POSTERIOR VERTEBRAL COLUMN RESECTION (PVCR): A SAFE AND EFFECTIVE OPTION TO TREAT KYPHOSIS FROM SPINAL TUBERCULOSIS

ABSTRACT Objective: To evaluate the clinical and radiological results of posterior vertebral column resection (PVCR) for the treatment of kyphosis from spinal tuberculosis. Methods: Retrospective study of a series of 14 cases, 9 female and 5 male, with a mean age of 28 years (1 to 64) at the time of surgery, with spinal tuberculosis with mean kyphosis of 56.42° (2° to 95°). Results: All patients underwent surgical treatment with PVCR, with a mean number of 2.57 (1 to 6) resected vertebrae and a mean number of instrumented vertebrae of 6.14 (4 to 8). The mean kyphosis correction was 64% after up to one year of follow-up. Prior to surgery, eight patients had signs of spinal cord impairment (ASIA score ranging from A to D), and six did not present any deficit of strength or sensory function (ASIA E). There was no neurological worsening and everyone with deficits improved by at least one degree on the scale. In the postoperative follow-up, bone consolidation was observed in all patients, except one. Among the intra- and postoperative complications, pleurotomy was the most prevalent. Conclusion: Posterior vertebral column resection has proven to be a safe and effective option for the treatment of kyphotic deformity from spinal tuberculosis. Level of evidence IV; Therapeutic studies - Investigation of treatment results.


INTRODUCTION
Tuberculosis, even today, is considered a global public health challenge, and Brazil, according to the World Health Organization (WHO), is one of the priority countries. 1 According to WHO data, it is estimated that about 6 million new cases occur worldwide and in Brazil, according to the Notifiable Diseases Information System (SINAN), 70,000 new cases emerge annually and 4,500 deaths occur as a result of this disease. 2 The National Institute of Traumatology and Orthopedics (INTO) is one of the reference hospitals in the Unified Health System (SUS) and frequently receives cases of bone tuberculosis. Bones are a common extrapulmonary tuberculosis site and the spine is the most prevalent location, 3 corresponding to about 50% of bone tuberculoses. 4,5 Although the gold-standard treatment for spinal tuberculosis is prolonged antibiotic therapy, 6,7 surgical treatment may be the only option in selected cases. Included among these are patients with important neurological deficit, with spinal instability, or with severe or progressive deformity.
There are various modalities of surgical treatment with proven effectiveness for spinal tuberculosis, among them the anterior approach 8,9 and the combined approach. 10,11 However, in cases of severe kyphotic deformity vertebrectomy is the routine choice. Since Suk et al. first described the posterior vertical column resection (PVCR), this technique has been shown to be safe, as it presents lower morbidity when compared to vertebrectomy via a combined approach. 12 PVCR permits circumferential decompression of the dural sac and ample debridement of the devitalized tissues, in addition to promoting significant correction of the kyphotic deformity that often accompanies complicated spinal tuberculosis.
The objective of this study was to evaluate whether posterior vertical column resection is a safe and effective option for the surgical treatment of spinal tuberculosis.

METHODS
This is a retrospective study of 14 patients with spinal tuberculosis who underwent corrective kyphotic deformity surgery via the PVCR technique at the Instituto Nacional de Traumatologia e Ortopedia Jamil Haddad, during the period from November 2012 to October 2018. All patients were operated by the same surgical team.
Patients with a confirmed microbiological or histopathological diagnosis of spinal tuberculosis, regardless of age, who were submitted to surgical treatment performed using the PVCR technique were included. Patients without a confirmed microbiological or histopathological spinal tuberculosis diagnosis, as well as patients who had undergone PVCR due to other pathologies, such as a tumor or congenital malformation, were excluded.
The patients were analyzed according to age, sex, the vertebral segments affected, neurological status, the magnitude of the kyphotic deformity at diagnosis and at the end of a minimum follow--up of one year, the number of instrumented levels, the number of resectioned vertebrae, the rate of consolidation, and complications. Collection of clinical and demographic data was conducted through analysis of the medical records of the selected patients.
The assessment criterion for the neurological status of the patients was the Frankel scale as modified by ASIA (American Spine Injury Association). 13 Anteroposterior and lateral panoramic spinal radiographs were analyzed prior to surgery, in the immediate postoperative period, and in outpatient follow-up consultations. The angle of kyphosis was measured by the Cobb method in the lateral radiograph. As most of the patients in our series had some degree of neurological deficit, we decided not to perform the deformity flexibility test. The measurements were taken using Surgimap Spine software (Nemaris Inc, New York, NY). The consolidation rate was evaluated by means of computed tomography performed at least 6 months after the surgery.
In the patients under 10 years of age, due to the small spinal dimensions and attempting to avoid skin complications, posterior cervical spine implants with screws and 3.5 mm rods were used ( Figure 1). In the other patients, conventional thoracolumbar spine fixation materials with screws and 5.5 mm rods were used.
Following the recommendations of the health authorities and with respect to the social isolation imposed due to the new coronavirus (Covid-19) pandemic, we opted to waive the application of the Informed Consent Form (ICF) for this research project. This study was approved by the Institutional Review Board of the Institute as number CAAE: 32791520.9.0000.5273.

Surgical Technique
The patient, under general anesthesia and multimodal neurophysiological monitorization, is positioned in ventral decubitus on cushions on a radiotransparent table. A medial longitudinal incision is made over the planned instrumentation levels. A subperiosteal medial to lateral dissection with meticulous hemostasis is performed and, at the level planned for the PVCR, the transverse process and 3 or 4 medial centimeters of the corresponding posterior rib are exposed. Facetectomies and bipedicular instrumentation of the levels to be operated, generally 2 to 3 levels above and below the PVCR, are performed. After, fluoroscopic confirmation, a total laminectomy and a costotransversectomy of the levels to be resected are performed. Before starting the resection of the vertebral body, a temporary contralateral rod must be placed as support to prevent translocation and consequent iatrogenic injury to the spinal cord. During this stage, greater care must be taken to avoid the occurrence of pleural injury. Subperiosteal dissection of the vertebral body is initiated and directed anteriorly from the lateral faces of this body. Malleable retractors are used to protect the pleura and large vessels during this maneuver. During the dissection and exposition of the lateral wall of the vertebral body, hemostasis must be meticulous to prevent lesion of the segmental vessels. The vertebral body and the intervening discs are removed. To complete the resection, the temporary fixation rod is moved to the other side and the remaining contralateral vertebral body is resected in the same way. Correction of the deformity is accomplished by in situ modelling of the rods and exchanging the hypomodeled provisionary rods for the desired form, one by one. To prevent an inadvertent distraction of the spinal cord and neural elements, the spine must initially be shortened, with light compression over the resected gap. After the correction is achieved, the temporary rods are substituted one by one by the definitive rods, thus preventing loss of correction. The remaining anterior gap should be filled with autologous bone graft contained in a titanium cage. Anterior or circumferential arthrodesis is performed at the level of the osteotomy and posterior fusion is performed at all the instrumented levels. Following placement of the posterior graft, the surgical wound is closed by planes and a subcutaneous suction drain is installed.

RESULTS
The mean patient age at the time of surgery was 28 years (1 to 64). All the patients had important segmental kyphosis or significant reduction in lumbar lordosis, with mean angulation of 56.42° (2° to 95°). Prior to surgery, eight patients had signs of spinal distress (ASIA score ranging from A to D). Six patients did not present strength or sensitivity deficits (ASIA E). There was no neurological worsening and all patients with neurological deficits improved by at least one scale grade. The mean number of resected vertebrae was 2.57 (1 to 6) ( Figure 2) and the mean number of instrumented levels was 6.14 (4 to 8). Mean residual deformity after correction in the sagittal plane was 20.28° (-31° to 61°), with a mean correction rate of 64%. (Figure 3) Computed tomography performed 6 months after surgery showed bone consolidation inside the titanium cage ( Figure 4) in all patients except one, who is still in outpatient followup and remains asymptomatic. The most common complication was pleurotomy, observed in 6 patients, in whom intraoperative thoracic drainage was performed. (Table 1)

DISCUSSION
Historically, the surgical technique described by Hodgson et al., the Hong Kong procedure, 12,13 alluding to the University of Hong Kong, the center where it was first performed, was used to treat vertebral tuberculosis surgically. The procedure consists of an anterior approach to the spine, combined with ample debridement and anterior arthrodesis with the use of a tricortical graft. The works published by the University of Hong Kong showed the superiority of the procedure as compared to focal debridement, performed alone. As regards the rate of pseudoarthrosis and resolution of the disease, the technique described in 1956 has a satisfaction rate of 93%. Our work presented a similar fusion rate of 92%.
The combination of anterior approach instrumentation and posterior arthrodesis, although with no difference in the eradication of    15 The rate of correction also is greater in association with the posterior approach. 16 Qureshi et al. indicated 47% correction with the advent of this technique and only 9% with the isolated anterior approach. 17 PVCR provides, through a single approach that is familiar to the spine surgeon, access to the anterior, middle, and posterior spines. This approach allows circumferential decompression of the spinal cord, a significant reduction in kyphotic deformity, and good preparation for interbody arthrodesis. [18][19][20] Hassan et al. obtained 26 degrees of kyphosis correction via the isolated posterior approach, despite the longer surgical time. 21 Zhou et al. reported a 68% kyphotic deformity correction rate in tuberculosis. In comparison, our study showed a mean of 20 degrees and a rate of 64%. 22 Another important advantage of PVCR when compared to double access is to avoid the morbidity inherent in anterior access, as the patient with tuberculosis many times suffers from significant pulmonary and systemic impairment.
Zhou et al. reported that kyphoses with Cobb angles greater than 60 degrees are not good indications for the anterior approach due to the difficulty of access to the apex of the deformity. 22 As demonstrated in studies of the anterior approach, thoracic spine deformities that involve multiple levels have higher failure and postoperative correction loss rates, which contributes to the choice of PVCR in cases of severe kyphotic deformity.
Despite excellent radiographic results, PVCR is a technically complex and challenging procedure, associated with a high rate of complications. Dorward and Lenke conducted a review and reported that the mean rate of neurological complications in performing PVCR was as high as 14.3%, mainly explained by biomechanical instability during surgery. 23,24 With the goal of preventing iatrogenic lesions of the spinal cord, the temporary rods must always be exchanged for new rods to ensure the mechanical integrity of the system and spinal stability. 25 The postoperative use of thoracolumbosacral orthosis for 3 months in patients with resection of one or two vertebral segments and for 6 months in patients submitted to resection of three or more segments is also indicated. 26 In our case series, no loosening of screws or rod breakage was observed, nor was there neurological worsening in any patient, and all who had deficit improved by at least one grade on the ASIA scale.
PVCR was applied to all cases in our study, including pediatric patients, as they satisfied the severe deformity criterion or had associated neurological deficit. It is important to perform pedicular instrumentation in these cases, despite the technical difficulty, to create effective post-osteotomy stabilization. According to Harms et al., pedicular instrumentation in very young children did not produce adverse effects to additional vertebral growth, which was comparable to the anteroposterior and craniocaudal growth of instrumented vertebrae and adjacent non-instrumented vertebrae. 27 Correlating our results with those in the literature, PVCR is a safe and effective surgical indication in the treatment of severe kyphosis from tuberculosis in all age groups, with proven success in deformity correction, the maintenance of that correction, and in the bone consolidation rate, especially when the patient has involvement of the thoracic spine or of 2 or more vertebral levels.