Modern anti-tuberculous chemotherapy has revolutionized the treatment of tuberculous spondylitis, but residual kyphosis with or without neurologic deficit is a possible complication after conservative treatment. The British Medical Research Council Working Party on Tuberculosis of the Spine performed a large scale, controlled trial of various treatment methods11-13
. They concluded that chemotherapy was an effective treatment for the majority of the patients with tuberculous spondylitis11,13
. However, kyphotic deformity and delayed bony union are possible complications of conservative treatment11,13
Since Hodgson and Stock1
first reported on radial debridement and anterior fusion, it has been advocated as the treatment of choice for the patients with tuberculous spondylitis. It has the advantages of direct access to the focus of infection, the ability to effectively decompress the neural tissue, and it can stabilize the spine by a strut bone graft. However, graft slippage, resorption, subsidence and fracture of the graft were observed by many surgeons and the subsequent increased incidence of kyphosis was reported3,14
. Kim et al.15
reported the results of radical debridement and anterior iliac bone grafting in 140 patients. In their series, the kyphosis correction was 55.1% just after surgery, and the correction of the initial angle was decreased to 7.5% at 2 years after surgery. Rajasekaran and Soundarapandian3
suggested that the patients having a graft longer than two disc spaces might benefit from additional measures for preventing graft collapse, such as an extended period of non-weight bearing, the performance of posterior arthrodesis and the prolonged use of a brace.
To prevent increasing kyphosis after anterior fusion, supplementary posterior fusion and instrumentation have been advocated by many surgeons5,7-9
. Moon et al.5
reported that posterior instrumentation after anterior fusion was helpful in providing early fusion and preventing the progression of kyphosis. They used various posterior instruments such as a hook and rod system or a rod and segmental wiring technique. The collapsed vertebrae regained their height with these procedures and the correction was successfully maintained. However, Harrington instrumentation or segmental wiring has bio-mechanically weaker holding power than the modern pedicle screw system, and the former techniques usually need more levels of instrumentation for the purpose of decreasing the abnormally high stress on the metal or bone lamina. Actually, pedicle screws have strong biomechanical holding power on the anterior column as well as on the posterior column, which can effectively share the anterior load on the spinal column. The posteriorly instrumented pedicle screw system could distribute the load to the anterior column and possibly decrease the number of fusion levels. We have started to use the pedicle screw system in patients with tuberculous spondylitis after anterior debridement and arthrodesis as early as the late 1980s. In the thoracic region, where the segmental motion is not so large, we preferred fusing two vertebrae above and below the affected lesion. In the lumbar region, where we try to save the segmental motion as much as possible, we usually fused one vertebra above and one below the lesion.
Güven et al.9
reported an average of 3.4 degrees loss of correction after posterior drainage and Cotrel-Dubousset instrumentation. Moon et al.5
reported on the results of anterior arthrodesis and the posterior Harrington or segmental wiring technique, and the loss of correction did not exceed 3 degrees in their series. Chen et al.8
used the Harrington technique or Luque instrumentation after anterior iliac strut grafting, and they observed less than 3 degrees loss of the initial correction. In our current series, the loss of correction during the follow-up was an average of 6.2 degrees despite performing pedicle screw instrumentation. It may be due to the relatively short length of the instrumented levels with our technique and the immediate ambulation of the patients without using a heavy immobilizing cast or brace. Even though the average loss of correction in our series was slightly higher than that of the previous studies that used posterior implants5,8,9
, we did not experience any unsightly increasing kyphosis or metal failure after the use of pedicle screws. A slight increase of the correction loss is a matter of making numeric measurement on the radiographs and it did not bring about cosmetic or functional problems in our patients. The patients in our series were able to ambulate wearing a light plastic brace immediately after stabilizing the spine with the implants.
We started to use the titanium mesh as early as the early 1990s in patients who had tuberculous spondylitis with a large bone defect and who were proven not to have active liquefied pus in the surgical field at the time of anterior debridement. Many recent studies have reported that titanium mesh cages were an effective and safe anterior column reconstruction method for tuberculous16
and even for septic spondylitis17,18
. We have found that usage of a mesh cage usage in patients with tuberculous spondylitis does not inhibit control of infection, but we are still reluctant to use it in patients with pyogenic spondylitis. Titanium mesh in patients with tuberculous spondylitis eliminated a need for iliac bone harvesting because we filled the mesh with resected rib bone that was acquired during the transthoracic and thoracoabdominal approaches. We sometimes did not perform additional posterior surgery after anterior mesh insertion. In the case of short segment involvement in the thoracic region, a titanium cage filled with rib bone was a good option for anterior reconstruction, and this eliminated the need for iliac bone harvesting or posterior instrumentation. A plastic thoracolumboscral orthosis brace was good enough to support the thoracic spine because a rib cage acts as the fourth column of the spine in this region. However this titanium mesh only option without posterior instrumentation seemed to be somewhat unstable in patients with multi-segmental involvement. We observed mild migration of the mesh in the patients who had multilevel involvement, but we had to omit posterior fixation due to the patients' poor general condition. We do not now recommend the titanium mesh only option for the multi-segmental involvement cases. Titanium mesh was an especially useful alternative to strut bone grafting for the patients who had multilevel involvement with a long segmental defect after anterior debridement and there was a need to harvest a long length of structural graft. It was also helpful for patients who had osteoporosis with poor iliac bone quality. For this population of patients, a strut bone graft for anterior column reconstruction can have suboptimal mechanical strength.
Statistical analyses showed that the amount of surgical correction and the loss of correction during follow up was not influenced by the operation method, the patient's age and the number of involved segments. The loss of correction was greater in the thoracolumbar region than that in the thoracic region. The thoracolumbar junction is well known as an area of concentrated stress due to the biomechanical transition from a stiff thoracic spine to a mobile lumbar spine. This biomechanical fact could contribute to the high loss of correction in the thoracolumbar region. Krompinger et al reported that 36% of the spine fractures at the thoracolumbar junction progressed 10 degrees or more at follow-up, whereas no thoracic facture and only one lumbar fracture progressed 10 degrees or more19
. There were no statistical differences among three kinds of operation methods, such as mesh only, posterior pedicle screws after anterior bone grafting, and pedicle screw after anterior mesh, regarding the surgical correction and the loss of correction. However, this may be due to the small number of patients in each group of our series, and further studies with a larger number of patients are needed.
Yilmaz et al.6
believed that anterior instrumentation is more effective than posterior instrumentation for reducing the deformity and stabilizing the vertebral column in patients who have kyphosis due to spinal tuberculosis. They used rods and screws placed in the vertebral bodies that extended one level cephalad and one level caudal to the affected vertebrae. The average correction of deformity was 64% in the patients with one or two levels of involvement and it was 81% for the patients with more than two levels of involvement. The maximal loss of correction was 3 degrees. They did not experience recurrence of disease. We put in the titanium mesh during the anterior surgery, but we did not insert an anterior screw device at the same time. For anterior instrumentation of the spine with using a screw system, a surgeon needs to extend the surgical dissection to the healthy vertebra for inserting the screw system in the cephalad or caudal direction. This usually requires more extended dissection of vital structures, including the large vessels adjacent to the vertebral column; moreover, if there is some infectious adhesion of the paravertebral structures, then the surgical dissection is even more difficult. Moreover, if the infection extends, after surgery, to the adjacent vertebra where the screws were inserted, then failure of fixation can lead to an anterior catastrophic result. For these reasons, we minimized the extent of the anterior surgery on the affected vertebra and we did not fixate the spine with using an anterior screw system for stabilization, even though single-staged anterior instrumentation has many benefits. Posterior instrumentation also can minimize the burden of metallic foreign bodies at the focus of infection and it can maximize the stability of the spine by the formation of a posterior tension band.
The anterior surgical approach has advantages of direct access and excision of the lesion with the ability to decompress the neural elements3,20
. Although some authors have reported a good clinical result for performing posterior transpedicular drainage and instrumentation in the patients who had tuberculosis of the spine without neurologic involvement9
, we believe anterior debridement and arthrodesis are indispensable for achieving successful surgical treatment of the patients with neurologic deficits. In our series, the neurologic deficits of the patients disappeared after anterior surgery in the most of the cases.
In principle, metallic implants or foreign bodies are contraindicated at the site of musculoskeletal infection. There is a fear of persistent infection around the metal that is caused by preferential adhesive bacterial colonization on an inert surface. Oga et al.4
reported that microorganisms such as Staphylococcus epidermis attach to the inert surface and colonize it, and then they produce an extrapolysaccharide biofilm that protects the organism from host defense mechanisms and antibiotics. However they found that mycobacterium tuberculosis was less adhesive, and they observed only a few biofilm-covered microcolonies around the stainless steel in vitro. They suggested that mycobacterium may have different adhesive properties and adhesive action to biomaterial surfaces than do other bacteria and mycobacterium may produce less biofilm. Clinically, they also did not experience persistent or recurrence of infection after anterior radical surgery followed by posterior instrumentation. Several other authors have subsequently reported clinical studies that have supported the microbiological safety of using metallic implants in patients with tuberculous spondylitis5,6,8,9,21
. In our series, we had no case of persist infection when the patients were treated by surgeries with using metallic implants and anti-tuberculous combination chemotherapy. The inflammatory parameters such as the erythrocyte sedimentation rate and the C-reactive protein level were eventually normalized in our patients and the use of metallic implants did not inhibit control of tuberculous infection. We also did not experience any late recurrence of infection.