Spinal tuberculosis primarily affects the anterior column of the spine. Once the microorganism has been eradicated, kyphosis remains the most important problem to be treated or prevented. The commonly used method of placing anterior strut grafts in the defect created by excision of the infected vertebral bodies corrects the kyphotic deformity, but the grafts are prone to failure or resorption especially when more than two vertebrae have been excised17,21. Posterior instrumentation is used to protect the graft, stabilize the segments, and prevent progression of the deformity.
In the current study, we assessed the results of use of anterior instrumentation in patients who had a moderate-to-severe kyphotic deformity or involvement of more than two vertebral levels in association with spinal tuberculosis.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Department of Orthopaedics and Traumatology, University of Ankara Medical School, Ibn-i Sina Hospital, 3.Cad. 55/3 Bahçelievler, Ankara 06500, Turkey. E-mail address for Dr. Yilmaz: cyilmaz@dialup.ankara.edu.tr.
Between January 1992 and December 1996, thirty-eight of 183 patients who had spinal tuberculosis had moderate-to-severe kyphosis or involvement of more than two vertebrae and were managed with débridement, anterior decompression, interbody arthrodesis with strut-grafting, and anterior instrumentation (Figs. 1-A, 1-B, and 1-C). The severity of the kyphosis was classified according to the criteria proposed by Kaplan2, with kyphosis of less than 30 degrees considered mild; 30 to 60 degrees, moderate; and more than 60 degrees, severe. The indications for operative intervention were progression of the deformity and neurological involvement. The type of instrumentation that was used depended on availability. A plate-and-screw-based system was used for patients who had one level of involvement, and systems similar to Kaneda instrumentation were used for those who had involvement of more than one level.
There were twenty male and eighteen female patients. The average age of the patients was forty-three years (range, eighteen to sixty-five years). The average duration of follow-up was two years and five months (range, two years to six years and four months).
Back pain, the most common symptom, was present in twenty-eight patients; thigh pain (of the radicular type), in eighteen; paraparesis, in eighteen; active pulmonary tuberculosis, in ten; paraplegia, in six; cauda equina syndrome, in four; vertebral hemangioma, in four; a draining fistula, in four; and a herniated disc, in two.
Manual muscle-testing of the patients who had paraparesis revealed grade-1, 2, or 3 strength in both lower extremities, according to the strength-grading scale of the Medical Research Council of Great Britain6. All patients had hypoactive deep-tendon reflexes, normal findings on sensory examination, a positive skin test with purified protein derivative, and an elevated erythrocyte sedimentation rate. Plain radiographs, computerized tomography scans, and magnetic resonance images were made for all patients.
Twenty-two patients had involvement at the thoracolumbar level (the twelfth thoracic vertebra was involved in twelve and the first lumbar vertebra, in ten), ten had involvement of thoracic vertebrae cephalad to the twelfth level, five had involvement of lumbar vertebrae caudad to the first level, and one had involvement at the lumbosacral level. One disc-space level was involved in eight patients; two levels, in fourteen; three levels, in ten; and four levels, in six. An epidural abscess was seen in twelve patients, and a paravertebral abscess was seen in twenty-four; the two remaining patients did not have an abscess (Table I). The abscess initially was noted on the imaging studies, and its presence was confirmed intraoperatively.
The kyphosis was measured on lateral radiographs as the angle between the superior end plate of the first uninvolved vertebra cephalad to the affected segment and the inferior end plate of the first uninvolved vertebra caudad to that segment. In the twenty-two patients who had involvement of one or two levels, the average kyphosis was 59 degrees (range, 34 to 77 degrees). In the sixteen patients who had involvement of three or four levels, it was 52 degrees (range, 15 to 87 degrees). The height of the vertebral body was decreased by more than one-third in twenty-six patients, and it was decreased by more than one-half in twelve. Nine patients also had scoliosis: seven had a left thoracic scoliosis that measured between 6 and 12 degrees, and two had a left lumbar scoliosis that measured between 14 and 18 degrees.
As mentioned, all patients were managed with anterior débridement, decompression, interbody arthrodesis with strut grafts, and anterior instrumentation. An intrapleural approach was used for the thoracic vertebrae, and a retroperitoneal approach was used for the lumbar vertebrae. Perioperative frozen-section specimens were examined to confirm the presence of tuberculosis. The histopathological sections showed Langerhans giant cells, granuloma, and caseating necrosis. The instrumentation consisted of rods and screws placed in the vertebral body and extended one level cephalad and one level caudad to the affected vertebrae. The strut grafts were obtained from the resected ribs in twenty-four patients, whereas fibular allograft was used in eight, autogenous tricortical iliac graft was used in four, and both resected rib and fibular allograft were used in two.
The average operating time was 4.7 hours (range, 3.5 to 6.8 hours), and the average blood loss was 1560 milliliters (range, 778 to 3450 milliliters).
Postoperatively, histopathological examination and cultures of material obtained with biopsy and from the débridement confirmed the diagnosis of tuberculosis. All patients received antituberculous chemotherapy that included streptomycin (one gram per day for two months for patients who weighed fifty kilograms or more and 750 milligrams a day for those who weighed less than fifty kilograms), rifampicin (600 milligrams per day for nine months for patients who weighed more than forty-nine kilograms and 450 milligrams per day for those who weighed less than fifty kilograms), isoniazid (five milligrams per kilogram of body weight per day, for a maximum of 300 milligrams per day for nine months), and pyrazinamide (two grams per day for two months for patients who weighed more than forty-nine kilograms and 1.5 grams per day for those who weighed less than fifty kilograms).
The patients were allowed to walk while wearing a molded thoracolumbosacral orthosis on the seventh postoperative day. The brace was worn for a minimum of six months.
Lateral and anteroposterior radiographs were made in the immediate postoperative period, at three and six weeks, and at three months; they then were made every three months during the first year, after which they were made every six months until the time of the latest follow-up evaluation. If there was no evidence of fusion at three months, radiographs were made monthly until fusion was noted. The presence of fusion was determined by the absence of localized pain and tenderness over the site of the arthrodesis, the maintenance of correction of the deformity, the absence of motion as seen on lateral bending radiographs, and evidence of fusion on radiographs and computerized tomography scans. Computerized tomography scans were made at six months to determine if there had been any recurrence of the disease3,20. The erythrocyte sedimentation rate was monitored for the presence of an active disease process.
Neurological examination was performed at each follow-up visit with use of the classification system of Frankel et al.5. According to this system, type A indicated a complete spinal cord injury; type B, a spinal cord injury with only sensation present; type C, an injury with motor function present but not useful; type D, an injury with useful motor function; and type E, an injury with intact neurological function.
Of the twenty-two patients who had involvement of one or two levels, sixteen had a preoperative kyphosis that ranged from 30 to 60 degrees and six, a kyphosis that was more than 60 degrees; postoperatively, the kyphosis ranged from 10 to 28 degrees in all twenty-two patients. Of the sixteen patients who had involvement of three or four levels, six had a preoperative kyphosis that ranged from 0 to 30 degrees; six, 30 to 60 degrees; and four, more than 60 degrees. Postoperatively, the kyphosis in these sixteen patients ranged from 0 to 30 degrees. The correction was maintained in twenty-one patients, and there was a maximum loss of 3 degrees in sixteen patients. One patient died on the second postoperative day because of pulmonary insufficiency. This patient had disseminated tuberculosis and also had paraplegia.
One of the six patients who had paraplegia died, one recovered useful motor function, and the remaining four had improvement but did not regain normal strength. Fourteen of the eighteen patients who had paraparesis recovered complete neurological function, and the remaining four regained grade-3 strength in the lower extremities. The four patients who had cauda equina syndrome had recovery of bowel and bladder function and sensation as well as relief of pain after the index procedure.
Two of the three patients who had a type-A injury, according to the system of Frankel et al.5, preoperatively had improvement by one level, to type B, postoperatively, and one died. Of the three patients who had a type-B injury preoperatively, two had improvement by one level, to type C, and one had improvement by two levels, to type D. Of the ten patients who had a type-C injury preoperatively, eight had complete neurological recovery, to type E, and two had improvement by one level, to type D. All eight patients who had a type-D injury had complete neurological recovery.
The twenty-two patients who had involvement of one or two levels had evidence of fusion on the radiographs made at the end of the third month, and the sixteen patients who had involvement of three or four levels had such radiographic evidence at the end of the sixth month. There were no recurrent or persistent infections. The thoracic scoliosis was corrected to 0 degrees after the operation, and the maximum progression at the time of the latest follow-up evaluation was 3 degrees. One patient who had a lumbar curve of 8 degrees had no change in the deformity, and another patient had residual scoliosis of 3 degrees.
Three patients had persistent pain that was believed to be related to malposition of the screws. In two of the four patients who had a draining fistula, a superficial wound infection developed, but it responded to appropriate antibiotic therapy. The fistula in the other two patients healed spontaneously without use of antibiotics. Postoperatively, six patients had a urinary tract infection, four had atelectasis, three had ileus, three had deep-vein thrombosis, and one had pneumonia. The urinary tract infection and the pneumonia responded to antibiotic therapy. The chest tube was kept in place until the atelectasis resolved. Ileus was treated with nasogastric suction, and deep-vein thrombosis was treated with anticoagulation.