Comparison of Three Interbody Fusion Methods Through Posterior Approach for Lower Lumbar Spinal Tuberculosis in Adults: a Study of Mid-Long Term Follow-Up

Background: This study was conducted to compare mid-long term outcomes of three interbody fusion methods by way of posterior approach in adults with lower lumbar spinal tuberculosis. Methods: A total of 126 lower lumbar spinal tuberculosis patients were treated by one-stage posterior debridement, interbody fusion, and instrumentation. Of the three types of interbody bone grafts, 41 patients (group A) were treated with autogenous bone graft for interbody fusion , 45 patients (group B) were treated with allogeneic bone grafting, and the remaining 40 (group C) patients were treated with titanium mesh cage. Clinical and radiographic data were gathered and analyzed. Results: At the nal follow-up, all patients were found to be completely cured. Neurological performance and quality of life were remarkably improved compared with those preoperative at the nal follow-up. The preoperative lordosis angles of the three groups were signicantly corrected compared with postoperative immediately values or those evaluated at the nal follow-up. The correction loss of the group C was lower than those of groups A and B. All the patients obtained bone graft fusion; the fusion period of group B was longer than that of the other two groups. No signicant differences among the three groups in adjacent segment degeneration rate were found at the last visit. Conclusions: This mid-long term follow-up study established that one-stage posterior debridement, interbody fusion, and instrumentation can effectively treat lower lumbar spinal tuberculosis. Moreover, intervertebral titanium mesh cage bone graft may provide better outcomes than autogenous or allogeneic bone graft.

was utilized to assess the quality of life. Back pain and radicular pain of lower extremity were estimated with the help of Visual Analogue Scale (VAS).
University of California at Los Angeles (UCLA) grading scale (Table 1) was applied to evaluate the adjacent segment degeneration (ASD) on radiograph.   All patients were administered anti-TB drugs 2 to 4 weeks prior to the surgery, including isoniazid (300 mg/day), rifampicin (450 mg/day), and pyrazinamide (750 mg/day), and ethambutol (750 mg/day). They were strictly advised bed rest, strengthen their nutritional intake, and get anemia and hypoproteinemia corrected, simultaneously. Only when the symptoms of TB subside or disappear surgery may be conducted. During the anti-TB period, surgery may be performed in the presence of acute paralysis or progressive aggravation of neurological impairment, even if the ESR value does not decline.

Surgical method
The surgery was conducted with the patient lying in a prone position under general anesthesia.
Posterior midline incision was made considering the diseased vertebral body to be the center, in group C, exposing bilateral lamina, facet joints, and transverse processes. Pedicle screws were xed in one or two vertebrae adjacent to the affected vertebrae, and short pedicle screws were also installed in the affected vertebrae if the pedicle screw channel was not destroyed by infection. A hemilaminectomy or laminectomy was conducted on the highly damaged side of the lesion segment. Then, the diseased vertebral bodies were exposed by removing the superior and inferior articular processes and pedicle. With the help of curettes of different angles the lesion tissues including the sequestrum, necrotic intervertebral disc, caseous necrosis, and pus was removed, through the transpedicular space, until blood exuded on the bone surface. Thereafter the silicone tube was carefully placed deep into the lesion along the sinus tract, and the pus was absorbed under negative pressure. The procedure was repeated on the other side of the lesion if required. Installation of permanent rods and exerting compression with the help of cantilever bending maneuver under vision to correct the deformity and scoliosis. Both the upper and lower bone surfaces of the vertebral body were repaired as bone graft beds. One or two ideally shaped titanium mesh cages lled with autogenous bone particles (from healthy lamina and spinous process) were used on both sides and allogeneic bone particles in the middle on the basis of the shape and size of the bone graft bed to reconstruct the anterior middle column. Moreover, autogenous and allogeneic granular bones were implanted between bilateral transverse processes, or a suitable allogeneic bone plate was placed between vertebral lamina. Streptomycin powder (1 g) and isoniazid (0.3 g) were applied in the lesion area, and the incision was closed in layers on placing a drainage tube.
Trimmed autogenous iliac and allogeneic bone blocks were implanted in the bone graft bed in groups A and B, respectively. The rest of the surgical procedures were the same as followed in group C.
Mycobacterium culture and histopathological examinations were carried out on the focus tissues of each patient during the operation.

Postoperative management
Routine antibiotics were administered and nutritional support provided post operation. The drainage tube was removed once the drainage volume collected in 24 hours was less than 30 ml. All the patients were continued to be administered with anti-TB drugs chemotherapy regimen post operation mentioned earlier for 12 to 18 months. Routine blood, liver function test, ESR and CRP evaluations were conducted to observe the adverse reactions and assess the e cacy of drugs. Following strict bed rest post operation for 4 weeks, patients were permitted to walk gradually with the help of an external brace for 6 months. Early rehabilitation training and physical therapy should be imparted to all patients to prevent thrombus and improve neurological function. Clinical and radiologic examinations were conducted once in every 3 months during the rst year post operation in all patients and once in every 6 months thereafter.

Evaluating standard and statistical analysis
Operation period, intraoperative bleeding amount, and fusion period for each group of patients were documented. Bone healing was gauged as per the radiologic criteria of Lee et al through CT [10]. SPSS 20.0 software was used for performing statistical analysis. The measurement data of the three groups were compared by way of variance analysis rst, followed by LSD t test to compare each group when the value of P < 0.05. The numeration data were statistically analyzed with chisquare test. P < 0.05 was considered statistically signi cant.

Clinical data
The follow-up periods for the groups A, B, and C were 75.4 ± 11.8 months, 76.5 ± 11.2months, and 76.0 ± 11.5, respectively. All the patients diagnosed with lower lumbar spinal TB were tested to be clinically cured at the nal follow-up.
The operation period and intraoperative blood loss were, respectively, recorded as 189.1 ± 27.2 min and 946.3 ± 185.2 ml in group A, 161.8 ± 24.6 min and 788.9 ± 139.8 ml in group B, 163.3 ± 23.3 min and 777.5 ± 130.6 ml in group C. The results indicated that the values of group A were greater than those of groups B and C (p < 0.05). The ESR and CRP values normalized at 3 months post-surgery.
Patients suffering from preoperative neurological dysfunction exhibited improvement post-surgery in both groups. At the nal follow-up, the JOA, ODI, and VAS values were recorded to be 27.1 ± 1.8, 9.9 ± 1.5, and 0.9 ± 0.8 in group A; 27.3 ± 2.0, 10.0 ± 1.7, and 0.9 ± 0.7 in group B; and 27.3 ± 1.9, 10.2 ± 1.7, and 0.9 ± 0.7 in group C. Statistically signi cant differences were found between preoperative and the nal follow-up values of JOA, ODI, and VAS (p < 0.05). Nevertheless, VAS values one day postoperatively was higher in group A than those of groups B and C. No signi cant differences in the values of JOA, ODI, and VAS were observed among the three groups at the nal follow-up.

Radiographic data
The immediately recorded postoperatively and nal follow-up lordosis angles were 29.8 ± 4.3° and 27.0 ± 3.8° in group A, 29.6 ± 4.2° and 26.8 ± 4.6° in group B, 30.7 ± 6.8° and 30.0 ± 6.6° in group C, respectively. The correction loss values were 2.9 ± 1.0°, 3.1 ± 0.8°, and 0.8 ± 0.7 in groups A, B, and C, respectively. The preoperative lordosis angles of the three groups were evaluated to be remarkably recti ed compared with immediately recorded postoperatively or at the nal follow-up. No signi cant differences were recorded in the correction rates among the three groups. Nevertheless, the correction loss of group C were observed to be lower than those of groups A and B.
The fusion periods of the groups A, B, and C, were 9.7 ± 2.4 months, 24.7 ± 4.2 months, and 9.5 ± 2.7 months, respectively, indicating group B's period to be longer than those of the other two groups (Table 3). During the nal follow-up, 11 patients in group A had degeneration as per UCLA grading scale. The same was observed with 13 and 10 patients in group B and C. There were no statistically signi cant differences in rate of ASD among the three groups (Table 4). Imaging examination conducted at the nal follow-up indicated that all the grafts were fused (Figs. 1-3).

Complications
Cerebrospinal uid leakage was observed to occur in 2 cases in each of the three groups, which were cured after strengthen postoperative rehydration and delaying removal of drainage tube. 9 patients were detected with super cial wound infection and cured by antibiotics (four in group A, three in group B, and two in group C). 3 patients in group A complained of postoperative pain in the bone extraction area, which was overcome by administering nonsteroidal anti-in ammatory drugs. Catheter drainage through minimally invasive incision and regular chemotherapy were procedures through which local abscess recurrence experienced by 2 cases in group B were treated. Pseudarthrosis was experienced by 1 case in group B, for which he underwent anterior titanium mesh cage bone grafting.

Characteristics of lower lumbar spinal TB and its surgical treatment
The lower lumbar spine is at the lowest position of the spine, at which region the body strength is concentrated, speci cally at the junction of the active segment and the xed end. The loading conditions are complex and easily cause force imbalance. The incidence of TB in the lower lumbar spine is insidious, and symptoms of some patients are atypical. Such individuals often suffer lower back pain, which is easily misdiagnosed as lumbar disc herniation, lumbar spinal stenosis, osteoporosis, or other degenerative diseases [11]. As the disease develops, Mycobacterium tuberculosis erodes the vertebral body to cause dead bones and abscesses, leading to instability or deformity of the spine, or invades the spinal canal causing neurological symptoms in lower limbs, even cauda equina syndrome. Therefore, solid bone graft fusion and favorable spinal stability are the prerequisites for curing lower lumbar spinal TB and the key to reduce spinal deformity.
The surgical treatment of spinal TB is speci cally performed to remove the infection focus, relieve nerve compression, and reconstruct the stability of the spine [12,13]. Various surgical approaches were adopted to reduce lower lumbar spinal TB [4,5,14]. Majority of TB lesions involve the anterior and middle column; therefore, anterior debridement and bone grafting were recommended by some surgeons [15,16]. Nevertheless, the long term effect of the anterior approach showed that the bone graft was susceptible to collapse or absorption, and kyphosis was more severe [17,18]. To overcome the shortcomings of this method some surgeons adopted anterior and posterior approaches for xation and fusion, which enhanced the fusion rate of bone graft and the effect of kyphosis correction; however, this surgical procedure signi cantly increased iatrogenic trauma and hospitalization time, especially the elderly, children, and other physically weak patients [5].

Advantages of posterior approach for lower lumbar spinal TB
The posterior-only approach has turned out to be an effective treatment for lower lumbar spinal TB with the advent of posterior spinal instrumentation technology, as reported by several scholars [19][20][21]. Pedicle screws allow the xation of the three columns of the spine, effectively restoring the normal physiological curvature of the spine, thus correcting kyphosis, and a better holding force can reduce the risk of loosening and fracture of the internal xation, and results in a strong biological xation in a short time post operation. Intervertebral and intertransverse or interlaminar bone grafting to achieve 360° fusion can ensure long-term stability of the spine. The rationality of the posterior approach lies in the basic removal of the necrotic tissues and the ossi ed bone from around the lesion that prevents the entry of anti-TB drugs, thus destroying the positive environment essential for the survival of Mycobacterium tuberculosis. The rest of the small amount of lesion and abscesses can be absorbed by long-term, standardized anti-TB chemotherapy post-surgery [22]. As reported by Pang et al [23] the posterior approach was comparatively more effective in correcting kyphosis and less traumatic than the anterior approach. The outcomes of the single posterior and anterior posterior approaches in the treatment of lower lumbar spinal TB were compared by Xu et al [24], and the posterior procedure was found to be a better one with fewer complications.
Three interbody fusion methods for anterior and middle column reconstruction As per the 3-columns theory of Denis [25], the long-term outcome of the posterior approach depends on the reconstruction of the anterior and middle columns through interbody fusion on removal of the TB debris. Autogenous bone was considered as the gold standard for bone grafting materials owing to good biocompatibility and absence of any disease transmission risk [26]. Nevertheless, autogenous bone transplantation cannot satisfy the need for interbody fusion due to complications in the donor site, and prolong the operation time and trauma. Moreover, the number of senile spinal TB patients increased, usually accompanied by osteoporosis and other systemic diseases, such as diabetes and cardiovascular disease, causing low osteoinductive activity and poor osteogenesis of the autogenous bone. The use of allogeneic bone decreases the related complications due to autogenous iliac bone, but lacks osteogenic induction ability. Theoretically, as a dead bone, allogeneic bone carries the risk of hiding Mycobacterium tuberculosis thus causing its recurrence. Furthermore, the bone block lacks blood circulation, thus hindering the supply of adequate anti-TB drug concentration locally. In this study, operation period and blood loss in group A were observed to be greater than those in groups B and C. Two patients of group B suffered local abscess recurrence, who were cured by minimally invasive surgery and regular chemotherapy. One patient in group B experienced pseudoarthrosis at the bone graft site, who was treated by revision surgery.
Of late, several scholars have evidenced titanium mesh cage bone graft carries the potential of reliable spine reconstruction, high fusion rate, effective sagittal imbalance maintenance, and low implant-related complications [27,28]. Depending on the size and shape of the intervertebral bone defect, one or two suitable titanium mesh cage lled with autogenous bone particles from healthy lamina and spinous process were implanted. If the bone mass is inadequate, allogeneic bone particles can also be used for lling the middle of the cage. This intervertebral bone grafting technique has its unique advantages. Initially, the cage has su cient support strength to achieve immediate stability, and is conveniently able to withstand compressive force to prevent it from fracture and displacement. Moreover, implantation of an ideally shaped titanium mesh cage can ensure a relatively large graft volume and bone contact surface between adjacent vertebral bodies, thus promoting graft fusion in an enhanced way. Eventually, intervertebral implanted titanium mesh can be shaped as per the speci c shape of the bone defect, which can retain more healthy bone and prevent complications such as decreased stability of the spine and non-fusion of bone graft as a result of large bone defect. Certain scholars were concerned that the implantation of titanium mesh cage in the lesion area may likely lead to TB recurrence. Nevertheless, it has been demonstrated that TB bacilli have weak adhesion to titanium material and do not in uence the bactericidal effect of anti-TB drugs [29]. In this study, all the patients belonging to group C successfully attained bone fusion, while the fusion period was signi cantly less than that in group B and lower correction loss than those of groups A and B.
Since the study encompasses a mid-long term follow-up of more than 5 years post-operation, it is assumed that ASD may occur over time. An earlier biomechanical study indicated that ASD was associated with loss of motor function in the fused segment, compensatory increase in adjacent segment mobility and mechanical stress, which resulted in augmenting load on the discs and articular processes [30]. Even though interbody fusion is the main cause of ASD, it can also restore the stability of the responsible segment. In this study, the rate of ASD was 26.8% in group A, 28.9% in group B and 25.0% in group C at the nal visit. The ndings were similar to the incidence of ASD (range 21.3-31.9%) after lumbar fusion reported by recent meta-analysis [31].

Conclusions
Generally, this mid-long term follow-up study established that one-stage posterior debridement, interbody fusion, and instrumentation can effectively treat lower lumbar spinal TB. Moreover, intervertebral titanium mesh cage bone graft may result in better outcomes than autogenous or allogeneic bone grafts.

Declarations
Ethics approval and consent to participate This study protocol was approved by the Ethics Committee of Xiangya Hospital and the written informed consent was obtained from all patients. Each author certi es that all investigations were conducted in conformity with ethical principles.

Consent for publication
All patients signed informed consent forms to publish their personal details in this article.

Availability of data and materials
The datasets and materials generated or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This work was supported by the National Natural Science Foundation of China (No.81672191). No bene t in any form has been or will be received from a commercial party related directly or indirectly to the subject of this manuscript.

Authors' contributions
Zhenchao Xu participated in the design of this study, performed statistical analysis, and drafted the manuscript. Zheng Liu collected the clinical data and follow-up details of the study and conducted the study. Xiyang Wang directed the study design and manuscript drafting. All authors read and approved the nal manuscript.