Clinical Signicance of PLIF in Restoring L4-S1 Segmental Lordosis of 2-level Isthmic Spondylolisthesis

Background: Spinal and pelvic sagittal plane balance is closely related to good clinical prognosis, so in the treatment of 2-level isthmic spondylolisthesis, attention should be paid not only to adequate nerve decompression, but also to the correction of lumbosacral sagittal plane parameters. The purpose of this study was to observe the clinical prognosis and sagittal parameters of patients with isthmic spondylolisthesis treated with PLIF, and to nd out the risk factors leading to poor prognosis. Methods: From January 2006 to August 2018, the clinical data of patients with 2-level isthmic spondylolisthesis treated with PLIF in the Second Xiangya Hospital of Central South University were retrospectively collected. The clinical symptoms (JOA score and VAS score) and the sagittal parameters of lumbosacral segment (PI, PT, LL, L4-S1 SL, LDI, PI-LL, LL and L4-S1 SL) were recorded before operation, immediately after operation and at the last follow-up. According to the improvement rate of JOA score, the patients were divided into two subgroups, poor(P) group and good(G) group. The parameters within and between the two subgroups were compared. Meanwhile, Pearson correlation analysis was conducted between sagittal parameters and JOA score improvement rate. Results: A total of 52 patients were enrolled in this study, the average age was (59.96 ±9.11) years, and the mean follow-up time was (31.88 ±8.37) months. Group G (n = 37) and group P (n = 15). In terms of clinical symptom improvement and sagittal plane parameters, except PI, the other parameters of the patients were improved compared with those before operation, and the difference was statistically signicant. In the intra-group comparison, except PI, other indexes in group G were signicantly improved, while in group P, there was no signicant difference in PI, LL, L4-S1 SL, LDI, PI-LL before and after operation. In the comparison between groups, there was no signicant difference in baseline data between group G and group P; postoperative VAS score(back pain) in group G was lower than that in group P, but there was no signicant difference in VAS score( leg pain); in terms of JOA score and JOA score improvement rate, group G was signicantly better; △ L4-S1 SL ,L4-S1 SL and LDI were larger in group G, and the proportion of patients with normal LDI was higher than that in group P. Pearson correlation analysis showed that postoperative △ L4-S1 SL, LDI and L4-S1 SL were positively correlated with JOA improvement rate. 2 patients with failed internal xation occurred in group P, and the postoperative LDI was less than 50%. Conclusion: PLIF is an effective method for the treatment of 2-level isthmic spondylolisthesis. ΔL4-S1 SL, L4-S1 SL and good LDI may be important sagittal parameters affecting the clinical prognosis of L4 and L5 isthmus spondylolisthesis. discs, and the nerve roots of L4, L5 and S1 were released. Intervertebral bone grafting was performed using autogenous bone or cage combination. Finally, the posterior compression operation was performed at the L4/5 and L5/S1 segments, and the bone graft surface was held tightly. LDI were more likely to have ASD than those with high LDI. In this study, the LDI of the 2 cases of internal xation failure was less than 50%, indicating that abnormal LDI will increase the incidence of mechanical complications. We found that the mean LDI and proportion of LDI within the normal range in the G group were higher than those in group P, and there was a positive correlation between LDI and JOA score improvement rate, indicating that for double level spondylolisthesis, increasing L4-S1 SL appropriately and paying attention to its matching with LL can improve the clinical prognosis of patients to a certain extent.


Introduction
Isthmic spondylolisthesis is one of the most common types of lumbar spondylolisthesis in adults [1]. It refers to the displacement of the upper vertebral body relative to the lower vertebral body caused by unilateral or bilateral interarticular fractures, resulting in low back pain, lower limb radiation pain, neurogenic claudication and other clinical symptoms. Lumbosacral 2-segmental isthmic spondylolisthesis is rare, with an incidence of 0.3-1.48% [2,3]. When conservative treatment fails, decompression, reduction and fusion are required to relieve nerve compression and correct deformity.
The safety and e cacy of transforaminal lumbar interbody fusion (TLIF) and posterior lumbar interbody fusion (PLIF) in the treatment of single-segmental spondylolisthesis have been con rmed. However, compared with single-segmental lesions, the sagittal deformities of double-segmental spondylolisthesis are more serious [4]. The sagittal balance of spine and pelvis is closely related to good clinical prognosis [5][6][7]. When the sagittal plane is out of balance, it will lead to compensatory mechanisms, such as cervicothoracic overextension, pelvic retroversion and knee exion, causing muscle fatigue and leading to chronic back pain and dysfunction [8,9].Therefore, in the treatment of 2-level spondylolisthesis, we should not only pay attention to adequate nerve decompression, but also pay more attention to the correction sagittal deformity.
In this study, the clinical data of 52 patients with L4 and L5 isthmic spondylolisthesis treated with PLIF in our department were analyzed retrospectively in order to nd out the risk factors leading to poor prognosis.

Study population
With consent of the ethics committee of the Second Xiangya Hospital of Central South University and informed consent of patients, we retrospectively collected the clinical data of 52 cases of L4 and L5 double-segmental isthmic spondylolisthesis who underwent surgical treatment in our hospital from January 2006 to August 2018. Inclusion criteria: (1) the diagnosis was con rmed as continuous L4 and L5 double-segmental isthmic spondylolisthesis; (2) there were de nite indications for surgery; (3) surgical treatment using PLIF technique and only L4-S1 internal xation and fusion. Exclusion criteria: (1) incomplete imaging data; (2) follow-up time < 24 months; (4) previous history of lumbar surgery, trauma, pelvis or lower limb disease, lumbar tumor, in ammation and coronal deformity (Cobb angle >10 °).

Surgical Procedures
All patients underwent PLIF by senior spinal surgeons. The patient was placed in prone position under general anesthesia. Along the posterior median incision, skin and fascia were incised in turn, paravertebral muscles were separated, and bilateral screw entry points were exposed. Then the L4 and L5 oating lamina, spinous process and facet joints connected with them were removed, and the soft tissue on the bone surface was removed as a subsequent bone graft material. Appropriate pedicle screws were implanted on both sides of L4-S1, and the L4/5 and L5/S1 disc spaces were treated with distraction and pull reduction, followed by the removal of L4/5 and L5/S1 discs, and the nerve roots of L4, L5 and S1 were released. Intervertebral bone grafting was performed using autogenous bone or cage combination. Finally, the posterior compression operation was performed at the L4/5 and L5/S1 segments, and the bone graft surface was held tightly.
Slippage percentage: the ratio of the sliding distance of the upper vertebral body relative to the caudal vertebral body to the length of the superior endplate of the caudal vertebral. LL: the angle between L1 upper endplate and S1 upper endplate. L4-S1 SL: the angle between L4 upper endplate and S1 upper endplate.
ΔLL and ΔL4-S1 SL were de ned as postoperative values minus preoperative values.

Data Collection and Clinical Evaluation
The clinical data of the patients were obtained from the medical records department and outpatient follow-up. Demographic data and Perioperative Parameters included age, gender, Body Mass Index (BMI), operation time, intraoperative blood loss, follow-up time, and fusion status.
The standard of bone fusion was that X-ray showed the formation of continuous bone bridge between adjacent vertebral bodies, or there was no obvious light transmission area between adjacent vertebral bodies on CT, or there was no abnormal activity on dynamic X-ray lm. VAS score and JOA score were used to quantify the clinical e cacy, JOA score improvement rate =100/100 (postoperative score -preoperative score) / (29-preoperative score).
In order to facilitate statistical analysis, according to the literature standard [10,11], we divided the LDI into two groups: the normal group (50% -80%) and the abnormal group (not within 50% -80%), and divided the improvement rate of JOA score into two subgroups: poor (P) group (<50%) and good (G) group (>50%).

Statistical Analysis
All statistical analysis was performed by SPSS 23.0 (SPSS, Chicago, Illinois, USA). The measurement data were expressed by mean ± standard deviation, the difference between groups was analyzed by independent sample t test, and the data before and after operation was analyzed by paired sample t test. The count data were expressed by frequency or constituent ratio, and the statistical analysis was performed by chi square test. Pearson's correlation coe cient was used to determine the correlation between imaging parameters and JOA score improvement rate. All tests were two-sided, and a P < 0.05 was considered statistically signi cant.

Demographic data and Perioperative Parameters
A total of 52 consecutive L4 and L5 isthmic spondylolisthesis cases were included in this study, including 38 females (73.08%). The age of the patients was (59.96 ± 9.11) years and the BMI was (23.12 ± 1.64) kg/m 2 . (Table 1) Among the perioperative parameters, the operation time was (216.00 ± 24.62) min, the blood loss was (810.38 ± 197.23) ml, and the follow-up time was (31.88 ± 8.37) months. The fusion rate was 86.54% at the last follow-up. ( Table 1) In terms of clinical symptom improvement and sagittal parameters, except that PI was comparable before and after surgery, other parameters were improved compared with that before surgery, and the differences were statistically signi cant (P < 0.05) ( Table 2 and 3).
Subgroup analysis according to JOA score improvement rate According to the improvement rate of JOA score, there were no signi cant differences in demographic data such as age, gender, BMI, preoperative sagittal parameters and symptom scale scores, and fusion rate between the two groups (Table 1,2,3).
In self-comparison before and after surgery, it was found that all the other indexes except PI in group G were signi cantly improved, and the differences were statistically signi cant (P < 0.05), while in group P, in addition to PI, we also found that there was no signi cant difference in LL, L4-S1 SL, LDI, PI-LL before and after operation (Table 3).
After operation at the last follow up, there were differences in the improvement of clinical symptoms and sagittal parameters between group P and group G.
We found that the VAS score of low back pain in group G was lower than that in group P (2.03 vs 4.60, P < 0.001), but there was no signi cant difference in VAS score of leg pain (P = 0.744); In the JOA score and JOA score improvement rate, group G was signi cantly better (P < 0.001). In terms of imaging parameters, L4-S1 SL and LDI were larger in group G (36.00 ° vs 29.33 °, P = 0.033 and 63.26% vs 52.51%, P = 0.015), and the proportion of normal LDI in group G was higher than that in group P; △L4-S1 SL was larger in group G, and the difference was statistically signi cant (11.89 °vs 5.80°, P = 0.029). Other parameters such as LL, PI-LL, PT, SS were not signi cantly different between group P and group G, but we can still nd that the improvement of parameters in group G is more obvious ( Table 1,3).

Correlation Analysis of JOA improvement rate
In the correlation analysis, we found no correlation between JOA score improvement rate and preoperative imaging parameters. After operation, we found that L4-S1 SL, LDI and △L4-S1 SL were positively correlated with JOA score improvement rate (r = 0.371, P = 0.042; r = 0.348, P = 0.011; r = 0.468, P = 0.035). Other parameters such as LL, PI, PT, SS were not signi cantly correlated with JOA score improvement rate ( Table 4).

Complication of internal xation
No internal xation failure occurred in group G, while 2 cases in the P group had postoperative internal xation failure, and their postoperative LDI was less than 50%. During the follow-up, 1 case had the phenomenon that the top screw migrated and cut the upper endplate ( gure 2), the patient refused to undergo reoperation and was under close follow-up. Revision operation was performed in 1 case of sacral screw fracture.

Discussion
The causes of isthmus spondylolisthesis are complex, including genetic, traumatic, mechanical and hormonal factors [12,13]. Among them, the abnormality of spine and pelvis parameters is an important factor in the occurrence and development of spondylolisthesis [14]. Therefore, for the surgical principles of spondylolisthesis, in addition to decompression, reduction, xation and fusion, attention should be paid to the reconstruction of lumbosacral sagittal balance. Zhou et al. [4] found that patients with isthmic spondylolisthesis have larger PI and L5I than asymptomatic healthy people, while patients with multisegmental spondylolisthesis have larger SVA, PT and smaller LL, L4-S1 SL than single-segmental spondylolisthesis, which means that surgery is more challenging in correcting sagittal deformities of 2-level spondylolisthesis.
Du et al. [15] treated 58 cases of L4 and L5 spondylolisthesis with TLIF. The postoperative PT, SS, LL, L4-S1 SL, SVA and other parameters were signi cantly improved as compared with those before operation, it was found that 2-level TLIF was more effective in restoring local and global sagittal balance. Song et al. [1] treated 32 cases of 2-level isthmus spondylolisthesis with PLIF, the excellent and good rate of JOA score recovery was 84.3%, and the fusion rate was 87.5%, which was similar to our results. In this study, the JOA score, VAS score and sagittal plane parameters were signi cantly improved compared with those before operation through more than 2 years follow-up, which once again demonstrated the effectiveness of PLIF in the treatment of 2-level spondylolisthesis.
We believe that the advantage of PLIF in the treatment of 2-level spondylolisthesis lies in [13,[15][16][17][18][19]: 1. after spondylolisthesis, the posterior lamina is almost oating, and is usually accompanied by severe articular process hyperplasia and osteophyte formation, removal of posterior structure is bene cial to the reduction; 2. fully enlarge the spinal canal and nerve root foramen to avoid squeezing the nerve root after reduction; 3. laminectomy can obtain more autogenous bone, which avoids the rejection of allogeneic bone and the complications of iliac bone removal, which is bene cial to the interbody fusion. 4.
double-segmental interbody fusion is bene cial to the restoration of intervertebral foramen height and lumbosacral sagittal plane; 5. discectomy, intervertebral height and sagittal plane reconstruction, fusion and internal xation can be completed in one stage.
In this study, the overall improvement rate of JOA after operation was (64.81 ±27.83) %. We divided the patients into two groups according to the threshold of 50% improvement rate of JOA score. Through comparative analysis, it was found that poor reconstruction of L4-S1 SL and LDI was an important factor affecting prognosis. Kuhta et al. [20] treated 57 cases of single-segment degenerative spondylolisthesis with TLIF and found that insu cient correction of SL during operation would lead to forward shift of center of gravity and affect clinical prognosis. Takahashi et al. [21] found that the insu cient increase of L3-L5 SL in the treatment of L3 and L4 double-segment degenerative spondylolisthesis by PLIF was related to the poor clinical prognosis. He et al. [22] believe that in the treatment of degenerative spondylolisthesis with PLIF, the recovery of SL and PT is related to the clinical effect, and the increase of SL and the decrease of PT may play a positive role in relieving low back pain after operation. In this study, Δ L4-S1 SL, L4-S1 SL in group G was higher than that in group P, and positively correlated with the improvement rate of JOA score, indicating the importance of reconstruction of L4-S1 SL in double-segmental spondylolisthesis for improving clinical prognosis. According to the concept of the "core economic cone" [23], when the body is within the smallest cone, the muscle consumes the least energy, when the radius of the cone increases, the muscles of the back and buttocks must be activated to maintain body balance. L4-S1 is the base of the spine, when it is not properly corrected, it may lead to overall imbalance, to compensate for torso tilt, extension of the remaining segments and pelvic retroversion may be the cause of low back pain.
In order to reconstruct L4-S1 SL, we think the following can be done during the operation: 1. the hip joint should be properly extended before operation; 2. the posterior hyperplastic structures should be removed completely, which is bene cial to the shortening of the rear and the opening of the front during compression; 3. a moderate-sized cage should be anterior placed within disc space [22,24,25]; 4. in order to avoid the di culty of compression, the anterior part of disc space can be grafted rst, and then the remaining bone can be implanted after compression; 5. the ideal range of L4 -S1 SL is 2/3 of LL, and LL is related to PI [11,26,27], so individual bending should be performed; 5. L4 and L5 screws toward the lower endplate is more bene cial to the compression and the recovery of segmental lordosis [28].
LDI is the ratio of L4-S1 SL to LL, which is an important parameter of lumbosacral. Janik et al. [29] Considered that the ideal LDI in asymptomatic population should be 2/3. Relevant studies have shown that small LDI in adult spinal deformity may lead to postoperative proximal junctional kyphosis [30]. Zheng et al. [10] found that patients with abnormal LDI after PLIF were more likely to have adjacent segment degeneration (ASD) than those with normal LDI, and patients with low LDI were more likely to have ASD than those with high LDI. In this study, the LDI of the 2 cases of internal xation failure was less than 50%, indicating that abnormal LDI will increase the incidence of mechanical complications. We found that the mean LDI and proportion of LDI within the normal range in the G group were higher than those in group P, and there was a positive correlation between LDI and JOA score improvement rate, indicating that for double level spondylolisthesis, increasing L4-S1 SL appropriately and paying attention to its matching with LL can improve the clinical prognosis of patients to a certain extent.
The shortcomings of this study are: 1. the limitations of single center retrospective analysis, the incidence rate of 2-level isthmic spondylolisthesis is low, and the number of cases is relatively small, some parameters may not get statistical difference.
Therefore, relevant conclusions need to be veri ed by prospective, multi-center, long-term follow-up of large sample of cases;2. taking JOA score improvement rate as the basis for grouping has some subjectivity; 3. there are many reasons that affect the prognosis of patients, we only discuss the radiological parameters as the risk factors of clinical prognosis. However, as far as we know, this is one of the few clinical and sagittal follow-up studies of PLIF in the treatment of 2-level isthmic spondylolisthesis, and it is the rst time to explore the relationship between ΔL4-S1 SL, L4-S1 SL, LDI and the clinical prognosis of L4 and 5 isthmic spondylolisthesis.

Conclusion
PLIF is an effective treatment for L4 and L5 isthmic spondylolisthesis, ΔL4-S1 SL, L4-S1 SL and good LDI may be important sagittal parameters that affect the clinical prognosis.

Declarations
Ethics approval and consent to participate: The study protocol was approved by the Institutional Review Board at the Second Xiangya Hospital, Central South University, Hunan, P.R. China. Written informed consent was obtained from each patient for publication of this study.

Consent for publication:
Not applicable.
Availability of data and material: All data relevant to the study are included in the article.

Competing interests:
Jing-Yu Wang; Fu-Sheng Liu; Xiao-Bin Wang and Jing Li declare that they have no con ict of interest. Authorship: All authors participated in data acquisition. JL and XBW conceived and designed the study. JYW and FSL collected the data, performed the statistical analysis and wrote the manuscript. XBW and JL contributed to revision of the manuscript. All authors read and approved the nal manuscript.  P-value for the comparison between group G and the group P JOA IR: JOA score improvement rate