- Email: [email protected]
Results of nonsurgical treatment of thoracic spinal tuberculosis in adults
The Spine Journal 5 (2005) 79–84
Results of nonsurgical treatment of thoracic spinal tuberculosis in adults Abhay Nene, MS (Orth), Shekhar Bhojraj, MS (Orth), FCPS (Orth), D Ortho* The Spine Clinic, P.D. Hinduja National Hospital and Medical Research Centre, Veer Savarkar Road, Mahim, Mumbai 400 016, India Received 15 January 2004; accepted 24 May 2004
Abstract
BACKGROUND: The indications for surgery in spinal tuberculosis have been controversial, and more so recently, in the era of renewed understanding of the concept of multi-drug-resistant tuberculosis along with newer modalities of spinal instrumentation. Indications for surgery need to be redefined in this context. PURPOSE: To assess the efficacy and results of nonsurgical treatment in thoracic spinal tuberculosis in adult patients, and redefine indications for surgery. STUDY DESIGN: We present a retrospective analysis of 70 adults with thoracic spinal tuberculosis, with varying presentations, including abscesses and neurological deficits, seen at our spine clinic, in a period between August 1998 and August 2000, treated largely nonsurgically, with rewarding results. METHODS: A retrospective study was made of 70 adult patients with thoracic spinal tuberculosis presenting at our spine clinic, between August 1998 and August 2000. All patients were subjected to medical management, unless there were specific indications for surgery, as per our protocol, wherein absolute indications of surgery in adults included advanced neurological deficit (less than Grade 3 by 5, by the 5-point grading system of the Medical Research Council), neurology worsening while on antituberculous chemotherapy, diagnosis in doubt on clinicoradiological evaluation and significant kyphosis (greater than 40 degrees) on presentation. Clinical and radiological assessment of results was made by an independent observer, at a mean follow up of 40 months. RESULTS: Forty-four patients presented with abscesses, 21 of which were epidural. Seven had neurological signs of cord compression on clinical examination at presentation. Over 98% of our patients (69 of 70) were successfully treated conservatively, and none of these had any residual instability, radiculopathy or neurological compromise. Seventy-four percent had excellent to good results, with no mechanical residues of the disease, and 23% had residual kyphosis, which was clinically obvious, but biomechanically irrelevant. CONCLUSIONS: We think that tuberculous spondylodiscitis in adults can be well managed conservatively in a vast majority of cases, and indications for surgery are few and specific. 쑖 2005 Elsevier Inc. All rights reserved.
Keywords:
Thoracic spine; Tuberculosis; Nonsurgical treatment
Introduction The treatment of spinal tuberculosis (TB) has always been controversial. With the advent of refined surgical FDA device/drug status: approved for this application (INH, Rifampacin, Ethambutol). Nothing of value received from a commercial entity related to this research. * Corresponding author. P.D. Hinduja National Hospital and Medical Research Centre, Veer Savarkar Road, Mahim, Mumbai 400 016, India. Tel.: 91-22-24447427/24451515/24309207; fax: 91-22-24440425. E-mail address: [email protected], [email protected] (S. Bhojraj) 1529-9430/05/$ – see front matter doi:10.1016/j.spinee.2004.05.255
쑖 2005 Elsevier Inc. All rights reserved.
techniques and modern spinal instrumentation, an increasing number of surgeries are being performed for spinal TB. Without a clear-cut protocol, surgery is being performed for a range of indications, including abscess drainage, neurological deficit, biopsy, drug resistance and instability [1–3]. Fundamentally, the treatment of tuberculosis is by chemotherapy, and surgery attempts only to extirpate the complications arising from the disease process. Hence, all uncomplicated cases of spinal TB merit medical rather than surgical treatment. On primary evaluation, “acceptable” fallouts of nonsurgical treatment should be weighed against the risks and other drawbacks of surgery.
80
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
Over the years, while dealing with a fairly high load of spinal TB, we have found that meticulously applied nonsurgical treatment is capable of much better results than is commonly believed or documented, even in controversial situations. This belief is strengthened by a clearer understanding of drug-resistant TB, and the advent of secondaryline antituberculous chemotherapeutic agents. We present our series of 70 consecutive cases of thoracic spinal tuberculosis, which fit into our inclusion criteria for nonoperative treatment, with varying presentations, including abscesses and neurological deficits, seen at our spine clinic, in a period between August 1998 and August 2000. Our results of nonsurgical treatment in these patients were rewarding.
Materials and methods A retrospective study was made of 70 consecutive adult patients with thoracic spinal tuberculosis, who fit into the inclusion/exclusion criteria, presenting at our spine clinic, between August 1998 and August 2000. The inclusion criterion was strong clinicoradiological suspicion of spinal tuberculosis. Exclusion criteria were advanced motor deficit at the time of presentation (motor power less than Medical Research Council [MRC] Grade 3 by 5), deteriorating neurologic function while on antituberculous chemotherapy, prepubertal children and kyphosis greater than 40 degrees at the affected level at the time of presentation. There were 43 females and 27 males. Their ages ranged from 14 to 68 years, with a mean age of 37 years. Patients who had not completed their pubertal growth spurts were excluded from the study. There were no cases of reactivation of an old tuberculous lesion. The follow-up period ranged from 18 to 73 months, the average being 40 months. Of the 70 patients studied, 11 had thoracolumbar junctional lesions, and the remaining 59 had purely thoracic spinal lesions. The average number of levels involved was 2.5 (range, 1 to 5 levels). T11–T12 was the most commonly affected area, and T1 was the least commonly affected level in our series. Ten patients had skip lesions in noncongruous areas of the spine. Two patients had associated pulmonary tuberculosis, and no other patient had tuberculous involvement of any other system. None of our patients were HIV positive. All patients were assessed clinically for signs of instability, kyphosis, radiculopathy and sensorimotor deficits at each follow-up. Eleven patients had demonstrable neurological deficit at the time of presentation, 7 of whom had upper motor signs of hypertonia and hyper reflexia in the lower limbs, with extensor plantars, and 4 of whom had lower motor type parapariesis, with a motor power of MRC Grade 3 by 5,
in both the lower limbs, with or without sensory deficits (Fig. 1). Four patients presented with radicular girdle pains, with matching root pressure resulting from tuberculous granulation tissue seen on magnetic resonance imaging (MRI) scans. Instability was judged clinically, using the clinical criteria of severe back pain, paraspinal muscle spasm, painfully restricted movements of the thoracolumbar spine and an “instability catch” [4–6]. Nine patients of our series had these clinical signs of instability at the time of first presentation. Seven of the nine patients with clinical instability had thoracolumbar junctional lesions (Fig. 2). In all of the cases there was a strong clinical suspicion of spinal tuberculosis based on history and clinical symptoms and signs [7–9]. In all cases the diagnosis was backed up by radiological studies. Routine laboratory studies, including a complete blood count (CBC) and an erythrocyte sedimentation rate (ESR by Westergren’s method), were done in all patients along with a plain roentgenogram of the chest. A high ESR, although not specific, was considered as corroborative evidence of TB. No other hematological parameters were considered to have diagnostic or prognostic value. On plain roentgenograms, features varying from osteopenia, paravertebral abscess, disc space reduction and end plate erosions to gross bony vertebral and costovertebral destruction and deformities were identified [10–13]. On MRI scans, 44 patients had abscesses, 21 epidural and the rest prevertebral or paravertebral. Of these, 20 were larger than 2.5 cm in any one diameter. CT-guided biopsies were performed in 11 cases. Five showed positive histopathology for TB, and three yielded a positive culture.
Fig. 1. Pre- and posttreatment magnetic resonance imaging (MRI) scans of a 40-year-old man presenting with tuberculous parapariesis. At presentation, the patient had Medical Research Council Grade 3 by 5 power in both lower limbs, with significant vertebral body destruction and cord compression by tuberculous granulation tissue on MRI. He showed excellent clinical and radiological response to nonsurgical treatment, with return to normal neurology, and radiological evidence of bony fusion without significant deformity by the end of 12 months of chemotherapy alone.
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
Fig. 2. Magnetic resonance imaging of a 45-year-old woman who presented with clinical instability and upper motor type of neurological deficit in the lower limbs. She recovered completely with nonoperative treatment, clinically as well as radiologically, by the end of 12 months.
Fifty of the 70 patients had the classical spondylodiscitic lesions, with affection of apposing end plates of contiguous vertebral bodies and the intervening disc, typical of tuberculosis. Seven had adjacent vertebral body involvement with sparing of the intervening disc. Seven patients had osteitis with no collapse or bony destruction, and four had central body lesions on MRI scan. Two had a miscellaneous presentation, including posterior element TB and pedicular lesions. Plain roentgenograms and MRI scans were performed before commencing treatment and were repeated within 3 and 12 months of starting treatment. CT scans with intravenous nonionic contrast were performed at the same intervals, instead of MRI scan, in the one case that was eventually operated, using MRI-incompatible implants. Conservative treatment was administered by the following protocol: 1. The standard four-drug therapy of isoniazid (INH) (5 mg/kg), rifampicin (10 mg/kg), ethambutol (15 mg/ kg) and pyrazinamide (25 mg/kg) was administered as a first-line treatment, continued for a total of 4 months, followed by three-drug antituberculous treatment (rifampicin/INH/ethambutol) for 8 months, or until regression of symptoms with laboratory and radiological signs of resolution, whichever was later [3,14,15]. Intramuscular streptomycin (30 mg/kg) may be used as a first-line drug, in place of ethambutol. Our chest physicians regularly supervised this treatment. 2. Complete blood counts, erythrocyte sedimentation rate, liver function test and serum uric acid levels were monitored every 6 weeks for the first 3 months, and every 3 months after that until the end of treatment, to judge the efficacy of treatment, and to monitor side effects, if any.
81
3. Plain roentgenograms and MRI scans/CT scans, as mentioned previously, were performed at 0, 3 and 12 months, to monitor response to treatment. 4. Patients who were ambulatory at the time of diagnosis were kept ambulatory during treatment, with restriction of heavy, load-bearing, contact activities. Patients who were bedridden at the time of diagnosis were kept in bed until clinical recovery (pain relief and neurological improvement) and then mobilized and ambulated with similar precautions as the ambulatory group, that is, with restrictions in heavy activity [16]. Ambulation was permitted in these patients at an average of 3 to 4 weeks, and return to activities of daily living, light work and noncontact activity was possible at 6 to 12 weeks, depending on the clinical response to treatment. 5. Bracing was not used in any of our patients, because we find that the compliance for a total contact plastic molded orthosis is extremely poor in our tropical climate. 6. If there was no clinical/laboratory response after 6 to 12 weeks of treatment, the patients were presumed to be clinically resistant to first-line antituberculous drugs, and additional drugs were started, as per a fixed protocol. Second-line drugs for drug-resistant tuberculosis included Kanamycin, Ciprofloxacin, Sparfloxacin, Ethionamide and so forth. 7. In case of drug intolerance, drug therapy was altered by our chest physicians.
Results After starting conservative therapy, 56 of the 70 patients showed definite clinical response within 3 weeks, which continued and was confirmed radiologically at 12 weeks of starting first-line antituberculous therapy. Thirteen of the remaining 14 patients remained unchanged clinicoradiologically after 3 months and were treated with a modification in the drug regimen, as previously mentioned. All of these showed a good response to secondary-line drugs. Thus, a total of 69 of the 70 patients responded fully to nonoperative treatment, either first line (56 cases) or second line (13 cases). The duration of antituberculous treatment ranged from 12 months to 27 months, until regression of symptoms and laboratory and radiological signs of resolution [14,15]. Clinical improvement was in the form of marked reduction in pain, tenderness, paraspinal muscular spasm and increase in the level of physical activity [16]. All the 69 conservatively treated patients eventually regained their predisease activity status and showed excellent spinal function, with no signs of residual instability or sensorimotor deficits. Radiological signs of recovery on MRI were regression in the size of the abscess and granulation tissue, total marrow
82
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
reconversion and, finally, fatty reconstitution of the diseased bone marrow [11,12]. Radiological resolution was complete in all of the 69 cases that responded to chemotherapy alone. The sagittal balance of the thoracic spine was measured as the angle of kyphosis, on the MRI plates, by the Cobb method. The Cobb angle before and after treatment was compared [17,18]. The kyphosis angles, both the angle at the lesion as well as the overall kyphosis in the thoracic spine, between T1 and T12 vertebrae were considered. Angles were calculated at the start of treatment and at last follow-up (minimum 12 months) At the start of treatment, 16 patients were found to have an apical Cobb angle greater than 10 degrees (all were less than 20 degrees). Of these 16 patients, 12 remained unchanged at the end of treatment. 8 of these were followed up for over 2 years after disease resolution, and none of these had any increase in their kyphosis angles. The average progression of kyphosis at the end of treatment was less than 5 degrees at the disease site and less than 2 degrees of overall thoracic kyphosis (given the compensation above or below the apex) Four of the 16 patients had progression of kyphosis to greater than 20 degrees while on treatment. Only one of these progressed to beyond the acceptable limit of 40 degrees at the end of treatment [17,18] (45 degrees kyphosis at end of treatment, unchanged at 2-year follow-up) (Fig. 3). None of these 16 patients had any symptoms related to their kyphosis. Two were bothered by cosmesis but were uninterested in any surgical intervention for the same. One patient from our series failed the trial of nonsurgical treatment. This 52-year-old woman showed neurological deterioration at 3 weeks of drug therapy and required surgical decompression and stabilization. She recovered well after surgery, with a complete neurological recovery within 3 weeks of surgery. Her tissue biopsy was positive for tuberculosis, and the culture did not grow resistant strains. This case was considered a poor result of conservative treatment.
While grading our results, we took into consideration the following criteria: 1. Clinical healing, that is, total reversal of all pretreatment symptoms. 2. Radiological healing, that is, complete regression of pathological soft tissue on MRI scan. 3. Primary-line four-drug anti-TB medications, used for 12 months. 4. More than four drugs used, for longer than 12 months. 5. Treatable medical side effects of medications while on treatment. 6. Kyphosis over 10 degrees at affected level at the end of treatment. 7. Failure of nonoperative treatment, requiring surgical intervention. We classified our results into four groups: Excellent result: complete resolution of disease with firstline antituberculous treatment for 12 months, with no residues or side effects of treatment (Criteria 1⫹2⫹3). Good result: complete resolution of disease with firstline treatment for longer than 12 months, or with secondline drugs for 12 months or longer, with or without the occurrence of medically manageable side effects of drug therapy, such as gastrointestinal intolerance, drug-induced jaundice and so forth, necessitating frequent manipulations in their doses and regimen, but with eventual complete resolution (Criteria 1⫹2⫹4/5). Fair result: complete resolution of disease, but with a posttreatment kyphosis of 10 degrees or more at the affected level, which was clinically obvious, but mechanically irrelevant (Criteria 1⫹2⫹3/4/5⫹6). Poor result: Cases that did not respond to conservative treatment, primary or secondary line, and had to be eventually operated (Criteria 7). The results of our series of 70 cases, which were given a conservative trial of treatment for thoracic spinal tuberculosis, are shown in Table 1. Thus, excellent to good results were seen in over 75% of our patients, with nonsurgical treatment. The 23% with fair results also resumed routine activities of work and daily living without physical restrictions or need for regular analgesic medications, despite residual kyphosis. Table 1 Results of 70 nonsurgically treated cases of thoracic and thoracolumbar tuberculosis
Fig. 3. Pre- and posttreatment magnetic resonance imaging scans of a 28-year-old woman with mid-thoracic spinal tuberculosis. Pretreatment kyphosis was 25 degrees. She was offered nonsurgical treatment despite the kyphosis, which progressed to 45 degrees at the end of treatment (12 months). Two years after completion of treatment, she showed no progression of her kyphosis angle.
Result
Number of cases (total⫽70)
Approximate percentage (total⫽100%)
Excellent Good Fair Poor
34 19 16 1
48.5% 27% 23% 1.5%
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
83
Of the 13 patients in this series who did not respond favorably to standard antituberculous drugs (but eventually healed totally with addition of second-line drugs), 3 had a positive culture for drug-resistant tuberculosis on CT-guided biopsy. The other 10 were labeled as clinically drug-resistant tuberculosis. All 13 required treatment with six or more drugs (including ciprofloxacin, ofloxacin, sparfloxacin, kanamycin, ethionamide and para-amino salicylate) before they eventually healed.
Discussion The resurgence of spinal tuberculosis in the western world has sparked a flurry of activity toward the prevention and treatment of this condition. Standardization of protocols for management of spinal TB becomes important to optimize health resources and expenses. We seem to have come in a full circle as far as surgical indications in the treatment of spinal TB are concerned: from conservatism resulting from lack of appropriate techniques, through aggressive surgical debridement and reconstruction for “disease control” and drug resistance, with the availability of modern spinal instrumentation, and we are now relearning nonoperative care, with a better understanding of drug-resistant TB and the natural history of spinal TB. Current literature seems to uniformly echo the fact that spinal TB is a “medical condition” unless accompanied by “surgical” complications, such as neurological deficits, kyphosis and abscess [14,19]. However, the interpretation of these “surgical complications” is not uniform. The threshold for surgery seems to be low while dealing with patients with early upper motor signs and abscesses and kyphosis of any proportion [1–3]. Our results clearly show that surgery can be circumvented in all these conditions and highlight the fact that long-term excellent to good results can be obtained by supervised and appropriate nonsurgical treatment in adult thoracic spinal TB. In adults, progression of kyphosis is not a routine as it would be in a growing spine. It is extremely rare to see biomechanically unstable spines resulting from post-TB kyphosis in adults. The results of our study are illustrative of this fact. The issue of late onset paraplegia resulting from longstanding kyphosis also remains more relevant in childhood TB [20–22]. Regarding concern for cosmesis, patients should critically weigh the aesthetic aspect of a visible deformity against a scar of surgery before opting for treatment. Abscesses resolve completely with medications, as has been observed in our series (Fig. 4). Forty-four patients from our series had paraspinal abscesses of variable types, and all of these resolved completely with chemotherapy alone. We reserve surgical drainage for those abscesses, which cause mechanical pressure–related symptoms because of their size and location [19,22].
Fig. 4. Illustration exemplifying complete resolution of a large tuberculous paravertebral abscess without surgical drainage, at the end of 12 months of chemotherapy.
As far as neurological deficits in thoracic spinal TB are concerned, without clear-cut guidelines several authors have recommended that surgery be performed for any neurological deficit in spinal tuberculosis. Nussbaum et al. [3] pointed out that “No patient with neurological deficit recovered or stabilized with non-operative management.” Moon [14], on the other hand, recommended that only advanced neurological deficit, such as paraplegia, is a surgical indication in this disease. Clearly, it is not easy to indicate a cutoff point for the neurological status, to advocate nonsurgical treatment. We recommend that all patients without major lower limb weakness, motor power of grade greater than or equal to Grade 3 by 5 by the MRC system, be given a trial of medical treatment under close observation [22,23]. Presentations with paraplegia, gross motor weakness (MRC Grade 2 by 5 or less motor power) or neurologic deterioration while on treatment under observation would remain absolute indications for surgical intervention. In our series, 11 patients presented with varying motor deficits, and all these recovered completely with medical treatment alone, without need for surgical decompression. The issue of drug resistance has to be kept in mind while treating spinal TB. Drug resistance has been quoted to be as high as 15% to 20% in extraspinal TB, and similar figures seem to be applicable to spinal TB also [24]. Tissue diagnosis for culture and drug sensitivity, apart from diagnosis, thus gains importance. We recommend that this be attempted by image-guided needle biopsies. However, in the event of unconvincing results of these biopsies, we continue to treat the patient in accordance with our protocol, based on clinicoradiological grounds [19]. Appropriate drug therapy, keeping in mind multi-drugresistant TB, remains the cornerstone of treatment of thoracic spinal TB. Analyzing assimilated data from relevant centers all over the world would be the first step in redefining surgical indications for the treatment of spinal TB. This article aims to highlight the excellent results possible by nonsurgical treatment of thoracic spinal TB.
84
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
References [1] J Boachie-Adjei O, Squillante RG. Tuberculosis of the spine. Orthop Clin North Am 1996;27(1):95–103. [2] Pun WK, Chow SP, Luk KD, Cheng CL, Hsu LC, Leong JC. Tuberculosis of the lumbosacral junction. Long-term follow-up of 26 cases. J Bone Joint Surg Br 1990;72(4):675–8. [3] Nussbaum ES, Rockswold GL, Bergman TA, Erickson DL, Seljeskog EL. Spinal tuberculosis: a diagnostic and management challenge. J Neurosurg 1995;83(2):243–7. [4] Lee TC, Lu K, Yang LC, Huang HY, Liang CL. Transpedicular instrumentation as an adjunct in the treatment of thoracolumbar and lumbar spine tuberculosis with early stage bone destruction. J Neurosurg Spine 1999;91:163–9. [5] Nachemson A. The role of spine fusion. Spine 1981;6:306–7. [6] White AA, Punjabi MM. Clinical biomechanics of the spine, 2nd ed. Philadelphia: J.B. Lippincott, 1990. [7] Mehta JS, Bhojraj SY. Tuberculosis of the thoracic spine. A classification based on the selection of surgical strategies. J Bone Joint Surg 2001;83(B6):859–63. [8] Davies PDO. Clinical tuberculosis, 2nd ed. London; New York: Chapman and Hall Medical, 1998. [9] Jain AK, Kumar S, Tuli SM. Tuberculosis of spine (C1 to D4). Spinal Cord 1999;37(5):362–9. [10] Desai S. Early diagnosis of spinal tuberculosis by MRI. J Bone Joint Surg [Br] 1994;76-B:863–9. [11] Bell GR, Stearns KL, Bonutti PM, Boumphrey FR. MRI diagnosis of tuberculous vertebral osteomyelitis. Spine 1990;15:462–5. [12] Kim NH, Lee HM, Suh JS. Magnetic resonance imaging for the diagnosis of tuberculous spondylitis. Spine 1994;19(21):2451–5.
[13] Weaver P, Lifeso RM. The radiological diagnosis of tuberculosis of the adult spine. Skeletal Radiol 1984;12:178–86. [14] Moon MS. Tuberculosis of the spine. Controversies and a new challenge. Spine 1997;22(15):1791–7. [15] Pertuiset E. Medical therapy of bone and joint tuberculosis in 1998. Rev Rheum Engl Ed 1999;66(3):152–7. [16] Medical Research Council. A 15-year assessment of controlled trials of the management of tuberculosis of the spine in Korea and Hong Kong. Thirteenth Report of the Medical Research Council Working Party on tuberculosis of the spine. J Bone Joint Surg [Br]1998; 80(3):456–62. [17] Stotts AK, Smith JT, Santora SD, Roach JW, D’Astous L. Measurement of spinal kyphosis: implications for the management of Scheuermann’s kyphosis. Spine 2002;27(19):2143–6. [18] Voutsinas SA, MacEwen GD. Sagittal profiles of the spine. Clin Orthop 1986;210:235–42. [19] Bhojraj S, Nene A. Lumbar and lumbosacral tuberculous spondylodiscitis in adults: redefining indications for surgery. J Bone Joint Surg 2002;84-B(4):530–4. [20] Rajasekaran S. The problem of deformity in spinal tuberculosis. Clin Orthop 2002;398:85–92. [21] Wimmer C, Ogon M, Sterzinger W, Landauer F. Conservative treatment of tuberculous spondylitis: a long-term follow-up study. J Spinal Dis 1997;10(5):417–9. [22] Jain AK. Treatment of tuberculosis of the spine with neurologic complications. Clin Orthop 2002;1(398):75–84. [23] Moon MS, Ha KY, Sun DH, Moon JL, Moon YW, Chung JH. Pott’s paraplegia: 67 cases. Clin Orthop 1996;323:122–8. [24] Drobniewski FA, Yates MD. Multiple drug resistant tuberculosis. J Clin Pathol 1997;50(2):89–90.
Results of nonsurgical treatment of thoracic spinal tuberculosis in adults Abhay Nene, MS (Orth), Shekhar Bhojraj, MS (Orth), FCPS (Orth), D Ortho* The Spine Clinic, P.D. Hinduja National Hospital and Medical Research Centre, Veer Savarkar Road, Mahim, Mumbai 400 016, India Received 15 January 2004; accepted 24 May 2004
Abstract
BACKGROUND: The indications for surgery in spinal tuberculosis have been controversial, and more so recently, in the era of renewed understanding of the concept of multi-drug-resistant tuberculosis along with newer modalities of spinal instrumentation. Indications for surgery need to be redefined in this context. PURPOSE: To assess the efficacy and results of nonsurgical treatment in thoracic spinal tuberculosis in adult patients, and redefine indications for surgery. STUDY DESIGN: We present a retrospective analysis of 70 adults with thoracic spinal tuberculosis, with varying presentations, including abscesses and neurological deficits, seen at our spine clinic, in a period between August 1998 and August 2000, treated largely nonsurgically, with rewarding results. METHODS: A retrospective study was made of 70 adult patients with thoracic spinal tuberculosis presenting at our spine clinic, between August 1998 and August 2000. All patients were subjected to medical management, unless there were specific indications for surgery, as per our protocol, wherein absolute indications of surgery in adults included advanced neurological deficit (less than Grade 3 by 5, by the 5-point grading system of the Medical Research Council), neurology worsening while on antituberculous chemotherapy, diagnosis in doubt on clinicoradiological evaluation and significant kyphosis (greater than 40 degrees) on presentation. Clinical and radiological assessment of results was made by an independent observer, at a mean follow up of 40 months. RESULTS: Forty-four patients presented with abscesses, 21 of which were epidural. Seven had neurological signs of cord compression on clinical examination at presentation. Over 98% of our patients (69 of 70) were successfully treated conservatively, and none of these had any residual instability, radiculopathy or neurological compromise. Seventy-four percent had excellent to good results, with no mechanical residues of the disease, and 23% had residual kyphosis, which was clinically obvious, but biomechanically irrelevant. CONCLUSIONS: We think that tuberculous spondylodiscitis in adults can be well managed conservatively in a vast majority of cases, and indications for surgery are few and specific. 쑖 2005 Elsevier Inc. All rights reserved.
Keywords:
Thoracic spine; Tuberculosis; Nonsurgical treatment
Introduction The treatment of spinal tuberculosis (TB) has always been controversial. With the advent of refined surgical FDA device/drug status: approved for this application (INH, Rifampacin, Ethambutol). Nothing of value received from a commercial entity related to this research. * Corresponding author. P.D. Hinduja National Hospital and Medical Research Centre, Veer Savarkar Road, Mahim, Mumbai 400 016, India. Tel.: 91-22-24447427/24451515/24309207; fax: 91-22-24440425. E-mail address: [email protected], [email protected] (S. Bhojraj) 1529-9430/05/$ – see front matter doi:10.1016/j.spinee.2004.05.255
쑖 2005 Elsevier Inc. All rights reserved.
techniques and modern spinal instrumentation, an increasing number of surgeries are being performed for spinal TB. Without a clear-cut protocol, surgery is being performed for a range of indications, including abscess drainage, neurological deficit, biopsy, drug resistance and instability [1–3]. Fundamentally, the treatment of tuberculosis is by chemotherapy, and surgery attempts only to extirpate the complications arising from the disease process. Hence, all uncomplicated cases of spinal TB merit medical rather than surgical treatment. On primary evaluation, “acceptable” fallouts of nonsurgical treatment should be weighed against the risks and other drawbacks of surgery.
80
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
Over the years, while dealing with a fairly high load of spinal TB, we have found that meticulously applied nonsurgical treatment is capable of much better results than is commonly believed or documented, even in controversial situations. This belief is strengthened by a clearer understanding of drug-resistant TB, and the advent of secondaryline antituberculous chemotherapeutic agents. We present our series of 70 consecutive cases of thoracic spinal tuberculosis, which fit into our inclusion criteria for nonoperative treatment, with varying presentations, including abscesses and neurological deficits, seen at our spine clinic, in a period between August 1998 and August 2000. Our results of nonsurgical treatment in these patients were rewarding.
Materials and methods A retrospective study was made of 70 consecutive adult patients with thoracic spinal tuberculosis, who fit into the inclusion/exclusion criteria, presenting at our spine clinic, between August 1998 and August 2000. The inclusion criterion was strong clinicoradiological suspicion of spinal tuberculosis. Exclusion criteria were advanced motor deficit at the time of presentation (motor power less than Medical Research Council [MRC] Grade 3 by 5), deteriorating neurologic function while on antituberculous chemotherapy, prepubertal children and kyphosis greater than 40 degrees at the affected level at the time of presentation. There were 43 females and 27 males. Their ages ranged from 14 to 68 years, with a mean age of 37 years. Patients who had not completed their pubertal growth spurts were excluded from the study. There were no cases of reactivation of an old tuberculous lesion. The follow-up period ranged from 18 to 73 months, the average being 40 months. Of the 70 patients studied, 11 had thoracolumbar junctional lesions, and the remaining 59 had purely thoracic spinal lesions. The average number of levels involved was 2.5 (range, 1 to 5 levels). T11–T12 was the most commonly affected area, and T1 was the least commonly affected level in our series. Ten patients had skip lesions in noncongruous areas of the spine. Two patients had associated pulmonary tuberculosis, and no other patient had tuberculous involvement of any other system. None of our patients were HIV positive. All patients were assessed clinically for signs of instability, kyphosis, radiculopathy and sensorimotor deficits at each follow-up. Eleven patients had demonstrable neurological deficit at the time of presentation, 7 of whom had upper motor signs of hypertonia and hyper reflexia in the lower limbs, with extensor plantars, and 4 of whom had lower motor type parapariesis, with a motor power of MRC Grade 3 by 5,
in both the lower limbs, with or without sensory deficits (Fig. 1). Four patients presented with radicular girdle pains, with matching root pressure resulting from tuberculous granulation tissue seen on magnetic resonance imaging (MRI) scans. Instability was judged clinically, using the clinical criteria of severe back pain, paraspinal muscle spasm, painfully restricted movements of the thoracolumbar spine and an “instability catch” [4–6]. Nine patients of our series had these clinical signs of instability at the time of first presentation. Seven of the nine patients with clinical instability had thoracolumbar junctional lesions (Fig. 2). In all of the cases there was a strong clinical suspicion of spinal tuberculosis based on history and clinical symptoms and signs [7–9]. In all cases the diagnosis was backed up by radiological studies. Routine laboratory studies, including a complete blood count (CBC) and an erythrocyte sedimentation rate (ESR by Westergren’s method), were done in all patients along with a plain roentgenogram of the chest. A high ESR, although not specific, was considered as corroborative evidence of TB. No other hematological parameters were considered to have diagnostic or prognostic value. On plain roentgenograms, features varying from osteopenia, paravertebral abscess, disc space reduction and end plate erosions to gross bony vertebral and costovertebral destruction and deformities were identified [10–13]. On MRI scans, 44 patients had abscesses, 21 epidural and the rest prevertebral or paravertebral. Of these, 20 were larger than 2.5 cm in any one diameter. CT-guided biopsies were performed in 11 cases. Five showed positive histopathology for TB, and three yielded a positive culture.
Fig. 1. Pre- and posttreatment magnetic resonance imaging (MRI) scans of a 40-year-old man presenting with tuberculous parapariesis. At presentation, the patient had Medical Research Council Grade 3 by 5 power in both lower limbs, with significant vertebral body destruction and cord compression by tuberculous granulation tissue on MRI. He showed excellent clinical and radiological response to nonsurgical treatment, with return to normal neurology, and radiological evidence of bony fusion without significant deformity by the end of 12 months of chemotherapy alone.
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
Fig. 2. Magnetic resonance imaging of a 45-year-old woman who presented with clinical instability and upper motor type of neurological deficit in the lower limbs. She recovered completely with nonoperative treatment, clinically as well as radiologically, by the end of 12 months.
Fifty of the 70 patients had the classical spondylodiscitic lesions, with affection of apposing end plates of contiguous vertebral bodies and the intervening disc, typical of tuberculosis. Seven had adjacent vertebral body involvement with sparing of the intervening disc. Seven patients had osteitis with no collapse or bony destruction, and four had central body lesions on MRI scan. Two had a miscellaneous presentation, including posterior element TB and pedicular lesions. Plain roentgenograms and MRI scans were performed before commencing treatment and were repeated within 3 and 12 months of starting treatment. CT scans with intravenous nonionic contrast were performed at the same intervals, instead of MRI scan, in the one case that was eventually operated, using MRI-incompatible implants. Conservative treatment was administered by the following protocol: 1. The standard four-drug therapy of isoniazid (INH) (5 mg/kg), rifampicin (10 mg/kg), ethambutol (15 mg/ kg) and pyrazinamide (25 mg/kg) was administered as a first-line treatment, continued for a total of 4 months, followed by three-drug antituberculous treatment (rifampicin/INH/ethambutol) for 8 months, or until regression of symptoms with laboratory and radiological signs of resolution, whichever was later [3,14,15]. Intramuscular streptomycin (30 mg/kg) may be used as a first-line drug, in place of ethambutol. Our chest physicians regularly supervised this treatment. 2. Complete blood counts, erythrocyte sedimentation rate, liver function test and serum uric acid levels were monitored every 6 weeks for the first 3 months, and every 3 months after that until the end of treatment, to judge the efficacy of treatment, and to monitor side effects, if any.
81
3. Plain roentgenograms and MRI scans/CT scans, as mentioned previously, were performed at 0, 3 and 12 months, to monitor response to treatment. 4. Patients who were ambulatory at the time of diagnosis were kept ambulatory during treatment, with restriction of heavy, load-bearing, contact activities. Patients who were bedridden at the time of diagnosis were kept in bed until clinical recovery (pain relief and neurological improvement) and then mobilized and ambulated with similar precautions as the ambulatory group, that is, with restrictions in heavy activity [16]. Ambulation was permitted in these patients at an average of 3 to 4 weeks, and return to activities of daily living, light work and noncontact activity was possible at 6 to 12 weeks, depending on the clinical response to treatment. 5. Bracing was not used in any of our patients, because we find that the compliance for a total contact plastic molded orthosis is extremely poor in our tropical climate. 6. If there was no clinical/laboratory response after 6 to 12 weeks of treatment, the patients were presumed to be clinically resistant to first-line antituberculous drugs, and additional drugs were started, as per a fixed protocol. Second-line drugs for drug-resistant tuberculosis included Kanamycin, Ciprofloxacin, Sparfloxacin, Ethionamide and so forth. 7. In case of drug intolerance, drug therapy was altered by our chest physicians.
Results After starting conservative therapy, 56 of the 70 patients showed definite clinical response within 3 weeks, which continued and was confirmed radiologically at 12 weeks of starting first-line antituberculous therapy. Thirteen of the remaining 14 patients remained unchanged clinicoradiologically after 3 months and were treated with a modification in the drug regimen, as previously mentioned. All of these showed a good response to secondary-line drugs. Thus, a total of 69 of the 70 patients responded fully to nonoperative treatment, either first line (56 cases) or second line (13 cases). The duration of antituberculous treatment ranged from 12 months to 27 months, until regression of symptoms and laboratory and radiological signs of resolution [14,15]. Clinical improvement was in the form of marked reduction in pain, tenderness, paraspinal muscular spasm and increase in the level of physical activity [16]. All the 69 conservatively treated patients eventually regained their predisease activity status and showed excellent spinal function, with no signs of residual instability or sensorimotor deficits. Radiological signs of recovery on MRI were regression in the size of the abscess and granulation tissue, total marrow
82
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
reconversion and, finally, fatty reconstitution of the diseased bone marrow [11,12]. Radiological resolution was complete in all of the 69 cases that responded to chemotherapy alone. The sagittal balance of the thoracic spine was measured as the angle of kyphosis, on the MRI plates, by the Cobb method. The Cobb angle before and after treatment was compared [17,18]. The kyphosis angles, both the angle at the lesion as well as the overall kyphosis in the thoracic spine, between T1 and T12 vertebrae were considered. Angles were calculated at the start of treatment and at last follow-up (minimum 12 months) At the start of treatment, 16 patients were found to have an apical Cobb angle greater than 10 degrees (all were less than 20 degrees). Of these 16 patients, 12 remained unchanged at the end of treatment. 8 of these were followed up for over 2 years after disease resolution, and none of these had any increase in their kyphosis angles. The average progression of kyphosis at the end of treatment was less than 5 degrees at the disease site and less than 2 degrees of overall thoracic kyphosis (given the compensation above or below the apex) Four of the 16 patients had progression of kyphosis to greater than 20 degrees while on treatment. Only one of these progressed to beyond the acceptable limit of 40 degrees at the end of treatment [17,18] (45 degrees kyphosis at end of treatment, unchanged at 2-year follow-up) (Fig. 3). None of these 16 patients had any symptoms related to their kyphosis. Two were bothered by cosmesis but were uninterested in any surgical intervention for the same. One patient from our series failed the trial of nonsurgical treatment. This 52-year-old woman showed neurological deterioration at 3 weeks of drug therapy and required surgical decompression and stabilization. She recovered well after surgery, with a complete neurological recovery within 3 weeks of surgery. Her tissue biopsy was positive for tuberculosis, and the culture did not grow resistant strains. This case was considered a poor result of conservative treatment.
While grading our results, we took into consideration the following criteria: 1. Clinical healing, that is, total reversal of all pretreatment symptoms. 2. Radiological healing, that is, complete regression of pathological soft tissue on MRI scan. 3. Primary-line four-drug anti-TB medications, used for 12 months. 4. More than four drugs used, for longer than 12 months. 5. Treatable medical side effects of medications while on treatment. 6. Kyphosis over 10 degrees at affected level at the end of treatment. 7. Failure of nonoperative treatment, requiring surgical intervention. We classified our results into four groups: Excellent result: complete resolution of disease with firstline antituberculous treatment for 12 months, with no residues or side effects of treatment (Criteria 1⫹2⫹3). Good result: complete resolution of disease with firstline treatment for longer than 12 months, or with secondline drugs for 12 months or longer, with or without the occurrence of medically manageable side effects of drug therapy, such as gastrointestinal intolerance, drug-induced jaundice and so forth, necessitating frequent manipulations in their doses and regimen, but with eventual complete resolution (Criteria 1⫹2⫹4/5). Fair result: complete resolution of disease, but with a posttreatment kyphosis of 10 degrees or more at the affected level, which was clinically obvious, but mechanically irrelevant (Criteria 1⫹2⫹3/4/5⫹6). Poor result: Cases that did not respond to conservative treatment, primary or secondary line, and had to be eventually operated (Criteria 7). The results of our series of 70 cases, which were given a conservative trial of treatment for thoracic spinal tuberculosis, are shown in Table 1. Thus, excellent to good results were seen in over 75% of our patients, with nonsurgical treatment. The 23% with fair results also resumed routine activities of work and daily living without physical restrictions or need for regular analgesic medications, despite residual kyphosis. Table 1 Results of 70 nonsurgically treated cases of thoracic and thoracolumbar tuberculosis
Fig. 3. Pre- and posttreatment magnetic resonance imaging scans of a 28-year-old woman with mid-thoracic spinal tuberculosis. Pretreatment kyphosis was 25 degrees. She was offered nonsurgical treatment despite the kyphosis, which progressed to 45 degrees at the end of treatment (12 months). Two years after completion of treatment, she showed no progression of her kyphosis angle.
Result
Number of cases (total⫽70)
Approximate percentage (total⫽100%)
Excellent Good Fair Poor
34 19 16 1
48.5% 27% 23% 1.5%
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
83
Of the 13 patients in this series who did not respond favorably to standard antituberculous drugs (but eventually healed totally with addition of second-line drugs), 3 had a positive culture for drug-resistant tuberculosis on CT-guided biopsy. The other 10 were labeled as clinically drug-resistant tuberculosis. All 13 required treatment with six or more drugs (including ciprofloxacin, ofloxacin, sparfloxacin, kanamycin, ethionamide and para-amino salicylate) before they eventually healed.
Discussion The resurgence of spinal tuberculosis in the western world has sparked a flurry of activity toward the prevention and treatment of this condition. Standardization of protocols for management of spinal TB becomes important to optimize health resources and expenses. We seem to have come in a full circle as far as surgical indications in the treatment of spinal TB are concerned: from conservatism resulting from lack of appropriate techniques, through aggressive surgical debridement and reconstruction for “disease control” and drug resistance, with the availability of modern spinal instrumentation, and we are now relearning nonoperative care, with a better understanding of drug-resistant TB and the natural history of spinal TB. Current literature seems to uniformly echo the fact that spinal TB is a “medical condition” unless accompanied by “surgical” complications, such as neurological deficits, kyphosis and abscess [14,19]. However, the interpretation of these “surgical complications” is not uniform. The threshold for surgery seems to be low while dealing with patients with early upper motor signs and abscesses and kyphosis of any proportion [1–3]. Our results clearly show that surgery can be circumvented in all these conditions and highlight the fact that long-term excellent to good results can be obtained by supervised and appropriate nonsurgical treatment in adult thoracic spinal TB. In adults, progression of kyphosis is not a routine as it would be in a growing spine. It is extremely rare to see biomechanically unstable spines resulting from post-TB kyphosis in adults. The results of our study are illustrative of this fact. The issue of late onset paraplegia resulting from longstanding kyphosis also remains more relevant in childhood TB [20–22]. Regarding concern for cosmesis, patients should critically weigh the aesthetic aspect of a visible deformity against a scar of surgery before opting for treatment. Abscesses resolve completely with medications, as has been observed in our series (Fig. 4). Forty-four patients from our series had paraspinal abscesses of variable types, and all of these resolved completely with chemotherapy alone. We reserve surgical drainage for those abscesses, which cause mechanical pressure–related symptoms because of their size and location [19,22].
Fig. 4. Illustration exemplifying complete resolution of a large tuberculous paravertebral abscess without surgical drainage, at the end of 12 months of chemotherapy.
As far as neurological deficits in thoracic spinal TB are concerned, without clear-cut guidelines several authors have recommended that surgery be performed for any neurological deficit in spinal tuberculosis. Nussbaum et al. [3] pointed out that “No patient with neurological deficit recovered or stabilized with non-operative management.” Moon [14], on the other hand, recommended that only advanced neurological deficit, such as paraplegia, is a surgical indication in this disease. Clearly, it is not easy to indicate a cutoff point for the neurological status, to advocate nonsurgical treatment. We recommend that all patients without major lower limb weakness, motor power of grade greater than or equal to Grade 3 by 5 by the MRC system, be given a trial of medical treatment under close observation [22,23]. Presentations with paraplegia, gross motor weakness (MRC Grade 2 by 5 or less motor power) or neurologic deterioration while on treatment under observation would remain absolute indications for surgical intervention. In our series, 11 patients presented with varying motor deficits, and all these recovered completely with medical treatment alone, without need for surgical decompression. The issue of drug resistance has to be kept in mind while treating spinal TB. Drug resistance has been quoted to be as high as 15% to 20% in extraspinal TB, and similar figures seem to be applicable to spinal TB also [24]. Tissue diagnosis for culture and drug sensitivity, apart from diagnosis, thus gains importance. We recommend that this be attempted by image-guided needle biopsies. However, in the event of unconvincing results of these biopsies, we continue to treat the patient in accordance with our protocol, based on clinicoradiological grounds [19]. Appropriate drug therapy, keeping in mind multi-drugresistant TB, remains the cornerstone of treatment of thoracic spinal TB. Analyzing assimilated data from relevant centers all over the world would be the first step in redefining surgical indications for the treatment of spinal TB. This article aims to highlight the excellent results possible by nonsurgical treatment of thoracic spinal TB.
84
A. Nene and S. Bhojraj / The Spine Journal 5 (2005) 79–84
References [1] J Boachie-Adjei O, Squillante RG. Tuberculosis of the spine. Orthop Clin North Am 1996;27(1):95–103. [2] Pun WK, Chow SP, Luk KD, Cheng CL, Hsu LC, Leong JC. Tuberculosis of the lumbosacral junction. Long-term follow-up of 26 cases. J Bone Joint Surg Br 1990;72(4):675–8. [3] Nussbaum ES, Rockswold GL, Bergman TA, Erickson DL, Seljeskog EL. Spinal tuberculosis: a diagnostic and management challenge. J Neurosurg 1995;83(2):243–7. [4] Lee TC, Lu K, Yang LC, Huang HY, Liang CL. Transpedicular instrumentation as an adjunct in the treatment of thoracolumbar and lumbar spine tuberculosis with early stage bone destruction. J Neurosurg Spine 1999;91:163–9. [5] Nachemson A. The role of spine fusion. Spine 1981;6:306–7. [6] White AA, Punjabi MM. Clinical biomechanics of the spine, 2nd ed. Philadelphia: J.B. Lippincott, 1990. [7] Mehta JS, Bhojraj SY. Tuberculosis of the thoracic spine. A classification based on the selection of surgical strategies. J Bone Joint Surg 2001;83(B6):859–63. [8] Davies PDO. Clinical tuberculosis, 2nd ed. London; New York: Chapman and Hall Medical, 1998. [9] Jain AK, Kumar S, Tuli SM. Tuberculosis of spine (C1 to D4). Spinal Cord 1999;37(5):362–9. [10] Desai S. Early diagnosis of spinal tuberculosis by MRI. J Bone Joint Surg [Br] 1994;76-B:863–9. [11] Bell GR, Stearns KL, Bonutti PM, Boumphrey FR. MRI diagnosis of tuberculous vertebral osteomyelitis. Spine 1990;15:462–5. [12] Kim NH, Lee HM, Suh JS. Magnetic resonance imaging for the diagnosis of tuberculous spondylitis. Spine 1994;19(21):2451–5.
[13] Weaver P, Lifeso RM. The radiological diagnosis of tuberculosis of the adult spine. Skeletal Radiol 1984;12:178–86. [14] Moon MS. Tuberculosis of the spine. Controversies and a new challenge. Spine 1997;22(15):1791–7. [15] Pertuiset E. Medical therapy of bone and joint tuberculosis in 1998. Rev Rheum Engl Ed 1999;66(3):152–7. [16] Medical Research Council. A 15-year assessment of controlled trials of the management of tuberculosis of the spine in Korea and Hong Kong. Thirteenth Report of the Medical Research Council Working Party on tuberculosis of the spine. J Bone Joint Surg [Br]1998; 80(3):456–62. [17] Stotts AK, Smith JT, Santora SD, Roach JW, D’Astous L. Measurement of spinal kyphosis: implications for the management of Scheuermann’s kyphosis. Spine 2002;27(19):2143–6. [18] Voutsinas SA, MacEwen GD. Sagittal profiles of the spine. Clin Orthop 1986;210:235–42. [19] Bhojraj S, Nene A. Lumbar and lumbosacral tuberculous spondylodiscitis in adults: redefining indications for surgery. J Bone Joint Surg 2002;84-B(4):530–4. [20] Rajasekaran S. The problem of deformity in spinal tuberculosis. Clin Orthop 2002;398:85–92. [21] Wimmer C, Ogon M, Sterzinger W, Landauer F. Conservative treatment of tuberculous spondylitis: a long-term follow-up study. J Spinal Dis 1997;10(5):417–9. [22] Jain AK. Treatment of tuberculosis of the spine with neurologic complications. Clin Orthop 2002;1(398):75–84. [23] Moon MS, Ha KY, Sun DH, Moon JL, Moon YW, Chung JH. Pott’s paraplegia: 67 cases. Clin Orthop 1996;323:122–8. [24] Drobniewski FA, Yates MD. Multiple drug resistant tuberculosis. J Clin Pathol 1997;50(2):89–90.