM of OLIF combined with UPSF is greater than BPSF, but the peak stress of the screw-rod and cage are far from reaching their yield strength [10]. In view of this, we combined the previous nite element research of our research group to further explore the application and clinical e cacy of OLIF combined with UPSF and BPSF in multilevel LDD.

Study design
This study was a prospective cohort study and was conducted at the General Hospital of Ningxia Medical University from May 2016 to June 2018. The Scienti c Research and Ethics Review Committee of the General Hospital of Ningxia Medical University approved this study on April 26, 2019 (NO.2019-038). All participants in this study provided informed consent. The patients were assigned to different groups using the random number table method. The surgical operation was completed by only one doctor (Zhaohui. Ge), and the surgeon remained blinded to the group assignment of the patients before the operation started. Two independent radiologists evaluated the radiologic data, and the nal result was the average of the two radiologists evaluations.

Patients
The inclusion criteria were: (1) the patient must have been over 40 years old; (2) diagnosed with lumbar degenerative diseases, such as degenerative lumbar scoliosis, lumbar spinal stenosis, lumbar disc herniation, lumbar spondylolisthesis (within °) and/or lumbar instability; (3) low back pain and/or radicular pain for more than 3 months and conservative treatment was ineffective, which seriously affected normal life; (4) need for multi-level fusion and xation; (5) and at least 24 months of complete follow-up. The exclusion criteria were: (1) responsible segments that required fusion included L5-S1; (2) severely obese patients (BMI ≥ 35); (3) a history of lumbar fusion internal xation and abdominal surgery.

Surgical procedure
After general anesthesia was successful, the patient took the standard right decubitus position, xed the armpit, pelvis and left lower limb with adhesive tape, and slightly bent the hip and knee. The operative segments were determined using a "C" arm X-ray uoroscopy machine, and a transverse incision was marked, with a length of about 3-5cm. After that, disinfection and laying of sterile towel were completed. The skin was cut, as well as the subcutaneous tissue and fascia sequentially, and the external oblique muscle, internal oblique muscle, and transversus abdominis were bluntly separated with a vascular forceps until the retroperitoneal space was reached. With the operator's index nger downward, and touching the internal wall of the iliac bone, the extraperitoneal fat was continued to be separated inward and downward bluntly, then the quadratus lumborum transferred and the extraperitoneal fat and ureter carefully separated forward. The upper and lower soft tissues were dissociated, and the extraperitoneal structures gently pushed forward after penetrating. The gap between psoas major and abdominal aorta was found by the end of index nger. Following the guidance of the index nger, the positioning guide needle was inserted into the anterior and middle 1/3 of the intervertebral space of the surgical segment for uoroscopy positioning.
Step by step, the casing was inserted and the working channel established, followed by the xing of the cold light source rmly, then the working channel opened appropriately, and then the intervertebral disc bluntly separated and exposed. After resection of the annulus brosus, an osteotome was inserted into the intervertebral space and the contralateral annulus brosus was punctured. Reamer and curette were used to treat the intervertebral disc tissue, including the cartilage endplate, but the intact bony endplate was retained. Different size cage models were used to expand the intervertebral space to a satisfactory height step by step. After lling the allogeneic bone with the cage (Clydesdale Spinal System, Medtronic, Inc., Minneapolis, MN, USA) of appropriate size, the cage was inserted into the intervertebral space after being xed with a silk thread to complete the segmental stretching, indirect decompression, and interface xation.
After decompression and fusion, the patient was changed to a prone position and the surgical area was disinfected. The space between the multi dus muscle and the longissimus muscle (Wiltse approach) was used for unilateral or bilateral pedicle screw xation (Medtronic, Inc., Minneapolis, MN, USA). Taking UPSF as an example: the "C" arm X-ray uoroscopy machine located the target segment. The skin and lumbodorsal fascia were cut longitudinally at 2.5cm next to the spinous process. The index nger was used to bluntly separate the gap between the multi dus muscle and the longissimus muscle. With the aid of the retractor, the unilateral facet joints were exposed. Weinstein method was used for positioning, a unilateral pedicle screw was implanted, a titanium rod of appropriate length was taken to connect the tail of the pedicle screw, and the tail cap was used for compression and xation. After washing the wound with saline, the wound was sutured layer by layer.

Postoperative treatment
Antibiotics were routinely used for 3 days after operation, and no drainage tube was placed in all the patients. The patients were instructed to exercise moderately on the rst 2-3 days after surgery. The two groups were given the same adjuvant medication and rehabilitation measures. The follow-up time was 6 months, 12 months, and 24 months after the operation.

Clinical and radiographical evaluation
The intraoperative blood loss, operation time, average length of stay, and hospitalization expenses were retrieved from the patient's electronic medical records. The data of VAS, ODI score, and intervertebral space height of fusion segment at 6 months, 12 months, and 24 months were collected. The fusion and cage subsidence of each segment at the 24th month of follow-up were recorded.
Disc height: the ventral and dorsal intervertebral space height of the fusion segment was measured on the standing neutral lateral radiographs, and the average value taken. Fusion status: mature trabeculae were observed between the endplate and cage. If the X-ray lm was not clear, the lumbar spine CT was used for evaluation [4].
Cage subsidence: this was the difference between the height of intervertebral space within 1 week after operation and that at the 24 months. Mild was de ned as the difference ≤ 1 mm, moderate was 1 mm to 3 mm, and severe was ≥ 3 mm [11].
Lumbar lateral radiographs with ≥1mm translucent bands on both sides of the screw were considered as screw loosening.

Statistical analyses
The Statistical Package for Social Science software was used for data processing (version 20, SPSS Inc., Chicago, IL, USA). The data are expressed as mean ± standard deviation. Fusion rate, cage subsidence rate, and complication rate are expressed as n (%). Student's t-test was used for comparison between groups and within groups. Chi-square test or Fisher's exact test was used for the comparison between rates. P < 0.05 was considered statistically signi cant.

Demographic data
There were no signi cant differences in age, gender, BMI, follow-up time, preoperative diagnosis, fusion segment, comorbidities, and smoking between the two groups (Table 1).   (Table 3). Complications occurred in both groups, such as pain and numbness in front of the left thigh, segmental vascular injury, iliopsoas/quadriceps weakness, and endplate damage. There was no statistical difference in the incidence of complications between the two groups, including single and overall complications (Table 4).  In terms of radiological outcomes, the height of the intervertebral space of the fusion segment after the operation in the two groups was signi cantly higher than before (p < 0.05). There was no signi cant difference between the two groups at each follow-up time point after surgery (Table 5). At the 24 months, the fusion rate of the BPS group was numerically higher than in the UPS group, but there was no signi cant difference between each segment (Table 6). Moreover, there was no signi cant difference in the cage subsidence rate of the fusion segment between the two groups at the 24 months (Table 7).

Discussion
With the aging of the population, the incidence of LDD has also increased signi cantly. Due to the asymmetric degeneration of the intervertebral disc and facet joints, the lumbar vertebral body has slipped scoliosis, and stenosis pathological changes. Patients often experience intractable low back pain, radicular pain, intermittent claudication, among other symptoms [12]. For LDD with severe symptoms, conservative treatment often fails to achieve satisfactory outcomes, and surgical intervention is required [13]. Posterior lumbar interbody fusion, such as PLIF and TLIF, has been widely used in LDD and good results obtained [14,15]. However, the paraspinal muscles need to be stripped for a long time during the operation, resulting in muscle brosis after the operation, which may cause intractable low back pain after the operation. Complications, such as cerebrospinal uid leakage and nerve root injury may also occur during the operation [16]. Since the process of OLIF decompression and fusion are performed through the natural gap between the psoas major muscle and the abdominal aorta, the posterior spinal canal structure is not disturbed, which greatly reduces the incidence of these complications [17]. For LDD that requires single-segment fusion, OLIF stand-alone can be selected, but multilevel disc degeneration needs to be xed with a posterior pedicle screw. At present, there are still controversies, whether the xation method should be UPSF or BPSF.
Cappuccion et al. found in the biomechanics of LLIF that BPS is more stable than UPSF, especially in the lateral bending and axial rotation of the spine [18]. Slucky et al. simulated TLIF combined with different pedicle screw xation on the cadaver, and found that BPSF reduces the lumbar ROM signi cantly than UPSF, but it can provide higher xation strength in the fused segment [19]. Harris et al. measured the overall and L4-5 ROM on the lumbar spine of the cadaver and found that the reduction in ROM of BPSF compared with UPSF was mainly re ected in the axial rotation and lateral bending [20]. In biomechanics research, we found that BPSF provides higher xing strength and stability than UPSF, especially in the lateral bending and axial rotation. In theory, although BPSF provides higher xation strength, it causes the ROM of the fusion segment to be too small, stress shielding, and compensatory increase the ROM and load of the adjacent segments. This causes bone loss in the fused segment and accelerates the degeneration of adjacent segments [21].
In vitro experiments cannot truly re ect the stress state of the lumbar spine under the condition of loadbearing, which requires further clinical evidence. After a follow-up of the clinical cases, some scholars believe that UPSF can provide su cient xation strength for a single segment. However, for multiple segments due to insu cient xation strength, UPSF easily causes cage displacement and is not suitable for multi-segment posterior xation [22,4]. Zhang et al. indicated that the clinical e cacy and fusion rate of UPSF is similar to that of BPSF, but with less operation time, blood loss, and hospitalization costs in patients with lumbar spine disease who underwent TLIF combined with posterior xation treatment [23]. Wen et al. found that OLIF combined with UPSF achieves similar fusion rate and clinical e cacy to that of BPSF in single-segment LDD, and the operation time and hospitalization costs are less [9]. However, the outcomes of OLIF combined with UPS in patients with multilevel LDD have not been investigated. Therefore, we conducted a retrospective study to compare the clinical and radiological outcomes of OLIF combined with UPSF and BPSF in the treatment of multi-level LDD.
In this study, the intraoperative bleeding, operating time, and hospitalization expenses of the UPS group were less than those of the BPS group. This could be because the UPS only exposed the posterior tissue of the lumbar spine on one side and implanted pedicle screws. Intraoperative bleeding was different from the results of Wen et al [9]. It is possible that all cases in this study had not used percutaneous xation. In OLIF multi-segment surgery, the decompression and fusion process were all performed under "C" arm Xray. If percutaneous pedicle screw xation was used, radiation exposure time would be increased again, which is a disadvantage to the doctors and patients.
Our study showed that the pain and lumbar function of the two groups were signi cantly improved at each follow-up time, but at 1 month and 6 months, the outcomes of the UPS group were better than those of the BPS group, and there was no signi cant difference between the two groups at the last follow-up. These ndings were inconsistent with previous results of a meta-analysis [24]. This could be because the meta-analysis included patients who received PLIF, TLIF, and MIS-TLIF. Compared with these surgical methods, OLIF could protect the function of the paraspinal muscles to the most signi cant extent [25]. The rate of complications between the two groups was not signi cantly different since the complications were related to the surgical approach of OLIF and were irrelevant to the posterior screw placement process. A total of 33 complications occurred in the two groups of the study, and the total incidence of complications was 40%, which was higher than that reported by previous studies [26,27]. We speculate that the all patients included in this study were multi-segment LLD patients, and the psoas major muscle was stretched for a long time during the operation [2]; hence, the high incidence of complications. Among them, degenerative lumbar scoliosis accounted for 64%, and patients with L2-5 accounted for 91%, which have been previously documented as factors that increase the incidence of lumbar plexus and segmental vascular injury [28,29].
The radiologic outcomes of this study showed that there was no signi cant difference in disc height, fusion rate, and cage subsidence between the two groups after surgery. These results were inconsistent with previous ndings [30,22]. We speculate that this could be closely related to the two surgical decompression mechanisms. OLIF is an indirect decompression, which does not interfere with the vertebral plate and lumbar facet joints during the operation. These tissues belong to the posterior ligament complex (PLC), which is important for maintaining the stability of the spine and affects the fusion rate of the cage [31]. The biomechanical study of Du et al. found that the length of the cage is very important for UPSF to maintain the stability of the spine, which had a small impact on BPSF. It was found that in order to make the UPSF xation strength of the segment close to that of BPSF, the length of the cage should be higher than the height of the intervertebral space [32]. The cage used for OLIF in this study was a Clydesdale type, and the length was much longer than the intervertebral space height of the fusion segment. Similarly, previous nite element studies have also con rmed that in a single segment, the xation strength of OLIF combined with UPSF is lower than that of BPSF, but when the lumbar spine moves in all directions, the stress peaks of its cage and pedicle screws are far from reaching their respective levels. Therefore, UPSF could provide su cient stability [10].

Conclusions
Overall, in more than 24 months of follow-up, OLIF combined with UPSF has similar clinical effect and segmental stability with BPSF. Compared with BPSF, UPSF has the advantages of short operation time, less intraoperative bleeding, and low hospitalization cost. Therefore, OLIF combined with UPSF is a feasible and effective method for the treatment of multilevel LDD.

Limitations
Firstly, as prospective single center study, the sample size in this study was relatively small. Secondly, the follow-up time was only about two years, relatively short to evaluate the long-term clinical results of these two different surgical procedures. Therefore, large sample sized studies with a long follow up period are required to provide more reliable clinical research data.

Consent for publication
Written informed consent for publication was obtained from the patient; each participant was informed that individual details and images would be published in our manuscript.
manuscript. RM, SL, and XD reviewed the manuscript. ZM,and PW are responsible for the integrity of the data. ZG approved the nal version of the manuscript. All authors have read and approved the nal submitted manuscript.

Figure 1
Case 1: A 68 year old man with low back pain accompanied with numbness of right lower limb for 1 year and aggravating for 3 months, underwent OLIF combine with UPSF. a, b Preoperative lumbar anteroposterior radiographs showed L2-L5 degenerative scoliosis, and MRI showed multilevel lumbar disc herniation. c, d Compared with preoperatively, the patient's symptoms were signi cantly reduced immediately after surgery. Lumbar radiographs showed that the cage was in good position, the height of the intervertebral space increased signi cantly, and the unilateral pedicle screw was xed in place. e, f 6 months after surgery, lumbar radiographs showed that the cage and UPSF were xed in place. g, h Computed Tomography (CT) at 24 months after surgery showed that all the surgical segments achieved bony fusion.

Figure 2
Case 2: A female, 72 years old, with low back pain and intermittent claudication for 1 year, underwent OLIF combine with BPSF. a, b The preoperative lumbar anterior radiograph showed degenerative scoliosis of L2-5 lumbar spine, and MRI showed multi-segment lumbar disc herniation. c, d Immediately after operation, anteroposterior radiographs showed that the endplate of L1-2 was slightly collapsed, the height of the intervertebral space in each segment increased signi cantly, and the BPSF was xed in place. e, f 12 months after operation, CT showed cage and BPSF xed in place. g, h 24 months after the operation, all the surgical segments achieved bony fusion and the cage had slight subsidence, but the patient did not complain of symptoms such as low back pain.