Tuberculous meningitis: The challenges
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.72178
Source of Support: None, Conflict of Interest: None
Tuberculous meningitis (TBM) is a serious meningitic infection commonly found to occur in the developing countries endemic to tuberculosis. Based on the clinical features alone, the diagnosis of TBM can neither be made nor excluded with certainty. Unfortunately there is still no single diagnostic method that is both sufficiently rapid and sensitive. Most factors found to correlate with poor outcome can be directly traced to the stage of the disease at the time of diagnosis. The only way to reduce the mortality and morbidity is by early diagnosis and timely recognition of complications and institution of the appropriate treatment strategies.
Keywords: Antituberculous drugs, elevated intracranial pressure, hydrocephalus, hyponatremia, tuberculoma, tuberculous meningitis, vasculitis
Tuberculous meningitis (TBM) is still one of the common infections of central nervous system (CNS) and poses significant diagnostic and management challenges, more so in the developing world.  Despite modern antituberculosis chemotherapy, 20% to 50%  of patients still die, and many of the survivors have significant neurological deficits. Death from TBM is strongly associated with delays in diagnosis and treatment. This review discusses some of these challenges.
Global burden of tuberculosis is still high, particularly in developing countries; and globally, there were an estimated 9.27 million new cases (139 per 100,000 population) of tuberculosis in 2007, and the number of prevalent cases was 13.7 million (206 per 100,000 population).  The incidence of CNS tuberculosis generally reflects the incidence and prevalence of tuberculosis in the community. About 10% of patients who have tuberculosis develop CNS disease.  HIV infection predisposes to the development of extra-pulmonary tuberculosis, particularly tuberculous meningitis.  With 206 per 100,000 prevalent cases of tuberculosis in 2007  and the projected incidence of cases of CNS tuberculosis being 20.6 per 100,000 population in the year 2007, most of it would be in the high-burden countries. Incidence rates of tuberculous meningitis are age specific and range from 31.5 per 100,000 (<1 year) to 0.7 per 100,000 (10-14 years) in the Western Cape Province, South Africa.  The estimated mortality due to tuberculous meningitis in India is 1.5 per 100,000 population.  HIV co-infection is associated with higher complication and case fatality rates. ,
TBM cannot be diagnosed clinically with certain and rapid diagnosis, particularly in patients with grade II and III disease,  is critical for instituting appropriate interventions to achieve good outcomes. In a significant number of patients, the diagnosis of TBM is empirical and is based on clinical, laboratory and neuroimaging data.
Clinical diagnosis of TBM is difficult as the clinical features are nonspecific and vary widely, and is often diagnosed when brain damage has already occurred. ,, The classical triad of meningitis, viz., fever (adults, 60%-95%; children, 67%), headache (adults, 50%-80%; children, 25%) and signs of meningismus (adults, 40%-80%; children, 98%), may not be present in all the patients.  Altered mental status is a more common presenting feature in children as compared to adults.  In the elderly, signs of meningismus may be absent and seizures occur more commonly.  Patients with HIV co-infection may less commonly have fever, headache and meningismus, and they are more likely to have an altered mental status. ,
Based on the clinical features alone, the diagnosis of TBM can neither be made nor excluded with certainty. In children the clinical variables predictive of TBM include symptoms persisting for more than 6 days, optic atrophy, focal neurologic deficit, abnormal movements, and a CSF leukocyte differential of less than 50% neutrophils.  The diagnostic sensitivity was 98%, and specificity was 44% when at least one feature was present; and sensitivity was 55%, and specificity was 98% if three or more features were present. An adult study in Vietnam identified five clinical variables predictive of diagnosis of TBM: age (<36 years: 2; >36 years: 0); white cell count (>15000: 4; <15000: 0); history of illness (>6 days: -5; <6 days: 0); CSF leukocyte count (>750: 3; <750: 0), and percent of CSF neutrophils (>90: 4; <90: 9).  A maximum score of four or more on admission was diagnostic of TBM. This diagnostic rule had a sensitivity of 86% and a specificity of 79%.
Laboratory diagnosis of TBM includes many diagnostic methods. Unfortunately there is still no single diagnostic method that is both sufficiently rapid and sensitive. Several of the diagnostic methods used for the diagnosis of TBM are relatively insensitive.Most often, the diagnosis is dependent on lumbar puncture and cerebrospinal fluid (CSF) examination. CSF of untreated TBM typically shows moderate lymphocytic pleocytosis, moderately elevated protein concentration and low glucose. However, the CSF profile of TBM mimics the profiles of a large list of both infectious and noninfectious meningitic processes. Acellular CSF has been reported in the elderly and in patients with HIV-co-infection. ,,, Bacteriological diagnosis, demonstration of acid-fast bacilli (AFB) of M. tuberculosis by Ziehl-Neelsen stain (sensitivity, 25%) and culture (sensitivity, 18%-83%) is highly specific (100%).  In the Vietnam study, volume of CSF, duration of symptoms, CSF neutrophil count, lactate and glucose were all independently associated with bacteriological confirmation.  Tuberculostearic acid is a fatty acid component of the M. tuberculosis cell wall, which has been detected in CSF of patients with TBM, and the method has good sensitivity (95%-100%) and specificity (91%-99%). But the limitations are that it requires an expensive equipment and considerable expertise. 
Tests to detect M. tuberculosis-specific antibodies and antigen in CSF of patients with TBM are rapid and less expensive. But these techniques are limited by the inability to differentiate acute infection from previous infection and by problems with cross-reactivity, in addition to variable and often poor sensitivity and specificity. For the antibody assays, the reported sensitivity and specificity are 52%-93% and 58%-99%, respectively. Similar are the rates for antigen assays; the reported sensitivity and specificity are 38%-94% and 95%-100%, respectively. 
Commercially available nucleic-acid amplification and other polymerase chain reaction (PCR) assays may provide a screening tool for the diagnosis of TBM. The value of these assays lies largely in the rapidity with which results can be obtained and the very good specificity of these tests.  A systematic review of the diagnostic sensitivity and specificity of commercially available nucleic-acid amplification assays for TBM revealed 56% sensitivity and 98% specificity, and the negative predictive value and positive predictive value were 44% and 35.1%, respectively. The sensitivity of these assays is too low - about half of those with a negative result will have the disease. 
Neuroimaging, both contrast computerized tomography (CT) and magnetic resonance imaging (MRI), reveals the pathology and the complications of TBM. The image characteristics are nonspecific and include basal meningeal enhancement, hydrocephalus, tuberculoma(s) and infarcts ,,,,,,,, ; but when correlated with the given clinical features, they may give a clue for the diagnosis.  Certain imaging findings on contrast CT seem to be specific for TBM in children: basal meningeal enhancement, tuberculomas, or both (sensitivity, 89%; and specificity, 100%)  ; and presence of hyperdensity in the basal cisterns in the noncontrast CT scan. 
Treatment of tuberculous meningitis
There is scarcity of controlled trials of antituberculous drugs in CNS tuberculosis. Most of the guidelines follow the model of short-course chemotherapy of pulmonary tuberculosis: an "intensive phase" of treatment with four drugs, followed by treatment with two drugs during a prolonged "continuation phase."  Infectious Disease Society of America, Centers for Disease Control and Prevention, and American Thoracic Society guidelines recommend an initial 2-month induction therapy with isoniazid, rifampin, pyrazinamide and ethambutol, followed by 7-to 10-month additional isoniazid and rifampin for an isolate that is sensitive to these drugs. Recent systemic review suggests that 6 months' regime might be sufficient if the likelihood of drug resistance is low. 
In 2008, an estimated 440,000 cases of multidrug-resistant (resistant to both isoniazid and rifampin) tuberculosis emerged globally, with India and China together accounting for almost 50% of the total cases worldwide.  Further, cases of extensively drug-resistant (EDR) tuberculosis (resistant to both isoniazid and rifampin plus resistant to a fluoroquinolone and injectable second-line drugs) have been reported. , In high-burden countries, the proportion of tuberculosis cases that are multidrug-resistant (MDR) may range from 1% to 14% or more.  Thus the probability of a patient with TBM in high-burden countries having MDR tuberculosis would be 0.1% to 1.4%. It would be extremely difficult clinically to suspect MDR TBM. In-hospital case-fatality rate was 57% in patients with MDR tuberculous meningitis, with a significant functional impairment in most of the survivors.  The mortality was near 90% in patients with HIV-associated MDR tuberculous meningitis.  In a prospective study of Vietnamese adults with TBM, isoniazid and/or streptomycin resistance was associated with slower CSF bacterial clearance but not with any difference in clinical response or outcome. However, combined isoniazid and rifampicin resistance was strongly predictive of death.  MDR tuberculosis requires extended treatment with second-line drugs that are less effective and have more adverse effects than isoniazid-based and rifampin-based regimens.  With the emergence of EDR tuberculosis, even the second-line drugs will be ineffective. Ethinomide and cycloserine have good CNS penetration and may be used as part of "intensive-phase" treatment regimen in patients with suspected MDR TBM.  TMC207, an investigational diarylquinoline compound, acts by specifically inhibiting mycobacterial ATP-synthase and inhibits drug-sensitive and drug-resistant M. tuberculosis isolates and is also bactericidal against dormant tubercle bacilli. ,,, In patients with newly diagnosed, smear-positive pulmonary infection caused by MDR M. tuberculosis, addition of TMC207 to standard therapy was found to reduce the time to conversion to a negative sputum culture as compared with placebo and to increase the proportion of patients with conversion of sputum culture. The drug was well tolerated except for significant nausea. 
Adjunctive steroids therapy
Cochrane systematic review concluded that overall adjunctive therapy with corticosteroids reduces the risk of death (relative risk (RR), 0.78). Data on disabling residual neurological deficits from three trials showed that corticosteroids reduced the risk of death or disabling residual neurologic deficit (RR, 0.82). The review recommends routine use of corticosteroids in HIV-negative people with TBM to reduce incidence of death and disabling neurological deficits amongst survivors. Corticosteroids should be used irrespective of patients' age and stage of the disease.  Recent Vietnam adult study of adjunctive dexamethasone therapy in TBM demonstrated a significant reduction in mortality but not in morbidity.  Further subgroup analysis revealed that this benefit occurred among all patients with severe grades of CNS tuberculosis, and this benefit was not seen in patients with HIV co-infection. This study also found that treatment with dexamethasone was associated with less severe adverse events, particularly in hepatitis.
Treatment of complications
In TBM, potential complications include associated elevated intracranial pressure (eICP), hydroencephalus, vasculitis, acute seizures, and hyponatremia. Aggressive and appropriate treatment of these complications can minimize the secondary brain injury and improve the chance of a good outcome. 
The frequency of fever in TBM has been reported to vary between 60% and 95%.  Despite considerable research, whether infection-related fever is globally beneficial or harmful remains unclear.  Fever exacerbates the degree of resulting neuronal injury in the presence of acute brain insult ,, and also raises ICP.  No data exist to determine the effect of fever on the pathology and ICP in TBM. Even in the absence of convincing data, achieving normothermia might be justified in patients with stage II and III TBM.  However, one has to exercise caution in patients with associated sepsis. Low or normal temperature during bacteremia has been shown to be associated with poor outcome.  Standard fever-management consists of antipyretic drug therapy and external/ physical cooling. Newer methods, viz., surface-cooling and intravascular-cooling devices, are more effective in decreasing fever than standard fever-management protocols. 
Disturbances of sodium, intravascular volume, and water are common in TBM. Hyponatremia occurs in 35% to 65% of patients with TBM. ,, In patients with TBM, hyponatremia is an independent predictor of death or severe disability.  The differential diagnosis includes central salt-wasting syndrome (CSWS), syndrome of inappropriate secretion of antidiuretic hormone (SIADH), and adrenal insufficiency. The available evidence suggests that the cause of hyponatremia in TBM is CSWS. ,,,,, CSWS involves renal salt loss resulting in hyponatremia and hypovolemia, whereas SIADH involves physiologically inappropriate secretion of antidiuretic hormone (ADH) or increased renal sensitivity to ADH, leading to renal conservation of water and euvolemic or hypervolemic hyponatremia. 
At presentation, many of the patients with TBM have compromised volume status. The first step is to assess the volume status and replace the volume with normal saline and simultaneously investigate for hyponatremia. The therapy in CSWS is volume restriction and sodium replacement (0.9% sodium chloride or 3% if necessary). Treatment of hyponatremia developing at a rate of ≥0.5 mmol/L/h should be aggressive, as it is a life-threatening complication.  Mineralocorticoid, fludrocortisone supplementation has also been shown to be effective in returning serum sodium levels to normal. , Volume restriction is the treatment in SIADH and in patients with symptomatic hyponatremia; 3% sodium chloride is usually combined with frusemide to facilitate free water excretion and correct hyponatremia.
Acute seizures occur in about 50% of children and in 5% of adults.  Rarely status epilepticus (SE), convulsive (CSE)  and non-convulsive (NCSE),  may complicate TBM. In patients with CNS infections, after the first acute seizure, recurrent seizures are common  ; thus probably these patients need antiepileptic drug (AED) prophylaxis to prevent seizure recurrence, at least for the period of resolution or stabilization of acute CNS insult. The treatment strategy in such patients would be acute abortive treatment with benzodiazepines, followed by loading dose of phenytoin/ fosphenytoin and subsequent maintenance therapy. The other alternative drug is valproate or levetericetam. ,, AEDs may be continued if there is high risk of recurrence for a period of 3 to 6 months.  While using AEDs, interactions with other co-medications, particularly anituberculous drugs, should be considered. , When combined with isoniazid, rifampicin counters the former's inhibitory effect on the metabolism of phenytoin. Isoniazid, rifampin, pyrazinamide and valproic acid are all hepatotoxic drugs; and when used together, they may potentiate hepatotoxicity. Preferably, valproic acid should be avoided; and if given, liver functions should be monitored at regular intervals.
Vascular pathologies associated with TBM, viz., arteritis, arterial spasm, intraluminal thrombus, and external compression of proximal vessels by the exudates in the basal cisterns, compromise cerebral perfusion and oxygen delivery to the brain. ,, Arteritis mostly involves the perforating branches of the major arteries at the base of the brain. , It is not clear how to treat this serious complication of tuberculous meningitis and also the compromised cerebral perfusion and infarction. Corticosteroids may be beneficial, probably due to their anti-inflammatory effect.  The Vietnam adult study suggests that probably dexamethasone might improve survival from tuberculous meningitis by reducing the incidence of infarction and speeding up the resolution of hydrocephalus.  Corticosteroids might antagonize vascular endothelial growth factor β and thereby reduce vasogenic cerebral edema. Gujjar et al. studied the efficacy of triple-H therapy in patients with tuberculous arteritis and suggested that triple-H therapy is safe and may be beneficial in tuberculous arteritis.
In patients with TBM, eICP is one of the predictors of poor outcome. The relative risk of poor outcome in children with clinical features of eICP was reported to be 1.7 (95% CI, 1.7-2.2; P=.002).  Presence of hydrocephalus usually signifies eICP; and in patients with hydrocephalus, the stage of the disease at admission is the predictor of poor outcomes. ,,, The pathological substrate of eICP includes (1) diffuse edema consequent to encephalitic process; (2) infarcts, micro and macro, secondary to vasculitis of both small and large vessels and the associated space-occupying effect edema; (3) hydrocephalus; (4) space-occupying effect of associated tuberculoma(s). , Other players in the pathogenesis of eICP in TBM include fever and hyponatremia. ,
Clinical presence of papilledema may help to diagnose eICP. Glasgow Coma Scale (GCS) is reliable in assessing severity of brain injury. Any GCS score less than 8 suggests serious pathology and possible eICP. In addition, neuroimaging provides a good idea about the possible pathological substrate of eICP and helps in identifying tuberculomas and space-occupying infarcts, hydrocephalus and severity of cerebral edema or the presence of brain shift. There are no established guidelines for when to institute ICP monitoring in patients who have tuberculous meningitis with eICP. It will be appropriate to monitor ICP in patients with tuberculous meningitis with feature of ICP and grade II and III disease with no pathology that requires surgery.
Management of eICP should be carried out in a stepwise fashion like in any other clinical setting. The administration of osmotic agents is one of the principal strategies to lower eICP, particularly in patients with no pathology that requires surgery. The commonly used osmotic agents are mannitol and hypertonic saline. ,,,, But none of the studies have systematically evaluated the efficacy of the osmotic agents. Similarly there is hardly any study reporting treatment of eICP on the basis of an ICP-targeted approach.  The safety and efficacy of hypertonic saline in the treatment of eICP in other clinical settings have been well established. However, caution is advised in cases of high osmolar loads because they carry increased risks for potentially deleterious consequences of hypernatremia or may induce osmotic blood-brain barrier opening with possibly harmful extravasation of the hypertonic solution into the brain tissue.  Hypertonic saline is without the risks of dehydration and tubular damage, as in the case of mannitol.
Hydrocephalus can be treated with diuretics, osmotic agents, serial lumbar punctures, external ventricular drainage or ventriculoperitoneal shunt. Addition of acetazolamide and furosemide was significantly more effective in achieving normal ICP than antituberculous drug treatment alone. , However, patients on medical treatment should be closely monitored to detect any worsening or lack of improvement, and shunt surgery should be considered in case of failure of medical management. In TBM, ventriculoperitoneal shunt is associated with favorable outcome. The stage of the disease at presentation is the predictor of outcome following shunt surgery. ,,, In mild and moderate hydrocephalus, early shunt surgery (2 days after diagnosis) was found to be associated with better outcomes compared to delayed surgery (3 weeks after diagnosis). 
In patients with TBM, associated space-occupying tuberculoma(s) may be the substrate for eICP. Growing evidence suggests that most often tuberculomas resolve with antituberculous treatment.  However, surgical excision is indicated in (1) tuberculoma causing obstructive hydrocephalus and significant eICP; (2) tuberculoma causing obstructive hydrocephalus and not resolving on medical treatment; (3) large space-occupying tuberculomas with eICP; (4) tuberculomas with associated compartmental shifts and not resolving with medical treatment. 
TBM is a serious CNS infection associated with significant mortality and high morbidity among the survivors. , Most factors found to correlate with poor outcome can be directly traced to the stage of the disease at the time of diagnosis. , The only way to reduce mortality and morbidity is by early diagnosis and timely recognition of complications and institution of the appropriate treatment strategies. However, still the most challenging aspect is early diagnosis with certainty, and the diagnosis is hampered by slow and insensitive diagnostic methods. The other major emerging challenge is treating MDR TBM.