A 270° Retention of Spinal Structure in the Treatment of Thoracic and Lumbar Tuberculosis via Posterior-Only Approach

From January 2013 to September 2014, we enrolled 21 patients with active spinal tuberculosis of the thoracic or lumbar spine with indications for surgery: (1) progressive neurological deficit; (2) persistent pain caused by instability; (3) severe kyphosis or kyphosis likely to progress; and (4) poor outcomes after conservative chemotherapy. Permission was obtained from the hospital ethics committee before starting this study, and informed consent was obtained from all patients or their legal guardians.

Medical records of clinical and operative reports, radiographic images, and pathology reports were extracted. The diagnosis of spinal tuberculosis is determined by nonspecific laboratory findings including blood routine examination, blood chemistry profile, and erythrocyte sedimentation rate (ESR) and by imaging findings such as spine X-ray, computed tomography (CT), and magnetic resonance imaging (MRI).The Cobb angle is measured on spinal X-ray. The involvement of contiguous vertebral end plates, the destruction of vertebrae, and the location of sequestrums were identified on CT. Abscess, spinal cord compression, and destruction of intervertebral discs are observed by MRI. In addition, neurologic status is also evaluated according to the Frankel grade scores.

The patients received antituberculosis drugs after the preliminary diagnosis for at least 3 weeks, together with strict bed rest and measures to improve general condition and nutrition. Antituberculosis drugs included ethambutol (EMB) 750 mg/d, isoniazid (INH) 300 mg/d, pyrazinamide (PZA) 1500 mg/d, rifampicin (RFP) 450 mg/d, and levofloxacin 300 mg/d. Surgery was performed under the cover of chemotherapy. Generally, surgery was performed when physical condition improved obviously.

Given general endotracheal anesthesia in prone position, we made the posterior median longitudinal incision and continued to incise until spinous process and vertebral plate were exposed clearly. In step 1, we set the internal fixation screws through the vertebral pedicle as planned and corrected the kyphosis of certain degree by a rod at the same time. In step 2, we identified the vertebrae according to imaging findings and chose the severely involved side of vertebrae, with the opposite side filled in with gauze. We stripped the unilateral vertebral plate and zygapophyseal joints by rongeur in the chosen side, during which we freed and protected the upper and lower lumbar nerve roots (Fig. 1a and 1c), but sometimes sacrificed the thoracic nerves (Fig. 1b and 1d).Then we entered the targeted space, drained the abscess with negative pressure, and curetted the central lesion (Fig. 1e), which was the involved vertebral body and the intervertebral disc in most cases, with satellite lesions retained. The kind of lesion resection was considered focal debridement instead of radical debridement. In step 3, we resected the appropriate autogenous bone from posterior superior iliac spine (Fig. 1f). After careful disposal, we implanted the bone graft into the debrided space. As the above steps were completed, 270° of the spinal structure remained (Fig. 2). Treatment with streptomycin (1.0 g) was locally administered. A negative pressure drainage tube was routinely fixed on the operated side. The debrided tissue and bone were kept for culture and histopathologic examination.

Antibiotics were administrated during the first week after the operation. The negative pressure drainage tube was pulled out when drainage flow was less than 30 mL/d. The stitches were normally removed 14 days after the operation. Patients were instructed to wear a brace for at least 6 months. The patients were given the antituberculosis drugs for 12 to 18 months postoperatively. ESR and liver and kidney function were examined once a month postoperatively to monitor any adverse reactions of the antituberculosis drugs, and all patients were followed up on an outpatient basis at 3-month intervals for the first 6 months and then 6-month intervals for the next 2 years, and annually for life thereafter. X-rays were performed once a month, and MRIs were taken every 3 months.

Mean and standard deviation were used to describe normal distribution data. Median and quartile percentile were used to describe skewed distribution data. Paired t test and Wilcoxon signed-rank test were used to analyze the comparison between 2 different times. Two-way analysis of variance (ANOVA) was used to analyze the comparison among 3 different times, and the LSD t test was used as the post hoc test. P ≤ 0.05 was considered statistically significant. All statistical calculations were performed by PASW Statistics, version 21.0.

Twenty-one consecutive patients with thoracic or lumbar tuberculosis were enrolled in this study, and the characteristics of the patients are described in Table 1. The population comprised 13 males and 8 females, with a mean age of 47.9 (range, 24–73) years. All patients presented with constitutional symptoms such as fatigue, weakness, night sweats, lower fever with weight loss, or other symptoms of different degrees. The most common symptom was dorsal spine pain (90.5%), followed by neurologic complaints (76.2%). Although abscesses of different size were identified in all patients, absolute paraplegia did not occur in this study. On admission, the neurologic compromise function was assessed: 4 cases of grade C, 12 cases of grade D, and 5 cases of grade E. The mean Cobb angle of the 21 cases at diagnosis was 17.24° (range, 8°–25°), and the median ESR was 28 (range, 2–113) mm/h.

All 21 patients underwent a 1-stage surgical treatment by internal fixation, focal debridement, and fusion via the posterior-only approach. The mean operation time was 230 (range, 120–300) minutes. The median of intraoperative blood loss was 500 (range, 200–1500) mL. Transfusion volume of blood or blood products ranged from 0 to 1500 mL, and 11 cases did not receive transfusions. The detailed surgery information is summarized in Table 2.

For all cases, the mean follow-up period was 29.95 months (range, 15–40 months). After the operation, anemia and hypoproteinemia were corrected in time. The pretreatment ESR ranged from 2 to 113 mm/h, and the postoperative ESR ranged from 5 to 22 mm/h at the 3-month follow-up. The ESR of all patients had significantly decreased (P < 0.05; Table 3). The mean Cobb angles distinctly reduced from 17.24° (range, 8°–25°) preoperatively to 10.38° (range, 5°–16°) in the immediate postoperative period and 12.67° (range, 6°–18°) at final follow-up. The loss of correction was only 2.29° (Fig. 3). The results of post hoc analysis showed there was statistical significance of Cobb angle between each time by LSD t test (P < 0.05; Table 3). During follow-up, the mean bony fusion time is 5.86 (median, 4–9) months. At the last visit, in the 16 patients with preoperative neurologic deficit, 2 with grade C recovered to normal, 2 with grade C recovered to grade D, and 12 with grade D recovered to normal. They had distinct improvement in the Frankel Grade scores postoperatively (P < 0.05; Table 3). Delayed wound healing affected 1 patient because of diabetes, and no sinus formation was observed in the patients. No complications regarding internal fixation and no tuberculosis recurrence were observed during follow-up. All 21 patients obtained complete intervertebral bony fusion (Fig. 4) with a mean time of 5.86 months (range, 4–9 months).

Over the last several decades, although various methods have been introduced and applied to treat spinal tuberculosis, therapeutic strategy for spinal tuberculosis is still controversial. As to this disease, supportive therapy and potent antitubercular chemotherapy are used to stop the infectious process, improve physical condition, and accelerate the patients' recovery process. In the meantime, surgical treatment still plays an important role in treating spinal tuberculosis, with the goals of decompression, debridement, and the rebuilding stability to correct deformity of kyphosis. However, the ideal surgical approach for thoracic and lumbar tuberculosis still appears controversial.

In 1983, Denis first introduced the 3-column model concept to describe the instability of spinal trauma.⁶  Spinal tuberculosis mostly affects the anterior and middle spinal column. Thus, the anterior approach, which was considered the gold standard for a long time, could provide direct access to involved segments and allow efficient decompression of the spinal cord.⁷  However, exposure of the involved segments via an anterior approach would increase surgical trauma, because the segments are blocked by complex anatomical structures especially in thoracic cavity.^8–11  Another practical problem was that a postoperative drainage tube was not easy to install through the thoracic and abdominal cavity. Moreover, it may lead to postoperative hydrothorax and pneumonia. Meanwhile, paravertebral exudate could not be easily drained out and could possibly form encapsulations, potentially leading to relapse. The posterior approach with smaller surgical incisions and minimum surgical trauma could reduce possible postoperative complications. In this study, only 1 patient experienced delayed wound healing because of diabetes, and no severe postoperative complications, including atelectasis, lung infection, and chylothorax, happened. Moreover, postoperative drainage of residual lesions outflowed more easily in the supine position because of gravity.

Another reason for the traditional gold standard of anterior approach is that posterior decompression in spinal tuberculosis should remove the healthy posterior vertebral structures and aggravate spinal instability. Nowadays, the modern pedicle screw could provide rigid fixation and enable adequate stability, which is much better and stronger than anterior fixation.⁹  Moreover, it could also reach a potent braking effect to correct deformity and avoid the loss of corrected degrees as far as possible.^12,13  Meanwhile, fixation and bone graft are far from the lesions in the posterior approach, which could decrease disease relapse.¹⁴  In our study, no breaking or loosening in internal fixation occurred during follow-up. The postoperative Cobb angle was significantly improved, and the loss of correction during follow-up was acceptable (2.29°).

The traditional viewpoint restricting the posterior approach comes from the uncertainty of whether the approach can achieve radical resection of the lesion. In a previous study, the posterior approach provides no advantage in radical debridement.¹⁵  However, with the development of multiple antituberculosis drugs, tuberculous lesions could be successfully treated by means of spontaneous fusion eventually, and the attempt to radical debridement should not be overemphasized.^16,17  When patients have severe spinal tuberculosis with involvement of consecutive multisegmental vertebrae or huge prevertebral abscesses, achieving satisfactory radical debridement by posterior approach is impossible. Nowadays, many surgeons such as Jain¹  and Tuli¹⁸  noted that posterior laminectomy with extensive structure removed may cause overmuch normal posterior spinal column damaged, potentially making the spine unstable. After radical debridement, the extensive damage to the bone graft bed creates difficulty in obtaining good osseous healing. In addition, Chandler and Cappello¹⁹  indicated that a great deal of scar tissue and abnormally proliferated fibrous connective tissue after laminectomy and radical debridement potentially adhered to nerve roots and the dural sac, causing a series of consequent compression symptoms. Focal debridement could overcome the disadvantages of extensive radical debridement by minimizing the damage to spinal integrity, retaining enough vertebral bone substance to guarantee fusion, preventing epidural adhesion, and decreasing the possibility of formation of scar tissue to protect spinal cord. In this study, the postoperative Frankel grade scores was significantly improved at the last visit, and bony fusion occurred with a mean time of 5.86 months. In contrast, autogenous bone grafting is also an important procedure to insure bony fusion. Autogenous bone would generate a satisfied effect of fusion with the minimal possibility of immunologic rejection and secondary infection, whereas 270-degree retention of the spinal structure can ensure the stability as much as possible.

Ideal spinal operations should be well tolerated and nearly minimally invasive, associated with minor postoperative complications and an almost near-normal spine. Thus, we describe the surgical method emphasizing 270° retention of the spinal structure. To the best of our knowledge, no other study has reported this kind of method previously. Thus far, the clinical outcomes and radiographic results of the patients are all satisfactory. Based on our clinical experience, we recommend the indications for the method as follows: (1) stability of the involved segment is lost, and potent internal fixation is in need; (2) patients, reluctant to tolerate or accept the immobilization of lying in bed or wearing the brace for a long time, hope to get out of bed as soon as possible; (3) patients in poor physical conditions have difficulty in accepting 1-stage surgical treatment by anterior approach or surgical treatment by anterior and posterior approach; (4) patients do not have a large abscess, serious spinal canal invasion, and spinal deformity.

Combined with multiple antituberculosis drugs, this method with 270° retention of the spinal structure via the posterior-only approach can maintain spinal column integrity, improve neurologic function, achieve satisfactory bony fusion, decrease postoperative complications, and lead to effective recovery in the end.