Spinal TB is defined as a chronic infection caused by Mycobacterium tuberculosis which most commonly affects the thoracolumbar junction but the whole spine could be involved [2,3,4,5]. It is usually caused by hematogenous spread of offending pathogen from the primary infection site, most common in pulmonary region or genitourinary system [14], to the vasculature of vertebral bodies. As the reputation of a great mimicker in medicine, the clinical pictures of spinal TB are varied with a wide range from insidious back pain to severe spinal deformity, neurologic deficit, and even paralysis [1,2,3]. The non-specific manifestation makes timely diagnosis, the most crucial part for successful treatment and avoiding complications, remain a challenge for clinicians. Typically, the presentations in images involve destruction of the vertebral bodies and intervertebral disk space, collapse of the spinal structures, and anterior wedging leading to kyphosis and gibbus deformity. However, many atypical features have been reported in the literature.

Pande et al. presented a new classification of atypical spinal TB and described that noncontiguous lesion is one of the atypical manifestations [15]. This atypical condition is not uncommon with the reported incidence ranging from 1.1 to 16.3% in the literature [16]. Kaila et al. even reported an incidence up to 71.4% by reviewing the whole spine magnetic resonance imaging (MRI) of 14 patients [17]. Several possible reasons had been mentioned to explain the high incidence of skipped spine TB. First, it may be related to embolic spread of bacteria to multiple levels of vertebrae which is similar to the phenomenon of multiple metastasis in tumor cases [18]. Second, Batson’s paravertebral venous plexus, a valve-less system in the vertebrae, allows the spread of infection along it and is considered as a reason in developing noncontiguous lesions [19]. However, this is still controversial and the real reason remains unclear.

Conservative management with multiple anti-TB drugs is the mainstay in managing spinal TB because varying categories of bacilli could exist. It could also reduce the instance of drug resistance. According to the recommendations of WHO, four drugs: isoniazid, rifampicin, pyrazinamide, ethambutol, or streptomycin, are administered for two months in the initiation phase followed by two drugs—isoniazid and rifampicin, for seven months in the continuation phase. Kanamycin, amikacin, capreomycin, levofloxacin, etc., are considered as second-line drugs and should be used carefully as they are expensive and have more side effects [13]. However, the adequate treatment should be individualized and should be based on patients’ conditions, response to treatment, and results of culture. In our case series, we followed a treatment protocol with multidisciplinary team care as it could improve the outcomes [20, 21]. Eventually, all the patients could be cured from spinal TB.

Surgery for spinal TB is usually reserved for patients with (1) refractory disease, (2) severe kyphosis, (3) pan-vertebral lesions, (4) progressive neurologic deficit, and (5) clinical deterioration [1,2,3,4,5]. The operation is commonly accessed through the anterior retroperitoneal or transthoracic approach for debriding the infective tissue and decompressing neurologic elements comprehensively. Then, the anterior column was reconstructed using autograft, allograft, or vertebral body spacers. For obtaining an immediate stability, supplemental posterior instrumentation after anterior column reconstruction in a single- or two-stage manner could be carried out through additional posterior approach [22, 23]. This combined surgery is the most secure procedure for patients with severe deformity, instability, and neurologic deficit. However, the major concerns for this combined approach are (1) longer operative time, (2) greater blood loss, (3) higher risk of complications, (4) greater surgical trauma for patients, (5) the need to perform diaphragm take down and rib cutting, and (6) technically difficult [24]. Due to the above-mentioned reasons, this is not an ideal technique, especially for patients who are older or with multiple comorbidities because it may hurt such fragile cases.

To decrease the morbidities and disadvantages related to the combined approach, the technique through a posterior-only approach in single-stage manner had been proposed. Circumferential debridement, decompression of neurologic elements, and even total or subtotal corpectomy followed by anterior reconstruction could be carried out through the posterolateral transpedicular approach without sacrificing the nerve roots. The blood loss, operative time, surgery-related complications, and length of hospitalization can be reduced. In addition, most of the surgeons are more familiar with the posterior instead of anterior approach, and revision surgery is much easier if necessary. This technique had been used to treat a wide range of spinal disorders and resulted in a satisfactory outcome [25,26,27].

Autograft, allograft, or vertebral body spacer could be used for anterior column reconstruction after circumferential debridement through transpedicular approach. Different pros and cons exist in each choice, and there is still no consensus in which one is better [28,29,30]. Autogenous iliac or fibula strut bone graft is an ideal choice for anterior column reconstruction during spine surgery as it is osteo-inductive and osteo-conductive, biomechanically stable, and biocompatible. However, the major concern of using autogenous strut bone graft is the high incidence of donor site morbidities [31, 32]. The need for creating another wound followed by an additional procedure for harvesting the bone graft can cause discomfort in these fragile patients. An expandable cage is another recommended choice as it could be passed through the route between the nerve roots and then easily be expanded to adequate length [33, 34]. However, it is not an ideal alternative due to the lack of osteo-inductive and osteo-conductive properties. In our case series, we used allogenic fibular bone graft to reconstruct the anterior column. We chose fibula because the size is suitable for inserting it anteriorly through the route between nerve roots. The major advantage of allograft is that it eliminates the need for additional surgical procedure and prevents donor site morbidities. However, there are some disadvantages related to allograft which included accelerated bone resorption, delayed vascular penetration, slow bone formation, and incomplete or delayed graft incorporation. Some authors also had reported a good outcome by using allograft to reconstruct spinal structure [30]. In the present study, none of our patients experienced an allograft-related complication and achieved good outcome.

Although the clinical outcomes are quite satisfactory in our patients, there are still some limitations in this study. First, there were only 12 patients enrolled; the sample size is too small to prove the efficacy and feasibility of using this technique to treat skipped spinal TB. Second, the retrospective nature of our study design lacked randomization. Hence, it was impossible to enroll patients who underwent different treatment methods for subsequent comparison. Third, the infectivity levels and patients’ health conditions are diverse which may influence the analysis. Fourth, computed tomography was not routinely used to evaluate the fusion of fibular allograft due to the policies of our national health insurance. As plain radiographic examinations alone are not enough to assess the union of allograft, some nonunion or pseudoarthrosis may be undetected. Finally, the quality of bone in each patient is unknown because we did not conduct bone mineral density examinations routinely which may affect the outcomes. For these reasons, further prospective randomized studies are required to prove the feasibility and efficacy of the technique.