Tuberculous meningitis: Challenges in diagnosis and management: Lessons learnt from Prof. Dastur's article published in 1970
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.246224
Source of Support: None, Conflict of Interest: None
Keywords: Antituberculous treatment, challenges, complication, tuberculosis, tuberculous meningitis
Nec minus a phlegmone et abcessu quam hujasmodi meningitis et tuberculis, cephalgiaelethales et incurabilesoriuntur
(Sometimes the headaches, fatal and incurable, follow abscesses and swellings of the envelopes of the brain, as well as plaques and tubercles of these membranes)
First described by Willis in the 17th century, tuberculous meningitis (TBM) is the most severe form of tuberculosis (TB). TB in all its forms remains a challenging clinical problem and a public health issue of considerable importance and magnitude, the world over. It continues to be a worldwide burden, with majority of new active cases occurring in underdeveloped countries. As per the World Health Organisation (WHO) statistics, five countries, viz., India, China, Pakistan, Indonesia and South Africa, account for over 70% of the global burden of disease. Mycobacterium tuberculosis (MTB) causes approximately 10.4 million new cases of TB and 1.5 million deaths annually, with an additional 0.4 million deaths in individuals co-infected with human immunodeficiency virus (HIV). Patients coinfected with HIV are at more than 20 times higher risk of developing TB compared to non-infected individuals. Although central nervous system TB accounts for 5-10% cases of extrapulmonary TB and only 1% of all cases of TB, it is responsible for more deaths than any other form of TB, owing to the inherent seriousness of this illness.
Almost 5 decades back, Dastur and colleagues in their seminal paper had highlighted the gross pathological changes in 100 patients of TBM, who succumbed to this illness.
The work of Professor Dastur and colleagues continues to be highly relevant in the modern era.
The landmark paper by Dastur et al., gave us an insight into the pathophysiology and pathological changes of CNS TB, which is being replicated in today's era by advanced imaging techniques. With the advent of magnetic resonance imaging (MRI), more so the newer sequences like magnetisation transfer (MT), gradient recalled echo (GRE), susceptibility-weighted (SW), diffusion weighted (DW) and fluid attenuating inversion recovery (FLAIR) images, the cause and extent of the disease, its underlying pathophysiology, its complications, and the favourable or adverse response to treatment can now be gauged with almost the same precision during life as what was reported in post mortem samples by Dastur et al. A careful interpretation of the data provided by Dastur et al., suggests that majority of their patients had one or more life-threatening complications in the form of hydrocephalus, infarcts, severe arachnoiditis, including spinal arachnoiditis and tuberculomas, which correspond to Stage 3 of Medical Research Council (MRC) staging [Table 1]., As almost all the cases analysed by Dastur et al., had advanced TBM, it is reasonable to expect that the findings presented by them may not be evident in some patients, especially those in an early stage of TBM. In these patients, magnetization transfer (MT) MR imaging detects the earliest evidence of meningitis in the form of hyperintense signal changes on T1 weighted MT sequences, whereas the conventional spin echo sequences may be normal. The common sites for basal meningeal enhancement include the interpeduncular fossa, the pontine, perimesencephalic and suprasellar cisterns, and the Sylvian fissures [Figure 1]a. The degree of enhancing exudates varies from a thin layer to a thick sheet [Figure 1]b, as documented by Professor Dastur and colleagues. In more advanced cases, as described by Dastur et al., these exudates may block the flow of CSF (cerebrospinal fluid) producing hydrocephalus, which may be communicating (dilatation of all the ventricles) [Figure 1]c primarily due to basal exudates, or non-communicating either due to narrowing of the cerebral aqueduct or obstruction to the CSF flow at the level of foramen of Lushka and Magendie, through various mechanisms. All these pathological findings can now be revealed during life with the available imaging modalities. While contrast enhanced computed tomography (CECT) scan is the imaging modality of choice for establishing the diagnosis of TBM in an emergency setting, MRI gives a much more objective information about the degree of exudates, degree of hydrocephalus, the associated periventricular ooze as well as other complications like the development of infarcts and borderzone encephalitis (BZE). The latter was a phenomenon described by Dastur et al., that occurs in the form of localised necrosis of the underlying brain parenchyma in relation to the infiltrative exudates [Figure 2]a, [Figure 2]b and [Figure 2]c, [Figure 3]a and [Figure 3]b.
Dastur et al., noted gross compression and narrowing of large arteries (especially the middle cerebral arteries) due to meningovasculitis at the base of the brain with resultant brain infarctions (mainly in basal ganglionic and thalamic regions) in almost 50% of children and 33% of adults. With the advent of new MRI sequences, these changes can be seen during life as areas of diffusion restriction on DWI [Figure 4]a and [Figure 4]b. Complications like optochiasmatic arachnoiditis [Figure 5]a, [Figure 5]b and [Figure 5]c, as well as spinal arachnoiditis and tuberculomas [Figure 6]a and [Figure 6]b can be documented with precision on the contrast enhanced MRI.
To summarise, newer imaging modalities have empowered the radiologists and treating physicians to diagnose TBM with a reasonable degree of certainty at an early stage, thereby providing an early treatment of TBM. All these imaging findings are in concurrence with the findings lucidly described in the gross pathology by Professor Dastur and his colleagues. This clinico-pathologico-radiological correlation has given an edge to the clinicians for establishing an early diagnosis of TBM and in treating the condition.
Despite availability of descriptive literature on the pathological findings in TBM for more than 5 decades, due largely to the seminal work of Professor Dastur and colleagues, the morbidity and mortality in TBM continues to be unacceptably high. Reported mortality figures, as per the MRC stage at the time of presentation, are 4% in Stage 1, 11% in Stage 2 and 50% in Stage 3 in a study from North India and 20% for stage 1, 30% for stage 2 and 50% for stage 3 in a study from Vietnam. Currently, the major challenges to the successful management of TBM include:
A. Early diagnosis of TBM:
This continues to be the biggest challenge in TBM due to following reasons:
A1: Vague and varied clinical symptomatology
A2: Poorly sensitive diagnostic laboratory parameters.
A1. Vague and varied clinical symptomatology
The clinical symptomatology of TBM is often not specific. A prodromal period with non-specific constitutional symptoms (irritability, anorexia, weight loss or sleep disturbance in children; and, fatigue, loss of appetite, loss of weight and night sweats in adults) can last from a few days to several weeks in a majority of patients. This is especially true in the underdeveloped world where rampant misuse of quinolones and other antibiotics often suppresses the initial symptomatology of TB.
The typical duration of symptoms in TBM is usually of a few weeks and even months, though some of the diagnostic criteria require a minimum duration of 5 days only. The typical symptoms include headache, fever, vomiting, meningism, neurological deficits and altered mental status. Photophobia is less common than headache and neck stiffness. In children, vomiting and convulsions are more common than in adults. The classical triad of meningitis, i.e., fever (adults: 60-75%, children: 67%), headache (adults: 50-80%, children: 25%) and vomiting (adults: 40-80%, children: 98%) may not be present in all patients. Neck stiffness is usually absent during the early disease., To overcome the challenges brought about due to the varied symptomatology, a high index of suspicion and performance of appropriate investigations (cerebrospinal fluid [CSF] analysis and neuro-imaging) are needed to clinch the diagnosis especially in the early stages of TBM. Various diagnostic algorithms have been proposed by Ahuja et al., and Marais et al., (including the modified algorithm by Ahuja et al.,) for a uniform case definition of TBM especially for research purposes, where definite TBM has been defined in only cases where Mycobacterium has been isolated from the CSF by any of the available microbiological techniques.
A2. Poorly sensitive diagnostic lab parameters 9
TBM is a paucibacillary disease; thus CSF microscopy using the Ziehl Neelsen stain [which can confirm the diagnosis quickly through demonstration of acid-fast bacilli (AFB)] has a very low sensitivity (0-20%). The results of culture of MTB are notoriously slow to obtain, with the conventional solid media (the Lowenstein-Jensen media) yielding results only after 10-35 days. To overcome this challenge, many new diagnostic techniques have become available in the recent years. These have shortened the lag period for establishing the diagnosis of TBM and have improved the sensitivity while retaining a good specificity. Currently, several liquid culture systems including BACTEC MGIT 960 system (Becton Dickinson Microbiology Systems, Sparks, Md), and MB/BacT system (BioMérieux, Durham, N.C) are available, which yield positive results in a shorter period of time with the added advantage of providing drug susceptibility testing (DST) of MTB at the same time.
Commercial NAATs (nucleic acid amplification tests) have shown potential as rapid ‘rule-in’ diagnostic tests for TBM, with a high specificity (98%), but a low sensitivity (56%)., A more recent review has shown the promise that multiplex polymerase chain reaction (PCR) techniques offer, which have better sensitivity compared to the commercial NAATs (sensitivity 71-94%, specificity 88-100%). Newer NAATs are showing promising results by increasing the sensitivity and specificity for the detection of MTB as well as for assessing drug resistance in a short duration. However, these tests require specialized laboratory settings with rigorous quality control.,
Xpert MTB/RIF (Cepheid, Sunnydale, CA, USA; an automated diagnostic test that can identify Mycobacterium tuberculosis [MTB] DNA and resistance to rifampicin [RIF]) is the most recently endorsed diagnostic test for TB by the WHO (World Health Organization) in 2010. The main advantages of this test are that its technique can be learned easily, the machine can be used in decentralized settings, the turnaround time is just 2 hours, and the closed disposable cartridge system reduces the risk of contamination. The sensitivity of Xpert MTB/RIF in a large Vietnamese study was 59.3% [(n = 108/182 (95% confidence interval (CI) 51.8; 66.5)], which is (significantly though only slightly) lower than that of the mycobacteria growth indicator tube (MGIT) culture [66.5% (n = 121/182), (95% CI 59.1; 73.3)]. However its specificity was comparable to that reported for other commercial NAATs [99.5% (95% CI 97.2; 100)]. An important advantage of Xpert MTB/RIF is that it can detect rifampicin (RIF) resistance within two hours. In the above study, four patients of RIF resistance (n = 4/109, 3.7%) were identified by Xpert, of which 3 were confirmed to be suffering from multi-drug resistance (MDR) TBM. These numbers are too small to draw robust conclusions on the positive or negative predictive value of rifampicin resistance testing by Xpert MTB/RIF. However, as rifampicin resistance is associated with a very high mortality in TBM, it is advised that a positive result for rifampicin resistance, in the context of a clinical picture that fits drug resistant TB, should prompt clinicians to consider immediate second-line treatment. Xpert MTB/RIF has the ability to improve tthe establishment of an early diagnosis of TBM, and in particular, to open the field for clinical trials researching on the treatment optimization for patients with MDR-TBM.
It is pertinent to note that Dastur et al., reported evidence of TB elsewhere in the body (especially in the lungs and lymph nodes) in almost all the autopsied bodies. Taking an important lead from this finding, we analysed 70 cases of TBM patients [(either definite (n = 26) or probable (n = 44)] with whole body FDG PET (flouro-deoxy glucose positron emission tomography) and found evidence of TB elsewhere in the body in 66 (94.3%) of the patients. The FGD-PET showed evidence of pulmonary involvement in 62 (88.6%) and lymph nodes involvement in 61 (87.1%) patients [Figure 7]. Therefore, if there is any doubt about the diagnosis, additional body imaging in the form of chest X ray, a computed tomographic (CT) image of the chest and abdomen, and if required, a whole body FDG PET, can help in taking a decision regarding the start of ATT. Taken together, it is reasonable to conclude that newer techniques such as Xpert MTB/RIF, NAATs, multiplex PCR for MTB as well as whole body FDG-PET have the potential to help in establishing an early diagnosis of TBM and thus, in aiding to change the management scenario of TBM completely.
B. Effective treatment of TBM
The drug regimens recommended for the management of TBM are derived from short course treatments of pulmonary TB. However, the current algorithms for the treatment of TBM have ignored an important fact, i.e., brain should be considered as a distinctive compartment, and thus, the dosage and duration of therapeutic regimen for TBM should be established taking into account the pharmacokinetic (PK) and pharmacodynamic (PD) evidence. Even today, the optimal drug dosage and the appropriate duration of treatment for TBM have not been unequivocally established by large clinical trials. The challenges surrounding the effective treatment of TBM continue to be related to:
B1. Drugs, dosages and duration of antituberculous therapy (ATT)
B2. Treatment of multi-drug resistance (MDR) TBM.
B1. Drugs, dosages and duration of ATT
The first-line anti-TB agents recommended for the treatment of TBM include rifampicin (RIF), isoniazid (H), pyrazinamide (PZA), ethambutol (EMP) and/or streptomycin (S). Among the first-line drugs used in TBM, RIF, EMP and streptomycin have poor penetration across the blood-CSF barrier. Currently, WHO recommends a 2-month treatment with 4 first-line drugs in the intensive phase, followed by a continuation phase with at least rifampicin and isoniazid for 4-10 months. However as mentioned earlier, there is no data from randomized controlled trials to form the basis for these recommendations. Some authors have advocated a longer course of therapy for up to 2 years or even more, whereas other authors have suggested that short term RIF based regimens for 6 to 9 months may be adequate., Accordingly, treatment of TBM varies from place to place. The current United Kingdom guidelines suggest treatment with rifampicin, isoniazid, pyrazinamide and a fourth agent (streptomycin, ethambutol, or prothionamide) for the first two months followed by rifampicin and isoniazid for 10 months for uncomplicated cases of TBM (including cerebral tuberculomas without meningitis). ATT should be given for 18 months if PZA is not a part of the initial treatment regimen.
The American Thoracic Society recommends the initiation of therapy with isoniazid (INH 10mg/kg/day up to 300 mg/day), rifampicin (RIF 10-20mg/kg/day up to 600 mg/day), pyrazinamide (PZA 15 mg/kg/day up to 2 gm/day) and ethambutol (EMB 20mg/kg/day up to 1.2 gm/day). After the initial 2 months, INH and RIF alone are continued for an additional 7-10 months, although the optimal duration of therapy is not defined.
Recently, several studies have shown that high doses of RIF in TBM regimens are associated with higher levels in CSF and improved survival, though these studies have enrolled only a small number of patients. Also, there is lack of data about the use of fluoroquinolones in TBM, which remain an attractive option as they are active against MTB, are well tolerated, have an extensive safety data, have relatively little resistance to MTB and have a good penetration in the CSF; their clinical use, however, has shown inconsistent results.,
Taken together, it is clear that there are still doubts regarding the optimal dosages, regimen and duration of ATT in TBM, and thus, large randomised trials are the need of the hour to sort out these issues.
B2. Treatment of MDR TBM
Drug resistant TB is a growing problem globally. Resistance to rifampicin is associated with a very high mortality in TBM, a finding which emphasizes the importance of including rifampicin in successful treatment regimens. Globally, approximately 20% of isolates in TB are resistant to at least one antituberculous drug, and 7% are resistant to at least isoniazid. Isolated resistance to isoniazid with or without resistance to streptomycin is more frequently found in high-burden settings. Isoniazid resistance has been associated with increased mortality due to TBM, especially in those infected with human immunodeficiency virus (HIV). An unfavourable outcome may be prevented by the use of pyrazinamide throughout the treatment period. [RRR] MDR-TBM (resistance to at least rifampicin and isoniazid in multi-drug resistance tuberculous meningitis) is highly lethal. Most patients die within two months after the treatment initiation. The management of MDR-TBM continues to be far from satisfactory mainly for two reasons. Firstly, timely detection of drug-resistance is limited by the lack of rapid diagnostics. Thus, the timing of an adjusted treatment schedule for drug resistant TBM based on the currently available culture and sensitivity results is often too late to prevent neurological disability or death. Second, due to the lack of good detection techniques, there are no studies evaluating the optimal regimen for drug resistant TBM. It is pertinent to mention that newer diagnostic techniques (Xpert MTB/RIF, multiplex PCR, liquid based culture medium) have shown promise for detection of MDR/DR in the early stages of TBM. Whole genome sequencing of MTB may help in defining the drug resistance related genes and may help in the rapid diagnosis of MDR TB and in the better management of patients., However, large studies remain to be done to validate their routine use. If found successful, these methods may reveal MTB strains resistant to rifampicin at an early stage and open the field for trials focusing on the second-line treatment for drug resistant TBM.
D. Management of complications
The optimal management of complications of TBM like optochiasmatic arachnoiditis, spinal arachnoiditis, vasculitic infarcts, tuberculomas (especially, paradoxical as well as persisting lesions) and hydrocephalus, remains to be defined. The paradoxical response (i.e., clinico-radiological deterioration after start of ATT) is frequently reported. This can be observed in all tissues, but most often in the lungs, lymph nodes and the brain. In the brain, the paradoxical response often occurs in the form of appearance of new tuberculomas or enlargement of pre-existing tuberculomas, once ATT has been started (usually within 1-4 months after the start of ATT). It may be accompanied by worsening of symptoms or the appearance of signs of a space occupying lesion. This paradoxical response is considered to be more related to a heightened immune response rather than a manifestation of drug resistance. Once other causes of deterioration (such as drug resistance or the development of new infarcts) have been ruled out, these patients should be managed by continuation of ATT and administration of high dose systemic corticosteroids. Some authors have suggested the implimentation of a more aggressive immunomodulation with drugs. The proposed therapies for various complications are:
C1. Use of immuno-modulatory drugs
C2: Treatment of stroke and vasculitis
C3: Surgical treatment of complications.
C1: Medical therapy including the use of immuno-modulatory drugs:
The various options being used to treat these complications include a longer duration of treatment, introduction of second line ATT, use of high dose steroids for a longer duration, use of immunomodulatory drugs like thalidomide, levamisole, adjuvant interferon gamma therapy and administration of intrathecal hyaluronidase, etc., In our personal experience of >50 patients, the use of thalidomide (in a dose of 2 mg/kg body weight) for 4-6 months in patients with TBM suffering from various sequel (such as optochiasmatic arachnoiditis, a paradoxical increase in the size of tuberculomas, an increase in size and number of lesions, and spinal arachnoiditis), has resulted in an improvement in the clinical condition of the patients and has resulted in radiological resolution of the lesions in >70% of patients, who had already received >2 months of ATT and steroids. Certain recently published studies on a small number of patients have also demonstrated resolution of lesions by using thalidomide. It is pertinent to mention that robust scientific evidence for the use of any of above options is lacking, and thus, the need of properly conducted large multi-centre trials to provide guidelines for better management of these patients cannot be overemphasized.
C2. Tuberculous vasculitis and stroke
The poor outcome from TBM primarily reflects the extent of ischemic damage to the brain resulting from inflammation, necrosis, and thrombosis of blood vessels involved in meningovasculitis. The most common clinical manifestation of TBM-related stroke is hemiplegia, which is more common in young children than in adults, and in patients with advanced disease, as had been highlighted by Dastur et al.
No adjunctive treatment has consistently reduced the incidence of stroke or changed the course of hemiplegia in TBM. Corticosteroids did not significantly affect the number of new infarcts seen on CT or MRI, or the extent of residual hemiplegia in children or adults. Only aspirin may help to reduce the incidence of stroke, but its role needs to be confirmed in larger studies.
C3. Surgical management of hydrocephalus
The mainstay of management of TBM with hydrocephalus is CSF diversion. Surgical management can be planned according to the Vellore grade at presentation [Table 2]. Grade I patients can be monitored on medical management but early surgery has shown a better outcome. Grade II patients definitely show improvement with early surgery. Some authors have advocated a brief period of external ventricular drainage (EVD) before shunt surgery in grade III patients but direct shunt surgery is better. An EVD has been recommended for grade IV patients followed by installation of a CSF diversion in the form of a shunt in only those patients who have shown an improvement following the installation of an EVD. However, some authors have found a good outcome in 20% of grade IV patients who have undergone shunt surgery without installation of a prior EVD. In the above context, it is important to note that while earlier studies showed 100% mortality in grade IV hydrocephalus, a more recent study found a mortality rate of 60% in stage IV hydrocephalus. The lower mortality in this study was attributed to an early shunt surgery in addition to the institution of ATT, steroids and supportive care.
Recently, more and more interest has been generated in the endoscopic CSF diversion procedures for TBM. The endoscopic third ventriculostomy (ETV) is not a favoured procedure in hydrocephalus, at least theoretically, as it is usually of a communicating type in TBM. There are however, several reports on the use of ETV in hydrocephalus due to TBM. The success rate of ETV has been reported to range from 68% to 77% in patients in different studies. The experience of a surgeon well-versed in endoscopic procedures determines the success rate, as TBM exudates and scarring cause difficulty in recognizing the anatomical landmarks at the floor of third ventricle, thus hampering the efforts to perform a successful third ventriculostomy. Patients with longer duration of symptoms and those who have received at least four weeks of ATT are more likely to benefit from ETV.
Financial support and sponsorship
Conflicts of interest
Authors declare that they do not have any conflict of interest. They do not have any source of funding.
[Additional file 1]
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]
[Table 1], [Table 2]