Treatment of Severe Bone Destruction in L5-S1 Spinal Tuberculosis With Anterior Combined Posterior Approach

Background: There are many studies on the surgical treatment of lumbosacral tuberculosis, but both the anterior and posterior approaches present some shortcomings. This study aimed to evaluate the therapeutic ecacy of anterior debridement and bone graft, posterior xation and fusion with navigation for L5-S1 tuberculosis with severe bone destruction. Methods: This was a retrospective study of 24 patients with severe tuberculosis in L5-S1 who underwent anterior interbody arthrodesis and posterior pedicle screw internal xation by open approach under computer navigation between February 2011 and Novenmber 2016. The erythrocyte sedimentation rate(ESR), C-reactive protein level(CRP),visual analogue scale(VAS), and lumbosacral angle were compared between before surgery, after surgery, and at the nal follow-up. The fusion status of bone graft was evaluated with computed tomography(CT). Results: The mean operation time was 244.58 minutes. The mean intraoperative blood loss was 413.75ml. The accuracy of screw placement was 96.43%. The mean follow-up period was 26.17 months. The average ESR, CRP, VAS, and lumbosacral angles were 65.96 mm/h, 52.93 mg/L, 4.96 points ,and 107.94°, respectively,at preoperative, 34.17 mm/h, 16.47 mg/L, 1.58 points, and 116.12°, respectively, after surgery, and 14.08 mm/h, 6.20 mg/L, 0.58 points, and 115.97°, respectively, at the nal follow-up period. The differences of ESR, CRP, VAS are statistically signicant (p < 0.05). The difference of lumbosacral angles before and after surgery is statistically signicant (p < 0.05)(cid:0)but there is no statistically signicant difference between after surgery and the nal follow-up period (p(cid:0)0.05). Nine patients with ASIA Grade D before surgery returned to Grade E by the nal follow-up period. All patients achieved bone fusion.There was no recurrence of the disease. Conclusions: Anterior debridement and bone graft fusion combined with navigated posterior pedicle screw xation is a safe and effective treatment option for patients with severe bone destruction in L5-S1 spinal tuberculosis.


Introduction
Spinal tuberculosis is the most common form of extrapulmonary tuberculosis, accounting for 50-60% of all bone and joint tuberculosis. 1 Lumbosacral tuberculosis is extremely rare, accounting for only 2-3% of all cases of spinal tuberculosis reported in the literature. [1,2] Similar to treatment of other forms of spinal tuberculosis, the treatment of L5-S1 spinal tuberculosis entails anti-tuberculosis drug therapy or combination surgery. However, for L5-S1 vertebral tuberculosis, the region has a complex anatomical structure and special biomechanical characteristics. Treatment is particularly di cult when severe bone destruction affects the stability of the lumbosacral segment, and both the anterior and posterior approaches present some shortcomings. Surgery via the anterior approach alone for the correction of kyphosis has little effect, barely maintains spine stability, and often leads to bone graft detachment. Surgery via the posterior approach alone for the complete debridement of lesions, such as anterior abscesses and damaged intervertebral discs, is di cult, and performing structural bone grafting to reestablish the stability of the anterior vertebral column is also di cult. [3,4] This study aimed to investigate the e cacy of debridement and bone graft fusion through a one-stage paramedian retroperitoneal approach combined with navigated posterior-approach pedicle screw xation for L5-S1 tuberculosis with severe bone destruction.

Study design and patients
At the time of admission, all participants provided written informed consent for their data stored in our hospital database and used for study purposes .This study was approved by our Ethical Committee. A retrospective review of 24 patients with severe bone destruction associated with L5-S1 spinal tuberculosis who underwent lumbosacral spine surgery was performed at a single university-based spine clinic from February 2011 to November 2016. There were 13 male and 11 female patients, with ages ranging from 25 to 70 years (average 43.75 ± 11.78 years). The patients had varying degrees of low back pain. Nine patients had American Spinal Injury Association (ASIA) Grade D nerve compression, and six had a history of tuberculosis at other sites, including four patients with pulmonary tuberculosis, one patient with renal tuberculosis, and one patient with a tuberculocele. All patients underwent X-ray, computed tomography (CT), and magnetic resonance imaging(MRI).
They presented with paravertebral abscesses, sequestra and kyphosis of various sizes, as well as signi cant vertebral destruction and intervertebral space collapse. The surgery procedure was performed by the same team.
All patients were pathologically diagnosed with lumbosacral tuberculosis.
(2) presence of necrotic bone and paravertebral abscess or segmental instability.

Preoperative preparation
Anti-TB oral drugs with the HREZ standard chemotherapy regimen that consists of isoniazid 0.3 g/d, rifampicin 0.45 g/d, ethambutol 0.75 g/d, and pyrazinamide 0.75 g BID were administered for over 3 weeks preoperatively, combined with nutritional supplementation in the form of a high-protein diet. After admission, 750,000 IU streptomycin IM QD was given. The course of treatment is two months. Active tuberculosis was excluded by preoperative routine chest X-ray examination and multiple acid-fast staining sputum examinations. After the general condition of the patient improved and tuberculosis symptoms resolved, surgery was scheduled

Surgical management
The patient was placed in the supine position. The left hip was slightly lifted. An incision was made through a left paramedian approach to reach operative region. After incision, paraspinal pus, granulation tissue, sequestra, and residual intervertebral discs were removed under direct vision and were sent for pathological examination and tuberculosis culture. The height of the residual intervertebral space was measured. The surgical instruments and gloves were replaced, and a tricortical iliac bone graft of appropriate size was harvested from the left anterior superior iliac spine and implanted into the L5-S1 intervertebral space without additional xation( Fig. 1A-B). The length of the selected iliac bone graft would be slightly longer than the measured height of the L5-S1 intervertebral space to reduce the possibility of detachment after bone grafting. Intraoperative C-arm uoroscopy con rmed proper bone graft placement. The patient was switched to the prone position. After incision, the pedicle screw was placed under computer navigation to improve the accuracy of the screw placement. Pedicle screws were inserted in 1 segment above and below the involved body if the pedicle screw channel of the involved body was destructed and the screw hold in the body was not strong enough. Pre-bent connecting rods were installed and appropriately extended longitudinally to correct kyphosis. The posterolateral bone grafting were performed with autologous iliac cancellous bone after full decortication of the transverse process and the lamina at the xation segments.

Postoperative management
The drainage tube was removed when the drainage volume was less than 50mL/24h. The patients were encouraged to walk with a lumbosacral brace to reduce the occurrence of complications related to immobilization. The brace was worn for 12 weeks after surgery. The patients continued the oral HREZ after surgery. Six months later, the oral pyrazinamide was stopped. The patients continued 1 year regimens of HRE chemotherapy. For patients with complicated neurological dysfunction, mecobalamin 0.5 mg po TID was given.
ESR and CRP levels, liver and kidney functions, and biochemical examinations were re-examined monthly for 3 months after surgery and every 3 months thereafter. X-ray and CT examinations of the lumbosacral segment were performed 3 months, 6 months, and 12 months after surgery and annually thereafter.

Observational indices and assessment criteria
Operation time, intraoperative blood loss, and surgery-related complications were recorded. The incidence of poor postoperative screw position detected by CT after surgery was calculated. Changes in ESR,CRP,VAS and lumbosacral angle measured by the Dubousset method[6](Because the inferior end plate has been destroyed in these patients, the lumbosacral angle measures the degree of kyphosis between the superior end plate of L5 and the posterior wall of the sacrum)were determined before surgery, after surgery, and at the nal follow-up. The difference in ASIA grades before surgery and at the nal follow-up period as well as tuberculosis recovery and bone graft fusion status at the nal follow-up period were recorded.

Statistical methods
Data were processed using SPSS 18.0 statistical software. All data were expressed as mean ± standard deviation (x̄ ± s).One-way MANOVA was used to compare the differences among the data collected before surgery, after surgery, and at the nal follow-up period. The SNK test was used for pairwise comparisons between groups. The signi cance level was set at α=0.05.
The operation times were 170-420 min (average 244.58 ± 59.53 min). Intraoperative blood losses were 100-1000 ml (average 413.75 ± 236.40 ml). Among the 112 screws, only four slightly broke through the medial wall of the pedicle, and the accuracy of screw placement was 96.43%. There were no vascular injuries, ureteral injuries, or new nerve injury complications. There was no retrograde ejaculation in male patients. All patients underwent realtime uorescence quantitative PCR pathological examination, all of which supported the diagnosis of tuberculosis. The results are summarized in Tables 1 and 2.

Discussion
Tuberculosis is an internal medicine disease, and its treatment is primarily based on anti-tuberculosis drugs that can achieve good therapeutic effects. [7,8] A long-term follow-up study showed that although conservative treatment achieved good outcomes, residual low back pain and kyphosis were more common in late stages. [9] Currently, it is generally believed that patients should be treated with surgical intervention as long as there is progressive aggravation of neurological dysfunction, massive paraspinal abscess and sequestrum formation, severe pain not signi cantly relieved with medication, or signi cant bone destruction that affects spinal stability. [10,11] The purpose of the operation is complete lesion removal, spinal canal decompression, correction of deformity, and re-establishment of spine stability through bone grafting and stable internal xation. After the operation, patients can quickly return to normal life as a consequence of rehabilitation exercise. [4,12] However, the methods to achieve these goals are controversial. [3,4] Since Hodgson rst reported the treatment of spinal tuberculosis using the "Hong Kong operation" in 1960, [13] tuberculosis debridement and internal xation through an anterior approach has become the preferred surgical approach for spinal tuberculosis. The advantages of the anterior-approach surgery are as follows: (1) direct access to the lesion for complete debridement and direct spinal canal decompression. [4,14] (2) anterior vertebral column stability can be reestablished immediately through intervertebral bone grafting, and the compressive stress is conducive to postoperative bone fusion. [15](3)it can appropriately improve kyphosis. Nevertheless, the disadvantages of anterior-approach surgery are that the bone graft bears high compressive stress after anteriorapproach lesion debridement alone, and the bone graft is prone to collapse and absorption, causing insu cient support for the lumbosacral segment. Furthermore, the lumbosacral segment is an in ection point of biomechanical change where the local stresses are large. Without internal xation, long-term bed rest after surgery is required, and the bone graft can be displaced or even become detached due to excessive stretching or torsion of the waist.
[16] Selection of anterior-approach internal xation therefore requires greater exposure, increasing the risks of surgery, and has greater requirements on the residual vertebral bone. It has been reported that the complication rate increased by 50% as the scope of surgical exposure was increased and internal xation was placed on normal vertebral bodies. [17] Weinstein pointed out that the pedicle itself provides 60% of the total pedicle screw anti-pullout strength; cancellous bone accounts for approximately 15-20%, and the anterior cortical bone accounts for approximately 20-25%.
[18] Sun et al. reported that the average improvement in lumbosacral angle after posterior-approach surgery was 44.3°. [19] whereas the average improvement in lumbosacral angle after anterior-approach surgery was only approximately 9.5°. [3] Therefore, the anti-pullout strength, the ability to correct kyphosis, and the ability to maintain corrective effect with posterior-approach xation is better than with anterior-approach xation. Nevertheless, posterior-approach tuberculous lesion debridement alone is often incomplete and is prone to recurrences after surgery, especially in patients with signi cant bone destruction of the anterior vertebral column and large abscesses in the psoas major muscle or iliac fossa. [20] Moreover, the stability of the anterior vertebral column is not easily reestablished by structural bone grafting via posterior-approach surgery alone; this leads to loss of lumbosacral height and the effect of kyphosis correction. [15,21]Therefore, the rst choice for L5-S1 spinal tuberculosis with severe bone destruction is a combined anterior-posterior approach. [4,11,21] With respect to the order, Zeng et al. believe that because performing posterior-approach xation leads to stiffness of the xed segment, it will be di cult for anterior bone graft to support the anterior vertebral column after anterior-approach surgery. [2] Therefore, performing anterior-approach surgery rst is recommended for patients with severe bone destruction in L5-S1 spinal tuberculosis. For anterior-approach bone grafting, the size of the iliac bone graft should be appropriately larger than the measured height in order to be conducive to reducing dislocation of the bone grafting interface under compressive stress and to promoting bone graft fusion. [15] For posterior-approach bone grafting with autologous iliac cancellous bone, the cartilage surface of the facet joints in the lumbosacral region should be removed, including the L5-S1 spinal facet joint articular surface, and full decortication of the transverse process and the residual lamina can be performed to improve postoperative bone graft fusion and maintain the effect of kyphosis correction.
In order to improve postoperative bone graft fusion rate and reduce surgical complications, it is necessary to preserve as many of the residual vertebrae as possible and to place screws on the residual vertebrae to reduce the xed segments. We achieved good surgical outcomes in 24 patients with severe bone destruction in L5-S1 spinal tuberculosis. Several limitations should be considered while interpreting the results of this study. This was a retrospective analysis of patients from a single-center, which limits generalizability of study ndings. The small sample size of our study (n=24) decreases the power of the results. Some prospective randomized control trial studies with larger sample sizes are needed to validate our results.

Conclusion
Anterior debridement and bone graft fusion combined with navigated posterior pedicle screw xation allows not only complete debridement of necrotic tissue in the anterior and intervertebral space of the lumbosacral segment under direct vision and reconstruct anterior column with autologous tricortical iliac bone, but also obtain postoperative stability and facilitate bone graft fusion, which is a safe and effective treatment option for the treatment of severe bone destruction in L5-S1 vertebral tuberculosis.

Declarations
Ethics approval and consent to participate