Posterior-only Approach with Unilateral Fenestration for the Treatment of Thoracic and Lumbar Tuberculosis

EXSCINDED Abstract Abstract Background Unilateral fenestration (UF) is a minimally invasive technique used for the treatment of spinal stenosis. But whether it applies to spinal tuberculosis (TB) is still unknown. We present a retrospective study to evaluate the clinical efficacy and feasibility of a posterior-only approach with unilateral fenestration (PAUF) for treatment of the thoracic and lumbar TB (T<B).Methods We performed a retrospective study for 83 patients with the T<B who were treated with PAUF in our hospital from January. 2007 to December 2016. All 83 patients with one or two consecutive functional spinal units involved were selected from 355 spinal TB patients. The operation time, blood loss volume, postoperative complication rate, correction of kyphotic Cobb angle, neurologic function improvement by American Spinal Injury Association (ASIA) classification, the Visual Analogue Scale (VAS) pain score, and bone fusion time were used to assess the safety and clinical efficacy of the PAUF.Results The average age of 83 patients (45male and 38 female) was (17-79) The mean follow-up time was 46.9±13.1 (24-72) months. The operation time, blood loss volume and postoperative complication rate were 156.2±19.9 minutes, 471.8 ±81.4ml and 33.7%. The Mean kyphotic Cobb angle significantly reduced from preoperative 23.0° ±15.3° to postoperative 8.3°±11.0°, and the mean correction of kyphotic Cobb angle was 14.8±5.5°(p<0.001). At the last follow-up, the mean postoperative kyphotic angle was 9.9°±11.4°. According to the ASIA classification, 92.1% (35 out of 38) patients with preoperative neurological impairment had good neurological improvement after the operation. The VAS pain score significantly decreased from preoperative 6.9±1.1 to 1.3±0.7 one year post operation (p<0.001). All the patients achieved solid bony fusion within 13 months after surgery. There were three relapse TB: tuberculosis; UF: Unilateral fenestration; POAUF: posterior-only approach with unilateral fenestration; T<B: thoracic and lumbar TB; anterior-only approach (AOA); posterior-only approach (POA); combined anterior and posterior approach (A&PA); HREZ: isoniazid, rifampin, ethambutol, and pyrazinamide; HRE: isoniazid, rifampin, and ethambutol; ESR: Erythrocyte sedimentation rate; CRP: C-reaction protein; CT: Computed tomography; MRI: Magnetic resonance imaging; ASIA classificationAmerican Spinal Injury Association classification; VAS: Visual Analogue Scale scores of pain

after the second operation and adjustment of anti-TB regimens.Conclusions For thoracic and thoracolumbar spinal TB patients with one or two consecutive functional spinal unit involvement, posterior-only approach with unilateral fenestration is an effective and feasible surgical treatment which can minimize the surgical damage to the posterior column, control infection, corrects the kyphosis, and maintains correction and neurological improvement over time.

Background
Spinal tuberculosis (TB) is the most common extrapulmonary TB, accounting for about 50% of osteoarticular TB [1]. And the thoracic and lumbar segments are the most commonly affected [2]. For spinal TB with severe spinal instability, neurological deficit, and spinal cord compression, surgical treatment is recommended. In recent years, although progress has been made in surgical techniques, surgery for spinal TB remains challenging [3]. Spine surgeons have not reached a consensus on surgical treatment for thoracic and lumbar TB (T&LTB) [4][5][6].
To date, there are three main surgical approaches for the treatment of T&LTB, which are the anterior-only approach (AOA), posterior-only approach (POA), and combined anterior and posterior approach (A&PA) [7]. As the anterior element destruction is the primary destructive form of spinal TB and the AOA can fully expose the lesion, so most surgeons use an AOA for the treatment of T&LTB [1,8]. However, various scholar reported several disadvantages of it, such as the need for long-time postoperative bed rest, unsatisfying correction of the deformity and challenging procedure at lower lumbar levels [9][10][11]. The combined A&PA seems to be a mature surgical method for T&LTB because it not only supplies a direct vision for debridement but also provide a strong internal fixation. But it always leads to massive surgical trauma and long operation time, which are related to severe complications and prolongation of recovery time [12]. Compare with the AOA and combined A&PA, the POA cause less surgical trauma and also achieve good clinical outcomes [7,13,14]. So the POA is widely accepted by more and more spine surgeon.
However, total laminectomy is often required in a POA, which may result in instability of the spine and postoperative back pain [15,16]. Hence, an ideal operation still needs to be explored. Unilateral fenestration (UF) is a minimally invasive operation used for the treatment of spinal stenosis. But there were few reports about whether it applies to surgical treatment of T&LTB. So we performed a study on T&LTB patient with one or two consecutive functional spinal units involved to evaluate its clinical efficacy and feasibility.

General information
This study protocol was approved by the Ethics Committee of our hospital (201507991). The diagnosis of spinal TB was based on clinical manifestations (back pain, lower extremity radiation pain, night sweats, weight loss, and neurological dysfunction), nonspecific laboratory findings such as increased erythrocyte sedimentation rate (ESR), elevation of C-reaction protein (CRP), anemia, and hypoproteinemia, radiologic findings (X-ray films, computed tomography, and magnetic resonance imaging), pathological analysis, and positive bacteria culture.

Preoperative preparation
Every patient routinely received HREZ chemotherapy regimen: isoniazid 5 mg/kg/day, rifampicin 10 mg/kg/day, ethambutol 15 mg/kg/day, and pyrazinamide 35 mg/kg/day for 2-4 weeks before the operation. For patients with severe anemia and hypoproteinemia, intensive nutritional treatment was given until patients' anemia, and hypoproteinemia was significantly improved. When progressive neurological deficits appeared during preoperative chemotherapy, we appropriately shortened the time of drug treatment.

Surgical procedure (Figure 2):
Every patient was placed in the prone position after general anesthesia with tracheal intubation. A median longitudinal incision was made on the skin where the destructive vertebrae located. Conventionally, we separated paraspinal muscles of destroyed vertebral until the vertebral plate and facet joints exposed. Pedicle screws were inserted into one or two adjacent normal vertebrae above and below destructive vertebrae, and Carm X-ray was used to indicate the exact location of pedicle screws. A titanium rod was placed on the non-decompression side for temporary stabilization, and then decompression was performed on the other side where damage rang was bigger and paraspinal abscess occurred. Focus on the lesion, 1/3-2/3 of the inferior part of the upper lamina and 1/3-2/3 of the superior part of the lower lamina should be removed, and then ligament flavum was removed by forceps to expose spinal canal. We expanded the spinal canal appropriately according to the scope of the lesion without involving supraspinal and interspinal ligaments. To expand the vision, the operating table was tilted 30° towards the non-decompression side. We used nerve root retractor to protect the spinal cord and the adjacent upper and lower nerve roots. To prevent the traction of the spinal cord, we cut off the unilateral thoracic nerve root and ligature it if it is necessary. After that, we removed the spinal canal abscess, decompress the spinal cord. Then we opened the destructive intervertebral space, gradually remove the abscess and necrotic tissue, disc, endplate cartilage, sequestrum, and sclerosis bone. We access the paraspinal abscesses via paravertebral sinus and carefully separated abscess wall and sucked out pus and scraped necrotic tissue. Lesion area and sinuses were repeatedly washed with large amounts of saline solution, hydrogen peroxide, and diluted isoniazid solution. After decompression and debridement, anti-TB drug (1.0 g of streptomycin powder and 0.2 g of isoniazid) was put into the lesion area. Allogeneic bone with appropriate length and size was grafted in intervertebral space after mixing with streptomycin powder. The defect area of the vertebral plate resulted from decompression, was repaired by an allogeneic bone plate that was trimmed to fit shape. Then, the rod on the decompression side and cross-linkage (if necessary) were connected to immobilize the grafted bone plate. For patients with kyphosis deformity, the bone graft should be performed after the correction of kyphosis. A drainage tube was placed after all the above operation.

Postoperative treatment
Antibiotic was regularly used during the operation and within three days post-operatively.
After the amount of the fluid drainage reduced to 20ml per 24 hours, we removed the drainage tube. Bracing protection was recommended to last 3months post-operation. All patients received oral HREZ chemotherapy postoperatively. Six months later, pyrazinamide was discontinued. Then, patients received nine to 12-month regimens of HRE chemotherapy (6HREZ/9-12HRE). For some patients with neurological damage, early rehabilitation treatment was allowed to improve neurological function.

Follow-up index
All patients were examined clinically and radiologically at 3, 6, and 12 months after surgery, and then once a year after that. Blood loss and operative time were recorded during the procedure. The following indexes were recorded preoperatively, postoperatively and during follow-up: (1) Cobb angle; (2)  Clinical outcomes (Table 2 and table 3) The Mean kyphotic Cobb angle significantly reduced from preoperative 23.0°±15.3° to postoperative 8.3°±11.0° and the mean correction of kyphotic Cobb angle was 14.8±5.5°( p 0.001). At the last follow-up, the mean postoperative kyphotic angle was 9.9°±11.4°a nd the average loss angle was 1.70±1.07. There were 38 patients with neurological impairments before surgery (four cases in grade B, 13 case in grade C and 21 cases in grade D), and at last follow-up point (at least two years after surgery), 35 patients showed a neurological function improvement (two case improved from B to C, 11 cases improved from C to D, two case improved from C to E, 20 cases improved from D to E). And the rate of neurological improvement was 89.2%. The VAS pain score significantly decreased from preoperative 6.9±1.1 to 1.3±0.7 one year post operation (p 0.001). The average preoperative ESR and CRP was 43.8±16.7 mm/h and 40.5±17.2mg/L, which returned to normal level at the six months after surgery. X-ray and CT were used to assess bone fusion, and all patients achieved definitive fusion within 13 (mean fusion time is 8.1±1.

Complications and adverse events (Table 4)
After the operation, cerebrospinal fluid leakage occurred in two patients, superficial wound infections occurred in three cases, mild intestinal obstruction happened in 11 cases, and deep vein thrombosis of lower limb happened in one patient. During the followup, three patients suffered no neurological function improvement, two patients suffered intercostal neuralgia, and three patients suffered impairment of liver & kidney's function.
All these complications were cured and relieved effectively. There were three cases of recurrent, who were cured after the second debridement and adjustment of chemotherapy regimens.

Surgery is an auxiliary treatment for spinal TB
Spinal TB is a local infectious disease caused by Mycobacterium tuberculosis. According to different pathological status, specific treatment such as rest, motion-suppression, for removal of the lesion. They also achieved good clinical outcomes in their study. But in their operation, the blood loss was both over 600ml, which was significantly more than that in our operation. And at last follow-up time point, the mean loss of correction was 4.3°and 3.2° in their study, which was more than 1.70° in our study. According to these data, our operation seems to cause less surgical trauma and maintain a better correction of kyphosis. In addition, our operation also has these advantages below: 1) minimize the surgical damage to the posterior column; 2) causing small interference to the spinal cord when doing decompression and bone graft; 3) surgery has little impact to organs in the chest and abdominal cavity. Although there were many advantages of POAUF, it also has its limitations. We don't recommend our operation for patients with extensive paraspinal abscess, multiple vertebral involvement, and rigid kyphoscoliosis.  Tables   Table 1 General data of patients  Neurological function by ASIA classification, Wilcoxon signed rank test was used to comparing the grade of neurological function, there was a significant difference between pre-operation and post-operation (P<0.001)   Figure 1 The distribution of involved thoracic and thoracolumbar vertebras