1 Introduction

Spinal epidural abscess (SEA) presents as a space-occupying lesion between the dura mater and the osseo-ligamentous components of the vertebral canal. Tuberculosis of the spine is a less common cause of SEA, with such cases being commoner in developing countries. Isolated SEA due to tuberculosis is known to have an insidious onset with patients reporting late, usually when the mass effect due to the abscess starts producing paraesthesias or weakness in the extremities. As there is absent to minimal destruction of vertebral elements in the initial stages, patients do not get significant back discomfort. Typically, Epidural abscesses secondary to Tuberculosis of the spine present in the space between the vertebral body and anterior aspect of the dural sac secondary to involvement of the intervertebral disc. However, there is limited literature concerning retro-thecal epidural abscesses which present posteriorly between the dura and posterior spinal elements, and do not have significant vertebral body destruction. This study is a systematic effort to understand and analyze the outcome of such retro-thecal epidural abscesses, over a long-term follow-up.

2 Methods

After obtaining the approval of the Ethics Committee of our institute and written informed consent from the patients included in our study, we performed a retrospective analysis of the data collected from the records of the cases with a retro-thecal epidural abscess on MRI, which were managed surgically or conservatively, with the addition of chemotherapy during 6 years from 2014 to 2020. We included the patients from the age group ranging from 10 to 70 years with at least 1 year of follow-up. We excluded the patients with significant spinal instability due to the destruction of the vertebral column. We evaluated the patients on admission, 2 weeks, 6 weeks, 3 months, 6 months, and 12 months post-initiation of treatment. The data reviewed by us included age, sex, level of the abscess on MRI, the MSCC ratio on MRI, the time interval between the appearance of neurological deficit and the initiation of management, presence of drug resistance, spasticity as per the Modified Ashworth Scale and the pre-operative and post-operative LEMS. We used the LEMS to assess the neurological status. We used the online Surgimap software for calculating the MSCC on MRI. It was calculated on sagittal T2-weighted MRI images as the ratio of the mid-sagittal diameter of the spinal cord at the level of maximum compression caused by the abscess, divided by the average diameter of the spinal cord at the regions above and below the site of maximum compression.¹ [Fig. 1]. Surgical management was chosen in the patients who had a progressive neurological deficit, bowel and bladder involvement, and lack of an adequate response to chemotherapy that was started after obtaining the sample for microbiological investigation either via a C-arm guided biopsy or during surgery. Surgical management included decompression, with or without instrumentation. Under general anesthesia, a prone position was given and a midline posterior incision was taken over the level of affection. Decompression was achieved via a midline laminectomy without compromising the adjacent facet joints/pars. The abscess was isolated and evacuated completely after gradually separating all adhesions from the underlying dura. In case of the presence of an epidural cuff of tissue, the same was removed by sharp dissection. All the samples obtained intra-operatively were sent for microscopy, culture, Gene-Xpert, and histopathological examination for detecting Mycobacterium tuberculosis. Posterior instrumentation using pedicle screws and rods was performed in cases that showed involvement or destruction of the facet joints with resultant instability. Post-operatively, intravenous antibiotics were given for 5 days, followed by oral antibiotics till suture removal. Active range of motion exercises were started by day 3 and patients were gradually mobilized with a brace by day 8 with supervised physiotherapy consisting of bedside mobilization, faradic stimulation, bladder and bowel training, spasticity reduction and proprioceptive training, and gait training. Patients who underwent conservative treatment received medical management in the form of chemotherapy, bracing, and regular clinical follow-up. These patients underwent periodic hematological and radiological evaluations. We evaluated the response to chemotherapy by assessment of the above parameters along with a subjective sense of well-being, improvement in weight, and appetite. We calculated the LEMS by grading motor function on a scale of 0–5, where grade 0 was assigned when there was no motor function, while a grade of 5 meant normal motor function. The motor function was graded for hip flexors (L2), knee extensors (L3), ankle dorsiflexors (L4), great toe extensors (L5), and ankle plantar flexors (S1). The maximum score for each side was 25. We considered there was a major improvement in the neurology or a neurological deterioration when the change in the score was more than 10 points as compared to our previous assessment.² Chemotherapy consisted of fixed-dose combination tablets containing isoniazid (H), rifampicin (R), pyrazinamide (Z), and ethambutol (E), with injection Streptomycin (S). Chemotherapy consisted of an intensive phase of 2 months (HRZES) followed by a continuation phase of 6 months (HRE). We used Microsoft Excel for assembling our data for analysis. The mean, standard deviation (SD), standard error (SE) were determined and the paired t-test was used to determine the correlation, after setting the level of significance at 0.05.

Close

Fig. 1 Method of calculation of MSCC ratio. It is a ratio of the mid-sagittal diameter of the spinal cord at the compression site (Line 2) divided by the average diameter of the spinal cord at the non-compressed regions above (Line 3) and below (Line 4).

Export to PPT

3 Results

The average follow-up in our study was 12 months. There were 3 males (30%) and 7 females (70%) in our study. The mean age of our patients was 22.30 with a SD of 17.08. All patients had retrothecal spinal epidural abscesses without significant destruction of the posterior vertebral column (facet joints and spinous processes) on MRI. The mean MSCC ratio was 0.503. The average time interval between the onset of neurological deficit and the initiation of management was 6 weeks. Decompression was performed in 8 patients, of which 2 patients required additional instrumentation. 2 patients were managed conservatively [Table 1]. Chemotherapy was given to all patients. The spasticity developing after the initiation of treatment was graded using the Modified Ashworth Scale. Two patients developed Grade 2 spasticity of which one had a retrothecal abscess extending from D5 to D7, while the other had a retrothecal abscess extending from L1 to L2. Another patient having involvement from D7 to D9 developed Grade 3 spasticity. The spasticity was seen in those patients in whom the time interval between the onset of neurological deficit and the initiation of management was more than 6 weeks (p value < 0.05). Drug resistance was detected in one patient. The patient defaulted after taking chemotherapy for 3 months. The mean LEMS on admission was 20.2, which improved to 38.5 at the end of 1 year (p-value <0.05) [Table 2]. Patients who had a time interval fewer than 6 weeks between the appearance of neurological deficit and the initiation of management had better LEMS, with the relationship being significant at a follow-up of 3 months, 6 months, and 1 year. (p-value<0.05) [Table 3] Spasticity was observed in the patients who were intervened after a period greater than 6 weeks after the onset of neurological deficit. (p-value<0.05) [Table 4]. The MSCC ratio did not have a significant correlation with the LEMS. (p-value >0.05)[Table 5].

4 Discussion

Tubercular SEA usually develops secondary to involvement of the vertebral body and rarely due to hematogenous spread from another primary focus in the body.³ Tuberculosis initially involves the anterior inferior portion of the vertebral body, following which it spreads into the central part of the body or disk. Common patterns of vertebral body involvement include para-discal, anterior, and central, of which para-discal type is the commonest and is seen commonly involving the junctional areas. A retro-thecal spinal epidural abscess in the setting of tuberculosis, without the presence of spinal instability, is a less common pattern. Due to the insidious nature of such a lesion, patients present late in the disease course, especially in developing countries. The classically reported triad of fever, back pain, and neurological deficit is rarely seen in cases of tubercular SEA. Most of the cases have a progressive neurological deficit and even quadri-paresis or para-paresis at the time of presentation.⁴ In patients who have a chronic infection, constitutional symptoms of fever, weight loss and, other systemic features may predominate over the neurological features leading to a delay in diagnosis and management.⁵ The investigation of choice in these cases is MRI as it is a soft tissue lesion that is responsible for the symptoms. The MRI findings of SEA include a hyperintense lesion seen in T2-weighted sequences with a peripheral rim enhancement, which may or may not be associated with signal changes in the vertebral body or the overlying spinal elements. As per Hasan et al.,⁶ it is difficult to distinguish epidural and intradural abscesses due to similar radiological features, and thus, differentiating them based on radiological investigations would be misleading. In a few cases, the true location of the abscess is revealed intra-operatively on performing a durotomy. Surgical management is indicated in cases presenting with progressive neurological deficit secondary to compression, failure of medical management, and mechanical instability secondary to the abscess. Surgical options include decompression with or without instrumentation.

There are anecdotal case reports available in the literature that have mentioned operative treatment as the primary modality of management of these lesions. Metta et al.⁷ presented a case of a 35-year old immunocompromised male who presented with spastic paraparesis, due to an epidural extension of the disease process at D6-D8 level. The patient was managed with decompression and debridement and had a good neurosurgical outcome. Intraoperative samples sent for examination yielded Mycobacterium tuberculosis. Esteves et al.⁸ presented a rare case of a cervical SEA secondary to Mycobacterium tuberculosis that did not show any osseous involvement and was managed with a 2-stage procedure involving decompression and arthrodesis. Arora et al.⁹ presented a case of a 35-year-old male who had a multi-segmental tubercular SEA extending from D5 to S2 without any osseous involvement. The patient was managed with surgical decompression and chemotherapy. These anecdotal case reports concluded that early diagnosis and intervention are essential for a successful outcome in these cases. However, our study differed from them in regards to that we had patients managed not only surgically [Fig. 2], but also conservatively. Being a tertiary apex center, our study had patients from various age groups and various backgrounds including neglected cases with a delayed presentation due to socioeconomic issues. This helped us to better understand the natural history of this infectious pathology as most of the patients had not received any disease-specific management at the time of presentation. The patients in our study did not show the presence of spinal instability at the time of presentation. Destruction of more than 2 spinal columns as per the Denis three-column model or segmental kyphosis more than 30° was considered to be the surrogate marker of spinal instability. This criterion is arbitrary and can thus be considered as a limitation of the study. In the present study, the time interval between the onset of neurological deficit and the initiation of management was the single most important predictor of a successful neurological outcome. Most of our cases were neglected and presented late. The mean time interval between the onset of neurological deterioration and the initiation of management was 6 weeks. However, these cases showed a satisfactory outcome when managed with decompression, even though treatment was delayed. Rathod et al.¹⁰ studied the neurological recovery pattern in patients with tuberculosis of the spine with progressive neurological deficit undergoing delayed decompression and fixation. Their study concluded that delayed decompression in cases of tuberculosis of the spine had favorable outcomes, a finding seen in our study as well. A difference between both the studies was that the former study included patients with spinal column involvement showing osseous destruction whereas our study group included epidural abscesses without significant osseous destruction. Muzzi et al.¹¹ reported a series of 8 cases of cervical spine epidural abscesses managed with a minimally invasive surgical approach consisting of discectomy and drainage, without arthrodesis. They concluded that in the absence of spinal instability, drainage of the abscess without arthrodesis resulted in neurological recovery. In the present study, 2 out of the 8 patients requiring decompression needed additional instrumented fusion as they showed significant destruction of the posterior column in the form of near total involvement of the facet joints, which was seen intra-operatively. This was done to provide spinal stability and prevent the development of kyphosis, as neurological recovery in the presence of spinal instability could be sub-optimal.

Close

Fig. 2 (a): Pre-operative MRI of a 13 year old female with a D1 epidural abscess, who was managed with laminectomy and fixation at C7, D2, D3 using pedicle screws; MRI of the same patient taken at a follow-up of one year shows complete resolution of the abscess Fig. 2(b): Immediate post-operative radiograph of the 13 year old female with a D1 epidural abscess managed with laminectomy and fixation at C7, D2, D3 using pedicle screws.

Export to PPT

The persistence of varying levels of spasticity was a common finding in almost all cases. This occurred even when the patients regained functional power, though the spasticity was of mild grade in most cases. This indicates the residual effect of spinal cord compression, even though the compression was relieved with medical and operative management. We followed a protocol of extensive physiotherapy and gait training as a part of our post-operative protocol, which helped to ameliorate the spasticity in most cases, while few resistant cases were managed additionally with oral baclofen. A study by Takahashi et al.¹² concluded that gait training was safe in patients of cerebral palsy, in whom spasticity is a major concern, and it produced immediate effects on walking ability in ambulatory patients with the disease. A study by Schiess et al.¹³ demonstrated the long-term safety of intrathecal baclofen in the treatment of severe spasticity of cerebral and spinal origin. However, as the spasticity was mild and responded to conservative measures in our patients, we have no experience of intrathecal baclofen. Corticosteroids can be given to ameliorate the edema caused by the mass effect of the abscess. However, long-term studies are required to evaluate the effects of steroids in these lesions as currently, there is insufficient evidence regarding the same.

In the authors’ experience, it is not surprising to find more advanced destruction of the posterior elements during surgery as compared to what is seen on preoperative MRI imaging, since MRI is a poor imaging modality for tuberculosis involving the posterior bony elements. However, due to the resource constraints and concerns for radiation exposure, a CT scan was not done, which would have been an ideal imaging modality for knowing the status of the posterior bony elements. We used MRI for pre-operative planning provided it was done within 4 weeks from the proposed date of surgery. If the MRI was older than 2 weeks, the patient underwent a fresh radiograph to look for the integrity of the spinal column.

5 Complications

One patient in our study showed a worsening of the neurological deficit on follow-up. Multi-drug resistant Mycobacterium Tuberculosis was detected in her biopsy samples and she was started on second-line chemotherapy. However, chemotherapy was stopped by her parents after 3 months, following which they refused any further treatment. There was one case of an intra-operative dural tear, due to the presence of extensive epidural adhesions. The limitations of our study included the limited sample size and the retrospective design of the study.

6 Conclusion

Retrothecal tubercular SEA requires expeditious management that ranges from conservative to surgery, with clinical assessment being of utmost importance as seen in our study where one patient with severe compression on MRI but a good LEMS was successfully managed conservatively. The MSCC ratio did not have a significant relationship with the LEMS. The time interval between the onset of neurological deficit and the initiation of management was the single most important predictor of a favorable outcome.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Consent for publication

Appropriate written informed consent has been taken from all the patients for inclusion in the study and publication process.

Individual contributions

Dr. Rudra Mangesh Prabhu – Data curation, Investigation, Methodology, Writing- Original draft. Dr. Tushar Narayan Rathod – Conceptualization, Data curation, Investigation, Formal analysis, Supervision, Validation. Dr. Shivaprasad Sharangouda Kolur – Data curation, Investigation. Dr. Bhushan Sunil Hadole – Software. Dr. Shital Chavan – Writing – review and editing. Dr. Nandan Amrit Marathe – Supervision. Dr. Abhishek Kumar Rai – Writing – review and editing.