Microscopic far lateral transforaminal lumbar/thoracolumbar interbody fusion for patients with kyphotic Kummell’s disease - a retrospective comparison study with bone cement enhanced screws plus vertebroplasty

Abstract

Background

Various surgical techniques for Kummell’s disease (KD) have been reported, but the optimal surgical treatment remains controversial.

Objective

We proposed microscopic far lateral transforaminal lumbar/Thoracolumbar interbody fusion (FL-TLIF) combined with bone cement enhanced screws (BCES) for the treatment of KD with severe kyphotic deformity. We assessed the safety and efficacy of this procedure and compared the clinical and radiological outcomes of this surgical strategy with those of BCES plus vertebroplasty (VP).

Methods

This is a retrospective controlled clinical study. A total of 55 patients with kyphotic KD in our department from July 2016 to August 2022 were included. Twenty-eight were treated with BCES fixation plus vertebroplasty (VP group) while 27 were treated with microscopic FL-TLIF combined with BCES (FL-TLIF group). The demographic data including age, gender, involved segments and bone mineral density (BMD) were collected preoperatively. The intraoperative blood loss, surgical duration, bone cement leakage and adjacent fractures were recorded. The visual analogue scale (VAS), Oswestry Disability Index (ODI), regional Cobb (RC) angle were measured before and after the surgery and in the follow-up periods. The kyphosis correction degree (KCD), correction loss degree (CLD), VAS and ODI improvement were compared between the two groups.

Results

The mean T-score on dual energy X-ray absorptiometry (DEXA) bone mineral densitometry in the lumbar area is no more than − 2.0, indicating osteopenia or osteoporosis. The minimum follow-up period was 18 months in both groups. The VAS, ODI, and RC angle were significantly improved after surgery in both groups (P < 0.05). The RC angle was maintained at the final follow-up in the FL-TLIF group (P > 0.05), but there was a significant correction loss at the final follow-up in the VP group (P < 0.05). Blood loss and operative time were lower in the VP group compared with the FL-TLIF group (P < 0.05). No significant differences were found between the two groups in terms of VAS and ODI score improvement (P > 0.05). The KCD in the FL-TLIF group was significantly higher than that in the VP group (P < 0.05), while the CLD in the FL-TLIF group was significantly lower than that in the VP group (P < 0.05). Bone cement leakage occurred in 54 screws (20 in FL-TLIF group, 34 in VP group) with no clinical significance. Newly developed vertebral compression fractures adjacent to the level of instrumentation were observed in three patients (10.7%) in the VP group and one case (3.1%) in the FL-TLIF group during the follow-up period. No hardware complications were found in either groups at the last follow-up.

Conclusion

The microscopic FL-TLIF combined with BCES fixation is an effective and safe treatment option for KD with severe kyphosis in terms of clinical and radiological outcomes. Compared with BCES plus VP, the microscopic FL-TLIF strategy provided with more kyphosis correction and possibly better correction maintenance tendency at a minor cost of about 40 min additional operation time and 50 ml blood loss.

Background

Kummell’s disease (KD) was first identified as a common complication of osteoporotic vertebral compression fractures (OVCF) by German surgeon Hermann Kummell in the 19th century [1]. It is mainly caused by minor trauma, present a progressive development from the occurrence of minor fracture to mild vertebral collapse and instability, and finally to severe kyphosis deformity, which would be likely to compress the spinal cord, resulting in neurological and functional deficits [2]. Although the definition of KD is sometimes controversial and some surgeons would like call it delayed union of OVCF, nonunion of OVCF, the pathomechanisms and treatment protocols are similar. Treatment strategies for KD may be nonoperative, with pain medication, bed rest and external fixation bracing [3]. But there are still about 10% of patients who develop fracture nonunion, intractable back pain and severe kyphosis deformity, which indicate surgical treatment.

The aim of surgery is to restore spinal stability and decompress the nerve root, especially for patients with severe kyphosis deformity and neurological deficits [4]. Several surgical procedures have been proposed for KD, including vertebroplasty, posterior decompression and fusion, posterior osteotomy, anterior spondylectomy and reconstruction, but the optimal surgical treatment remains controversial [5]. Doctors struggled to balance the surgical trauma itself and the long-term clinical and radiological results for every individual.

In this study, we proposed a novel surgical strategy for KD with severe kyphosis, which combines microscopic FL-TLIF and BCES fixation. We confirmed the efficacy and safety of this procedure by comparing the clinical and radiological outcomes of this strategy with BCES plus VP.

Materials and methods

This is a retrospective controlled clinical study. A total of 55 patients with OVCF in Yantai Affiliated Hospital of Binzhou Medical University from July 2016 to August 2022, who experienced failed conservative treatments for at least 3 months (diagnosed as KD) with the RC angle of more than 30° on standing lateral radiograph [6] (or more than 20° on fulcrum CT sagittal reconstruction) were included in this study. We revised the dynamic instability from the difference of Cobb angle during dynamic radiography greater than 10°to more than 20°RC angle on fulcurm CT for accurate measurement. Patients with the following conditions were excluded from this research: (1) patients with normal BMD; (2) secondary osteoporosis, such as endocrine system, gastrointestinal, and parathyroid disease; (3) vertebral fracture due to high-energy trauma, and (4) poor general condition made the surgical procedure impossible. (5) Five cases that met the radiological criteria with neurological deficit were excluded to balance the grouped comparison.

The 55 patients were divided into 2 groups based on the surgical method used which depended on surgeon preference and time period when the surgery was conducted. Twenty-eight patients were treated with BCES fixation plus vertebroplasty (VP group) while 27 patients were treated with FL-TLIF combined with BCES (FL-TLIF group). Thirty-seven of the patients had a history of falling and 18 patients suffered from sudden back pain in daily activities. The duration of their symptoms ranged from 3 to 8 months. Demographic data (patient age, sex, damaged segment, BMD, and duration of follow-up) for the 2 groups are presented in Table 1. The blood loss, surgical duration, and adjacent fracture were recorded (Table 1). The visual analogue scale (VAS), and Oswestry Disability Index (ODI) were evaluated before and at the 7th day after surgery, and at the last follow-up. The improvement of VAS and ODI scores were calculated: preoperative VAS/ODI score − VAS/ODI score at 7th day after operation. This study was conducted with the approval of the Institutional Review Board of Yantai Affiliated Hospital of Binzhou Medical University. Written consent was obtained from all patients.

Table 1.

Clinical data of patients

Parameter	VP group	FL-TLIF group	P
Gender
Female	17	15
Male	11	12
Age, average, y The mean T scores	65.71 ± 6.16 -2.74 ± 0.47	66.44 ± 6.44 -2.84 ± 0.55	0.3996 0.6085
Affected levels
T11	5	5
T12	15	12
L1	7	7
L2	1	3
Follow-up (months)	22.68 ± 2.42	22.85 ± 3.64	0.4184
Operation time (minutes)	87.04 ± 10.31	128.59 ± 12.13	0.000
Blood loss (mL)	100.46 ± 12.22	153.89 ± 18.67	0.000
leakage screws/total screws (leakage rate )	34/168 (20.24%)	20/162 (12.35%)	0.052
Adjacent fracture	3	1

Radiological investigations included standing anteroposterior and lateral radiographs, CT scans and MRI were serially done for the evaluation of spinal canal preoperatively and the response to the treatment postoperatively. Preoperative CT scans were performed routinely in the fulcrum position with a pad under the back to primarily assess the stiffness of the kyphosis. Dynamic radiographs were required at the final follow-up to evaluate the fusion status. The sagittal deformity was evaluated by regional Cobb (RC) angle, which is the angulation of the upper endplate of the vertebra above the fracture and the lower endplate of the vertebra below the fracture (Fig. 1). The RC angle was measured before and 7 days after surgery, and the last follow-up. The kyphosis correction degrees (KCD) (preoperative Cobb angle − Cobb angle at the 7th day after surgery) and the correction loss degrees (CLD) (Cobb angle at the last follow-up − Cobb angle at the 7th day after surgery) were calculated based on CT sagittal reconstruction. We defined a “stable fusion” as having < 5° of movement on flexion-extension radiographs at the last follow-up [7].

Fig. 1.

Regional Cobb angle (RC) is the angulation of the upper endplate of the vertebra above the fracture and the lower endplate of the vertebra below the fracture. RC = θ

Surgical procedure

After general anesthesia, the patient was carefully placed in the prone position on an Allan table, which was adjusted to avoid compression of the intra-abdominal vessels and to decrease thoracolumbar kyphosis. A fluoroscopic lateral view of the fracture region was obtained immediately after the positioning to evaluate the RC angle. Then, a standard posterior midline incision and exposure were made and the pedicle screws were placed from one level above the collapsed vertebra to the vertebra below it. Fenestrated screws were used at all vertebrae for cement injection due to the poor bone density. The amount of bone cement in the end vertebrae was controlled to 6–8 ml to ensure the strength of the screws. However, we recommend 0–4 ml in the fractured vertebral body in the FL-TLIF group just to avoid much leakage to the disc space, which may be an obstacle to the interbody fusion. While in the VP group, we recommended more cement in the fractured body to support the anterior column without the concern of the leakage to the disc space.

In the VP group, the screws on the fractured vertebra were directed into the fracture site for later vertebroplasty. After the setting of the cement into the vertebra, the rods were contoured and assembled to maintain the correction of the kyphosis by intraoperative position and manual reduction (Fig. 2). Instrumental reduction was not performed in this group to avoid the possible cantilever effect caused by the over-traction of the anterior column.

Fig. 2.

A 67-year-old female suffered back pain for 8 months: A Pre-op X-ray showed T12 compression fracture with an RC angle of 41.7°. B, C The intravertebral vacuum cleft sign on CT reconstruction indicated Kummell disease with a fulcrum RC angle of 31.1°. D Pre-op MRI showed slight spinal canal involvement without symptoms clinically. E, F BCEs fixation plus VP was performed with acceptable kyphosis correction

In the FL-TLIF group, the trajectory of the screws on the fractured vertebra were directed to the lower endplate to leave room for later FL-TLIF. Most of the time, the high torque of the pedicle screw in the fractured vertebra made the cement injection unnecessary, but the cement enhancement was always necessary in the upper and lower vertebrae. The rod on the right side was then assembled for temporary stabilization before the FL-TLIF.

The FL-TLIF began from the microscopic dissection of the lateral margin of the lamina and the superior articular process (SAP). After removing the lateral part of the lamina and the SAP, the Kambin triangle and the disc annulus fibrosus was exposed [8]. If the window was spacious enough at this time, interbody fusion could be done in the routine process without visualizing the exiting root. When the disc space was narrow, frequently at thoracic levels or in cases with severe deformities, the exiting nerve root had to be exposed and protected after removing the lateral part of the ligamentum flavum (Fig. 3). In most cases with spinal canal involvement, the intraoperative position reduction of the kyphosis was adequate to decompress the canal if there was no neurological deficit preoperatively. For several cases with mild neurological symptoms but no neurological signs in both group, the crescent window of the lamina could be enlarged more medially to visualize the dura without breaking the isthmus and the inferior articular process (Fig. 4). The canal decompression if needed could be realized by removing the ligamentum flavum posteriorly and flattening the vertebral body protrusion anteriorly using a L-shaped depressor. Two segmental FL-TLIFs were needed in cases with collapse of both endplates and/or cases with an RC angle of more than 25° on preoperative CT scans, while only one FL-TLIF was enough in cases with one endplate damage and a less than 25° RC angle (Figs. 5 and 6).

Fig. 3.

The exiting nerve root was exposed after removing the lateral part of the ligamentum flavum (yellow point: exiting nerve root; yellow star: the lamina; white arrow: the disc; SAP: superior articular process; IAP: inferior articular process)

Fig. 4.

For cases with neurological symptoms, the crescent window of the lamina could be enlarged more medially to visualize the dura without breaking the isthmus and the inferior articular process (yellow point: exiting nerve root; yellow triangle: lateral margin of the dural sac; yellow star: the lamina; yellow rectangle: the isthmus; SAP: superior articular process; IAP: inferior articular process)

Fig. 5.

A 57-year-old female with L1 fracture suffered back pain for 6 months: A, B, C Pre-op X-ray and CT showed L1 burst fracture with an RC angle of 23.2° and 23.9° respectively. D Pre-op MRI showed spinal canal involvement with the posterior bony fragment. E Post-op X-ray showed the BCES fixation of the fracture with an RC angle of 10.3°. (F, G) Post-op CT showed one level interbody fusion using autogenous and allogeneic bone grafting with an RC angle of 10.2°

Fig. 6.

A 71-year-old female with L1 fracture suffered back pain for 4.5 months: A, B, C Pre-op X-ray and CT showed L1 collapsed fracture with an RC angle of 37.6° and 25.9° respectively. D Pre-op MRI showed no obvious spinal canal involvement. E Post-op X-ray showed the BCES fixation of the fracture with an RC angle of 14.3°. F, G Post-op CT showed two levels of interbody fusion with an RC angle of 13.2°

After the evacuation of the intervertebral space, the majority of the autogenous bone from the lamina and the SAP was first inserted into the anterior space and then plenty of allogeneic bone was stuffed into the central and posterior space. Interbody cage was not recommended because the disc space was open-wedged anteriorly in Kummell spine, and the posterior entry in FL-TLIF was quite narrow. The pre-contoured left rod was installed, and the kyphosis was re-corrected by manual and instrumental forces. If the correction was not satisfactory after this manoeuvre, the affected facet joints could be released by completely removing the capsule and the cartilage, followed by re-contoured rods correction. There was no need to do Pontes in these group of patients. Autografting with minor lamina bone was performed at the facet joints to promote the posterior fusion. The wound was closed in a traditional way with a drain for two days after surgery.

Statistical analysis

All values are expressed as mean ± SD. One-way ANOVA was used to evaluate the changes in the VAS, ODI, and RC angles based on the data obtained preoperatively, the 7th day postoperatively, and the final follow‑up within the groups. The KCD, CLD, and VAS and ODI improvement were compared between the two groups using an independent sample t-test. All statistical evaluations were performed with SPSS 22.0 (IBM, Armonk, New York). P < 0.05 was considered statistically significant.

Results

The demographic data (patient age, gender, damaged segment, BMD, and duration of follow-up) of the patients are presented in Table 1, and no statistically significant difference was observed between the two groups. The T scores measured by DEXA was − 2.74 in the VP group and − 2.84 in the FL-TLIF group, indicating osteopenia or osteoporosis. The fracture sites were all in the thoracolumbar region with a majority of lesions at T12 and L1. Two patients underwent two segmental FL-TLIFs while 25 cases had only one segmental fusion in the FL-TLIF group. The mean intra-operative blood loss was 100.46 ml in the VP group and 153.89 ml in the FL-TLIF group (P < 0.05). The mean duration of surgery was significantly lower in the VP group compared with the FL-TLIF group (87.04 min vs. 128.59 min, P < 0.05). The follow-up period ranged from 18 to 26 months (mean, 22.68 ± 2.42 months) in the VP group and from 18 to 28 months (mean, 22.85 ± 3.64 months) in the FL-TLIF group (Table 1).

All patients had significant improvement in back pain after surgery in both groups. Intra-group comparisons showed that the mean VAS, ODI score, and the RC angle were significantly improved seven days after surgery compared to the preoperative state (P < 0.05) in both groups (Tables 2 and 3). The improvement of RC angle was maintained at the last follow-up in the FL-TLIF group, but there was a significant correction loss at the final follow-up in the VP group (P < 0.05). There was no significant difference in VAS and ODI improvement between the two groups at 1 week after operation (P > 0.05). The KCD in the FL-TLIF group was significantly higher than that in the VP group at 1 week after operation (P < 0.05), and the CLD in the FL-TLIF group was significantly lower than that in the VP group (P < 0.05, Table 4). Early ambulation was encouraged the next day after the surgery for all patients. All the cases obtained solid fusion according to the dynamic radiographic images at the last follow-up. Bone cement leakage occurred in 54 screws (20 in FL-TLIF group, 34 in VP group) without clinical significance. Newly developed vertebral compression fractures adjacent to the level of instrumentation were observed in three patients (10.7%) in VP group and 1 case (3.1%) in the FL-TLIF group during the follow-up period, which were all treated by PKP with good satisfactory results. There were no wounds and hardware complications. No neurological deterioration after surgery occurred in the 2 groups.

Table 2.

VAS、ODI and RC angle (‾x ± s ) in FL-TLIF group

	VAS	ODI	X Cobb angle(°)	CT Cobb angle(°)
Pre-op	7.44 ± 0.80	66.85 ± 6.49	33.04 ± 2.65	24.10 ± 2.02
7th Post-op	2.70 ± 0.61a	24.70 ± 1.56 a	14.31 ± 2.52 a	11.82 ± 0.74 a
FFU	2.48 ± 0.51a, b	24.48 ± 1.85a, b	14.58 ± 2.51a, b	12.22 ± 0.92a, b
F P	500.996 0.000	1006.351 0.000	475.007 0.000	719.192 0.000

Pre-op pre-operative, Post-op post-operative, FFU final follow-up,

^aP<0.05 vs. preoperative values, ^bP>0.05 vs. 7th post-operative values

Table 3.

VAS、ODI and RC angle (‾x ± s ) in VP group

	VAS	ODI	X Cobb angle(°)	CT Cobb angle(°)
Pre-op	7.79 ± 0.69	66.64 ± 6.73	30.71 ± 2.19	22.76 ± 1.72
7th Post-op	2.82 ± 0.67a	24.75 ± 2.25 a	23.61 ± 2.05 a	15.23 ± 1.23 a
FFU	2.71 ± 0.60a, b	24.46 ± 1.91a, b	25.8 ± 2.67a, c	17.48 ± 1.42a, c
F P	551.364 0.000	915.851 0.000	68.968 0.000	193.203 0.000

Pre-op pre-operative, Post-op post-operative, FFU final follow-up,

^aP<0.05 vs. preoperative values, ^bP>0.05 vs. 7th post-operative values, ^cP<0.05 vs. 7th post-operative values

Table 4.

Improvement of VAS、ODI, Correction and loss degree of Cobb angle

	ΔVAS	ΔODI	KCD	CLD
FL-TLIF	4.96 ± 0.76	42.37 ± 6.50	12.27 ± 1.86	0.40 ± 0.89
VP	5.07 ± 0.90	42.18 ± 7.03	7.53 ± 1.17	2.25 ± 1.82
T P	-0.482 0.631	0.105 0.917	11.284 0.000	-4.823 0.000

KCD kyphosis correction degree, CLD correction loss degree

Discussion

The surgical indications for KD in this clinical study

The development of KD involves the following 3 stages: Stage I, the vertebral body is slightly injured by minor trauma, and the loss of vertebral height is less than 20%. Stage II, dynamic instability of vertebral bodies leads to fracture, and the loss of vertebral height is more than 20%, but the posterior wall of the vertebral body remains intact; and Stage III, the posterior wall collapses and leads to continuous severe back pain, with or without neurological deficits [9]. For stage III KD, PKP or PVP is usually hard to meet the goals of the treatment including the segmental stability, and the sagittal balance of the spine, especially for the patients with obstinate kyphosis. Various open surgeries were recommended for these situations in recent years. Zhang recommended short segmental pedicle fixation combined with vertebroplasty for KD with neurologic symptoms [10]. Huang concluded that bone cement augmented short-segment fixation with posterior decompression was beneficial in elderly patients with stage III KD [11]. Okuda applied posterior subtraction osteotomy in patients with the severe kyphosis [12]. Kostuik reported anterior instrumentation and decompression for patients with the spinal cord compression originated from the ventral aspect [13]. The surgical indications in the study are KD patients with kyphotic deformity more than 30° (or more than 20° on fulcrum CT), suffering from chronic back pain, with or without neurological deficits.

Although some studies showed excellent results by percutaneous screw fixation with a VP at the fracture level [4]. We did open BCES fixation plus VP for comparison instead of percutaneous screws due to the following considerations. In our opinion, the inevitable loosening of the percutaneous pedicle screws in the osteoporotic bone helps little to the VP in the long term, and the good results reported might come mainly from the VP itself. In addition, since the goal of the pedicle fixation should be the posterior joint fusion in the long run, the untreated facet joints need a longer time to fuse spontaneously compared with open surgery. Moreover, the release of the posterior elements in the VP group contributed a little to the intra-operative correction, which simultaneously balanced the comparison with the TLIF group. The benefits mentioned above to a kyphotic spine are worth the minor iatrogenic trauma of the posterior exposure.

BCES fixation

Patients with advanced KD are often elderly people and have severe osteoporotic spines. For osteoporotic vertebrae, pedicle screw loosening is a common postoperative complication. A variety of internal fixation methods such as the use of expansion screws and increased segmental fixation have been proposed to reduce the risk of hardware failure [14]. BCES fixation has been paid more attention by scholars in recent years since biomechanical tests have revealed that BCES can improve the pull-out resistance of the screw by 147–278% [15]. All the patients in our study used fenestrated screws for the injection of bone cement into the vertebral body, which end up with no screw loosening at the last follow-up. We also recommended the use of BCES to ensure the intraoperative instrument correction of kyphosis. In addition, the strong anchorage of the screw into the vertebral body ensured the anterior cantilever effects at the early stage after the fixation until the interbody callus comes into support at later time. On the other hand, the cantilever effects by the strong fixation in return favors the anterior fusion in the early period. Thus, the use of BCES is essential for the success of this proposed surgical strategy in terms of early cantilever effect and later anterior fusion support.

Bone cement leakage is a common concern when BCESs are used. Jassen reported 66.7% of bone cement leakage without neurological symptoms, and 7.9% of pulmonary embolism in the patients with fenestrated screws [16]. In our study, 20 screws in FL-TLIF group and 34 screws in VP group showed bone cement leakage beside the vertebral body or along the veins anteriorly, and a few into the disc space, with a leakage rate of 12.35% and 20.24% respectively. It was numerically higher in VP group but not statistically significant. All the leakages in our group had no obvious clinical symptoms probably because our leakages usually happened at the a later stage of the injection and the leakages were meticulously monitored when they began to emerge. The leakage difference rate between the two groups was not significant statistically or clinically. The different recommended cement volume to the fractured body in the two groups may be the reason of the difference of leakage rate.

Kyphosis correction

Since the residual kyphosis could lead to back pain, adjacent OVCF, or even sagittal imbalance, the correction of high-degree kyphosis is one aim of the proposed surgery. In order to correct the deformity, some scholars recommended pedicle subtraction osteotomy (PSO), which can provide 25° to 35° correction, but rarely beyond 40° [17, 18]. Further developments of PSO such as the bone–disc–bone closing‑wedge osteotomies (BDBO) increase the correction angle of 38–49° [19]. Yang recommended modified posterior vertebral column resection (PVCR) for the treatment of KD with kyphotic deformity and the average kyphotic angle was improved by about 30° [20]. However, major posterior shortening after PSO might cause central stenosis and nerve root twisting with entrapment [21]. In addition, the PSO, BDBO and VCR involve extensive dissection of the vertebra, in which much blood loss is unfavorable for elderly patients. Moreover, since the posterior elements are damaged in these osteotomies, long segmental fixation is usually needed to ensure the deformity correction and posterior or anterior fusion.

In this clinical study, we found that in most cases the intraoperative position reduction was able to correct a big part of the kyphosis, and so we did BECS fixation plus vertebroplasty in the earlier study period or for patients with bad general condition in the later study period. However, the degree of correction by position and instruments is sometimes limited to achieve a satisfactory correction (< 15°). Then the FL-TLIF was proposed to address these issues. The intraoperative position reduction combined with the intervertebral space release were able to correct most of the kyphosis. If the correction was not satisfactory after the management of the disc, the affected facet joints could be released by removal of the capsule and the cartilage of the joint followed by re-contoured rods kyphosis correction. After all these manipulations, we could achieve 15–25°correction totally, of which one level release contributed 6–10° on average. Since most of the KD in clinic are not stiff, they could be treated by this technique with an acceptable kyphosis of less than 15°.

Adjacent fractures after instrumentation had been reported as a postoperative problem correlated with the surgery. Theoretically, the fractures are not only caused by osteoporosis and injury, but might be due to the mechanical changes after the surgery including the increased shearing force by the instrumentation and the left kyphosis. The rates of adjacent fractures varied from 2.3% to 79% in literatures in different patients with different surgical techniques [4, 12, 22, 23]. In this study, there were three adjacent fractures in the VP group but only one in the FL-TLIF group during the average 18 months follow-up. Intuitively, it seemed to be a trend that the more kyphosis left, the more fractures would happen. However, the small numbers of the adjacent fractures made statistical analysis impossible. We need a larger patient population to study the correlation between adjacent fractures and the surgeries in order to exclude the multifactorial interferences.

FL-TLIF

We never perform a conventional TLIF for an OVCF because Conventional TLIF is more invasive compared with FL-TLIF+BCES fixation and BCES fixation + VP mentioned in this paper considering the removal of facet joints and interference with the spinal canal. The intact facet joints are of great importance for the stability of the fixation which we call it tension band fixation. The tension band fixation means the body weight exert a load on the screws and the facet joints but only a little tension on the rod. The posterior tension band mechanism combined with the anterior BCES cantilever mechanism makes the hardware failure less likely in an old patient who need a longer time for interbody fusion or spontaneous facets fusion. We use BCES fixation + VP as the control group because this technique also uses the posterior tension band mechanism and the anterior BCES cantilever mechanism, the only difference is the expectation of longer spontaneous facet fusion rather than shorter interbody fusion in the FL-TLIF group. In addition, the spinal canal was not much disturbed in FL-TLIF because it goes into the disc from the lateral margin of the lamina and SAP. Even in cases with protruded bony fragments, the exploration and decompression could be performed from the lateral side of the canal without damaging the continuity of the facet joint.

Posterior pedicle screw fixation combined with vertebroplasty (PV) was used for KD for its less invasive property, especially for patients who are not able to tolerate major surgery with massive blood loss such as PSO, BDBO or VCR [24, 25]. However, the lack of anterior reconstruction in this procedure raises concerns about the late-stage hardware failure and kyphosis correction loss. In this study, the CLD is significantly larger in the VP group than that in the FL-TLIF group, which might indicate that the anterior bony fusion plays a role for correction maintenance in the long run. However, the CLD is only 2.25 degrees for the VP group versus 0.4 for the TLIF group, the statistical difference may be of little clinical relevance. Although BDBO or VCR provided with strong anterior support, they are technically challenging and are risky for old patients. Moreover, the removal of the posterior elements makes it necessary the long segmental exposure, which additionally increases the risk of the procedure with more bleeding. FL-TLIF avoids the concern about late-stage hardware failure by interbody fusion support in about three months, which is achieved from the side of the lamina with less blood loss. Although the possible risks of FL-TLIF are the bleeding from the small blood vessels around the foramen and the injury to the exiting nerve, the procedure could be performed clearly and easily with the help of a microscope.

From this study, we could conclude that both surgical strategies (BCES fixation plus VP, BCES fixation plus FL-TLIF) were effective for KD patients with kyphosis in terms of VAS and ODI improvement. FL-TLIF demonstrated superior radiographic correction with comparable clinical outcomes. We recommend microscopic FL-TLIF for the treatment of KD with severe kyphosis because of the better kyphosis correction (KCD) and possible long-term correction maintenance (CLD) tendency compared with those of the VP group.

However, this study has some limitations. First of all, the current study is a small sample retrospective controlled analysis in a single institution. We need a prospective study in a large number of patients to more powerfully identify the effectiveness of this technique in the future. Second, although the learning curve of FL-TLIF is not very steep, it really needs some observation and experience to perform this microscopic surgery successfully. Third, the assessment of bony fusion is highly complex and controversial. We only used the simplest dynamic spinal radiographs to define the fusion. For greater accuracy, CT fusion grade scale, independent radiologist, interobserver reliability should be considered.

Conclusions

Favorable outcomes were obtained in KD patients with severe kyphosis treated by BCES plus VP and BCES plus microscopic FL-TLIF, and the latter is a better solution for this group of patients in terms of more kyphosis correction and possible long-term correction maintenance tendency at a minor cost of about 40 min, additional operation time and 50 ml blood loss.

Authors’ contributions

JW T conceived and designed the study. XJ T and TS X wrote the manuscript. XW D performed the data analysis. CX W and QY S contributed to the manuscript revise. XJ T, JW T and CX W contributed to literature search and data extraction. All authors have read and approved the final version of the manuscript.

Funding

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

The data used to support the findings of this study are available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

Ethical approval was obtained from the ethics committee(Yantai Affiliated Hospital of Binzhou Medical University ethical review document of Ethics Committee; Serial number: F-KY-0022-20220325-01). Patients provided informed consent before the treatment.

Consent for publication

The Author confirms:

·that the work described has not been published before;

·that it is not under consideration for publication elsewhere;

·that its publication has been approved by all co-authors,

Competing interests

The authors declare no competing interests.