Treatment of lumbar and lumbosacral spinal tuberculosis with minimally invasive surgery

Abstract

Objective:  To evaluate the outcome of computed tomography (CT)‐guided percutaneous puncture and local chemotherapy for tuberculosis (TB) of the lumbar or lumbosacral spine.

Methods:  From January 2002 to August 2006, 49 patients with lumbar or lumbosacral spinal TB in whom conservative treatment had failed were treated by minimally invasive surgery. There were 21 female and 28 male patients. The mean age of the patients was 47 ± 18 years (range from 12 to 78 years). There were 40 cases with lumbar TB, and 9 with lumbosacral TB. From L₁ to S₁, the involved vertebral numbers were 8, 26, 19, 19, 21 and 10, respectively. In 3 patients the spinal TB involved a single vertebra, in 39 patients two, in 6 patients three and in only one case four vertebrae. According to the Frankel scale, three patients had grade E incomplete paralysis preoperatively. The outcome, after treatment by CT‐guided percutaneous puncture and local chemotherapy, was retrospectively analyzed according to the preoperative and postoperative kyphotic angles and the Oswestry disability index (ODI).

Results:  Forty‐nine patients were followed up for from one year to 5 years and 8 months (average 35 months). All cases recovered, and there was no recurrence. Preoperatively, the kyphotic angle was 16.47°± 8.74°, and it had decreased to 13.35°± 8.02° by the final follow‐up (t= 5.79, P < 0.001). The average ODI score had improved from 70.12 ± 22.24 to 12.72 ± 8.62 (t= 21.42, P < 0.001).

Conclusion:  The majority of cases of lumbar and lumbosacral spinal TB can achieve satisfactory results with minimally invasive surgery. Its clinical application is therefore strongly recommended.

Introduction

When lumbosacral spinal tuberculosis (TB) patients show signs of abscesses, cavities, sequestra, and sinus formation, routine treatment by thorough debridement is often recommended. Based on this treatment, bone grafts and/or internal fixation are frequently implemented. As technologies have developed, minimally invasive surgery (MIS) has become an important clinical technique. The General Hospital of PLA has used MIS to treat lumbosacral spinal TB since 2002. Through retrospective analysis of cases during the past five years with a minimum follow up of one year, the clinical value and significance of MIS is here discussed.

Materials and methods

Clinical data

Sixty‐one patients with lumbar or lumbosacral spinal TB in whom conservative treatment had failed were hospitalized from January 2002 to August 2006. Four cases underwent open surgery while 57 were treated by MIS only. Forty‐nine patients with sufficient follow up data were retrospectively studied. Twenty‐eight male and 21 female patients were included, their mean age being 47 ± 18 years (range, 12–78 years). There were 40 cases with lumbar spine TB, and 9 with lumbosacral spine TB. The duration of the disorder ranged from 6 months to 14 years (average, 2.3 years).

One hundred and three vertebrae of 49 patients were affected. From L₁ to S₁, the involved vertebral numbers were 8, 26, 19, 19, 21 and 10, respectively. In 3 patients the spinal TB involved a single vertebra, in 39 patients two, in 6 patients three and in only one case four vertebrae. The mean number of vertebrae involved was 2.1. All cases had evidence of vertebral destruction and narrowed intervertebral spaces. Thirty‐eight cases were complicated by abscesses, 25 showed evidence of cavity formation, and 19 cases had sequestra. The mean erythrocyte sedimentation rate (ESR) was 51 ± 25 mm/h.

Four patients had a history of previous surgery, which consisted simply of recurrent resection of lesions. One patient had undergone surgery once, two patients twice and the remaining one had undergone surgery four times. Three patients had Frankel D incomplete paralysis preoperatively ¹ .

Surgical technique

Preoperative preparation

A definite diagnosis should be made based on the patient's clinical symptoms, ESR, X‐ray, magnetic resonance imaging (MRI) and computed tomography (CT) findings, pathology results, Mycobacterium tuberculosis (M. tuberculosis) culture and clinical outcomes.

Double‐lumen tubes were prepared for lavage (Fig. 1): the bigger tubes were 14 or 16 silica gauge gel catheters, and the smaller tubes were 0.3 mm epidural cannulas. A side hole, 1/3 of the bigger tube's diameter in size, was cut within 5 cm of the top. An epidural cannula with a diameter of 0.3 mm was inserted into the bigger tube to a depth of 25 cm from the top. A double‐lumen tube for lavage was thus assembled. Several spare epidural cannulas 0.3 mm in diameter were also prepared.

Figure 1.

The double‐lumen tube for lavage.

The patients in the group who underwent MIS were those had not responded significantly to pharmacotherapy alone. These failed cases were usually accompanied by definite vertebral destruction, abscesses, cavities, sequestra, sinus formation and many other complications, which prevented them from undergoing open surgery. No preoperative chemotherapy was needed.

Surgical method

Punctures and catheterization of the target lesions were performed in accordance with the findings of the preoperative CT scan and CT guidance during surgery.

1. Placement of the double‐lumen tubes

Placement of epidural tube: The lesion or abscess was punctured using an epidural needle. Having reached the lesion's centre, the inner core was replaced by an epidural tube and then the epidural needle was extracted. The epidural tube was attached to nearby skin using medical tape. The tube with then fitted with a No.7 scalp needle with a tail‐cap to enable future use.

Placement of double lumen: Once puncture had been successfully accomplished, the inner core was replaced by a guide wire, a guiding rod of 5 mm outer diameter with a cone head being inserted under wire guidance. Once the position had been confirmed by CT scan, another cannula of 5 mm inner diameter was inserted through the guide wire. Again, once the position had been confirmed by CT, the guide wire and guiding rod were pulled out, and the prepared double lumen tube inserted. Finally, the cannula was extracted, and the double lumen tube connected with No. 7 scalp needle and drainage bag, respectively. The procedure of tube placement was then complete. The pus obtained before chemotherapy was sent for routine bacterial culture and culture for M. tuberculosis, both with chemosensitivity testing.

2. Treatment of the primary lesion

If the primary lesion was narrow, a simple epidural tube was placed through puncture. If the primary lesion had caused serious damage and formed a large cavity, a double‐lumen tube was put in place as described above.

3. Therapy for any secondary abscess

In the case of a psoas major abscess, an approach was made posteriorly and vertically; with an iliac fossa abscess the approach was through the anterior superior iliac spine. The double‐lumen tube was then put in place.

4. Indications for extraction of the double‐lumen tube

The double‐lumen tube was extracted when the irrigated fluid was clear and no inflammation was apparent near the wounds. When removing the lavage tube, only the bigger tube (the outer one) was pulled out, so that the inner injecting cannula remained in place for future injections of medication over the coming few months, until the body temperature, ESR and C‐reactive protein had been consistently normal for 2–3 months. The drug used for local chemotherapy was 0.1 g isoniazid: once or twice per tube per day.

5. Duration for retaining the epidural tube

The double‐lumen tube was kept in place for 1–3 months, and the injecting cannula for about a further 2 months after extraction of the lavage tube.

6. MIS for any sinus after spinal TB surgery

There were two steps: (i) One or two epidural tubes were inserted into the target lesion under CT guidance; (ii) the same number of double lumen tubes or epidural cannulas was placed within the sinus tract towards the target lesion.

Postoperative treatments

All patients received individualized antiTB treatment with 3 months of isoniazid, rimfampicin, ethambutol and pyrazinamide followed by isoniazid, rimfampicin and ethambutol for as long as required (3HREZ/xHRE). The duration of systemic chemotherapy ranged from 12–24 months. Requirements for bed rest depended on the individual patient's condition during the lavage period. Patients with low back pain were strictly confined to bed. Patients with less pain were permitted to turn freely in bed. Patients with little or no pain were allowed to move to a greater extent and walk with the support of specially designed stabilizing braces.

Patients received systemic chemotherapy while taking local chemotherapy. The plan was 3HREZ/xHRE, where ‘x’ was the consolidation time needed according to each patient's condition, his varying from 12 to 24 months. Antibiotics were used for the first 3–5 days of the lavage period after surgery.

Evaluation standard of curative effect

Measurements of the patient's kyphosis angle between the top and bottom vertebrae were recorded before surgery and at the most recent follow‐up. The interspace between L₅ and S₁ was measured from the top edge of L₅ to the gap between S₁ and S₂. The included angle of the top and bottom vertebrae before and after surgery was measured so that any change in the kyphosis angle of the lumbosacral vertebrae could be determined.

The Oswestry disability index (ODI) was administered to evaluate function before surgery and at the final follow‐up ² .

The clinical remedy standard referred to is the Qi‐qiu Wu method ³ .

Data analysis

All recorded data was processed with the paired t‐test using SPSS 12.0 statistical software (SPSS, Chicago, IL, USA). A significant P‐value was defined as 0.05.

Results

In all 49 patients who received MIS, the lavage tubes were smoothly placed and no intra‐ or post‐operative complications were detected. No sinus tracts or cross infection developed near the wounds. The mean follow‐up time was 35 months (range, 12–68 months). Postoperatively, the patients' low back and leg pain all resolved, and all patients achieved clinical healing. The mean healing span was 4 months. The ODI pre‐ and post‐surgery was 70.12 ± 22.24 and 12.72 ± 8.62, respectively. The difference was statistically significant (t= 21.42, P < 0.001). The mean kyphosis angle before treatment was 16.47°± 8.74°, while the average was 13.35°± 8.02° postoperatively. The difference was statistically significant (t= 5.79, P < 0.001).

Three patients had nerve compression. They were Frankel D level before operation, and recovered to Frankel E level after MIS.

The X‐ray films of one of our patients, a 48‐year old female patient with low back pain radiating into the right leg pain are shown in Figure 2, a–g. An MRI before MIS showed destruction resulting in a narrow intervertebral space at L_4,5, vertebral collapse and a paravertebral abscess (Fig. 2, c–e). X‐ray films after MIS showed the intervertebral space of L_4,5 to be narrowed on an anterior view, and a bone bridge to be connected on the lateral view (Fig. 2, f, g).

Figure 2.

A 48‐year old female patient of with low back pain radiating into the right leg.
(a, b) Anteroposterior and lateral X‐ray films showing that, prior to MIS, the left L₄ vertebrae had collapsed creating a huge intervertebral space at L_4‐5 (c–e) MRI before MIS showing destruction resulting in a intervertebral space at L_4,5, vertebral collapse and a paravertebral abscess. (f, g) X‐ray films after MIS showing the intervertebral space of L_4,5 to be narrowed on an anterior view, and a bone bridge to be connected on the lateral view.

Discussion

The core purpose of MIS is to maximize drug concentration in the lesion when treating lumbar and lumbosacral vertebral TB.

Instead of employing complicated and advanced surgical techniques, MIS allows local application of large doses of chemotherapy drugs (isoniazid) to kill pathogenic bacteria, thus preventing the development of further pathological change. The otherwise problematic question when using drugs to cure spinal TB is how to maximize the amount of drug reaching the lesion. Due to the particular characteristics of local pathological change and frequency of drug‐resistant bacteria, long‐term high doses, drug combinations and daily single dosage drugs have traditionally been applied to overcome this difficulty. MIS is the most direct way to solve this problem.

With MIS it is simple to insert a cannula directly into the target lesion and any secondary foci under CT guidance. This technique does not provide the capability to remove lesions, cavities or sequestra, nor to correct deformity. Then why is it suitable for the treatment of most patients with lumbosacral vertebral TB? The core purpose is to maximize the amount of drug in the lesion, making MIS the most powerful weapon available for killing pathogenic bacterium. Continuous lavage of necrotic wastes, such as those found in abscesses, quickly improves the general condition of adjacent organs, and provides a sound basis for combating the disease. For those patients with sinuses, increasing the amount of drug in the lesion through MIS can greatly improve the chances of recovering without open surgery. One patient with a 14‐year history of lumbosacral TB, who had undergone surgery using an open technique four times and had many sinus tracts, achieved complete recovery after MIS.

Rediscover characteristics of pathological change in spinal TB

Resection of lesions can remove abscesses and necrotic tissue. We can clear foci, and resect sinus tracts to cure relapse which is complicated by sinus formation. However, is it necessary to perform such surgery? Through our clinical research, we believe there is a necessity to reconsider these pathological changes.

Abscess

A simple abscess does not affect spinal stability, and the majority of affected patients have no neurological symptoms, especially when it is lumbosacral vertebrae that are involved ⁴ . Puncture can be performed accurately under CT guidance. Almost all abscesses, from the upper neck to the sacroiliac joint, can be reached and properly treated ⁵ , ⁶ . In addition, the damage done to adjacent organs and requirements for anaesthesia are less than with open surgery, enabling patients to recover more rapidly ⁶ , ⁷ . More than two centuries ago, Pott observed that when sinuses formed and abscesses was drained, spine TB symptoms would be alleviated ⁸ . Therefore, no simple abscess in any part of the spine should considered an indication for an open technique, particularly not when it involves the lumbosacral vertebrae ⁵ .

Necrotic wastes and sequestra

Muzii et al. believe that thorough lesion resection should be used to treat spinal TB ⁷ . This opinion needs careful consideration. The advantages of such a technique are much less than might be anticipated. MIS places no emphasis on clearing lesions, but rather only pays close attention to setting up cannulas to prevent blockage. This method does not propose radical treatment for inflammatory change. The pathology and prospects for future recovery in spinal TB are far less serious than in the case of tumors, therefore the therapy should differ from that for spinal neoplasm. In the absence of neurological lesions, there is no need to completely remove either lesions or the involved vertebrae ⁵ , ⁹ .

Sequestra are not the main factor contributing to recurrence of spinal TB. The surgeon's fear of sequestra comes from a conventional misunderstanding: a belief that sequestra are the primary cause of recurrence. Sequestra are one type of pathological change which occurs in spinal TB, and the crucial source of this pathology is spinal M. tuberculosis organisms which have not been completely eradicated. Through continuous local chemotherapy to kill these bacteria, sequestra become encapsulated by fibrous tissue, then further scarred and ossified, thus becoming incorporated into stable lesions. Sequestra which have not resulted in nerve compression are not an indication for open surgery.

Sclerotic bone around lesions

Sclerotic bone around lesions, bone formed after the patient's condition has stabilized or bone bridges near lesions are not indications for open surgery, no more than are abscesses, sequestra and sinuses. Moreover, it is not recommended that sclerotic bone around lesions and the egg‐shell bone bridges on vertebral surfaces be resected, since this can cause iatrogenic instability.

Nerve involvement

Nerve compression is an absolute indication for open surgery for spinal TB ¹⁰ . Most cases of compression are caused by abscesses or, occasionally, movable sequestra. After simple local percutaneous puncture and abscess drainage, most patients with mild nerve compression recover. Only progressive spinal TB above Frankel C level requires surgical treatment. MIS is to increase the concentration of drug in the lesion. Even though this is ineffective in some cases, it still constitutes good preoperative preparation and will not cause treatment delay. If nerve compression occurs when the lesion is stable, it indicates that the deformed spine is in poor condition. The primary indications for open surgery are to release spinal cord compression, correct spinal deformity and rebuild spinal stability. Outer stabilizing braces and percutaneous bone grafting may also be indicated.

Pain

Back pain is a symptom of spinal instability. Absolute bed rest can help to relieve most patients' pain. Pain and neurological deficit should be considered separately, so that we can avoid applying internal fixation simply because of pain. This principle is especially true for lumbosacral vertebral TB, as this is the region where surgeries are frequently carried out and where complications most often occur. In fact, many instances of open surgery and internal fixation can be avoided. Once a lesion has entered a stable stage, spinal instability caused by non‐fusion of neighboring vertebrae is an indication for internal fixation.

Kyphosis deformity

No pathological change in spinal TB, other than nerve involvement, is an absolute indication for internal fixation. Paired t‐tests of the lumbar lordosis angle of our patients before and after surgery demonstrated a significant difference, indicating that the patients' lordosis angles definitely changed after treatment of lumbosacral vertebrae TB. The improvement was so pronounced that that no correction was required.

Relationship between disc failure, stenosis of the disc space, sequestra, cavities, and kyphosis deformity

Vertebral destruction and abscess are common in spinal TB; and often involve two adjacent vertebrae. Disc failure plus external weight further enhances collapse of the disc space. Severely affected patients will develop more damage to their vertebrae, and show symptoms of cavities, sequestra and collapse as well. There may also be deformities of the whole spine, such as scoliosis, spinal canal compression and kyphosis deformity. Abscesses, cavities and sequestra all belong to the initial and middle stage, and the worst manifestation of spinal TB is whole spine deformity. Through appropriate treatment, resolution of abscesses and cavities and fibrosis of sequestra can be achieved so that the spine recovers clinical stability. MIS, which does not involve an open technique for management of cavities and sequestra, also achieves good results. When confronted with whole spine deformity and neurological deficits, open surgery should be applied. Our research subject is the lumbosacral spine, which differs in many ways from the thoracic and cervical spine, because of its larger vertebrae, stable cauda, strong support for the cauda equina and nerve roots in the spinal canal, and good capacity to compensate for kyphosis. Our 49 research patients all achieved good results with MIS ¹¹ .

Radiological appearance and prognosis of healed spinal TB lesions

When talking about fusion and stability, there has always been an emphasis on removing lesions and on bone graft or fusion after internal fixation. While bone grafting is beneficial, we cannot ignore the following advantages of MIS: stable lesions, normal or slightly narrow intervertebral spaces, non‐ossification between vertebrae, and fibrous stability of paravertebral bone bridges.

Conventional systemic drug therapy for spinal TB (especially for the thoracolumbar segment) has some disadvantages, such as the long treatment period required and low compliance by patients. Some patients do not keep taking medicine, and have no, or not enough, bed rest. For patients who have received surgical treatment, over‐removal of sequestra, absence of bone grafts within the lesions and of reconstruction of the spine are all factors contributing to advanced kyphosis deformity. MIS rapidly kills M. tuberculosis bacteria and requires an insistence on bed rest patients for a certain period. It can be seen on subsequent MR that the vertebrae around lesions appear the same as normal ones, and can support normal body weight. Therefore, except for growing children and teenagers, most patients are not at risk of developing kyphosis deformity in the near future. The long‐term prognosis is still under observation.

Although there is a loss of kyphosis angle during treatment, and a tendency towards kyphosis deformity, the difference is not statistically significant. This may attribute to lumbosacral vertebrae, especially the adhesion muscle near lower waist. In addition, because of the line of gravity in humans and the physiological curvature of the spine, the risk of kyphosis deformity caused by lumbosacral vertebrae TB is far lower than for the thoracolumbar segment.

Indications and contraindications for MIS in lumbosacral vertebrae TB

Our research suggests that MIS is suitable for most lumbar and lumbosacral vertebral TB. However it can neither correct spinal deformity nor relieve canal compression. The spinal canal has a relatively large volume; furthermore the peripheral nerves in the canal have relatively good tolerance for compression, therefore nerve compression rarely occurs in lumbar and lumbosacral TB, making MIS appropriate in most cases. With MIS the emphasis is the following sequence of therapy: improve the patient's condition, eliminate any abscess and determine future therapy in accordance with the patient's progress. Most patients recover smoothly. For patients with severe nerve symptoms, it makes further open surgery safer. For patients with severe nerve compression and kyphosis deformity, MIS is not contraindicated. During MIS, we can observe how the patient responds. Where improvement is apparent, the non‐open technique can be retained; if deterioration occurs, open surgery should be instituted. Although MIS has no capacity for correcting kyphosis deformity, it can quickly improve the patients' general condition, remove any abscess around the lesion and create good conditions for a second stage of open corrective surgery.

New thoughts and methods provided by MIS for spinal TB

Conventional views are as follows:

• 1Local medication is opposed because it is deemed to increase drug resistance.
• 2The lesions of spinal TB cannot be resolved with a drainage tube and attempts to do so will lead to sinuses formation.
• 3Sequestra are the source of recurrence of spinal TB.

In fact, local medication has been a widely applied intervention in clinical practice for treatment of tumors and local wounds. Drainage tubes do not cause sinuses after surgical therapy. On the contrary, inadequate eradication of M. tuberculosis organisms in the lesion is the primary factor resulting in sinus formation. Sequestra are simply radiological evidence of recurrence. Our experience with MIS has proved that lesions can be cured even when sequestra exist. Therefore, it is surviving M. tuberculosis organisms that cause recurrence.

MIS allows the insertion of cannulas under CT guidance and the maintenance of large volumes of lavage. Compared to open surgery, it has the following advantages: recovery from the small wound is rapid; further development of complications is minimized because the number of bacteria is reduced; and inflammatory mediators and exudate in the lesion are diluted, inhibiting any allergic reaction that could lead to further tissue damage. Furthermore, MIS can prevent hematomas, bacterial reproduction and inflammatory diffusion, and reduce pathological scar tissue.

No recurrence was found during follow‐up. For patients who were allergic to isoniazid, or whose organisms were drug resistant as shown by bacterial culture and drug sensitivity testing, the same therapy was still applied and did achieve good results. The reason for this might be that the concentration of drug achieved by local treatment is far higher than the low inhibitory concentrations applicable in drug sensitivity tests.