Tuberculous Meningitis Workup

Updated: Nov 10, 2021
  • Author: Gaurav Gupta, MD, FAANS, FACS; Chief Editor: Niranjan N Singh, MBBS, MD, DM, FAHS, FAANEM  more...
  • Print

Approach Considerations

The diagnosis of tuberculous meningitis (TBM) cannot be made or excluded on the basis of clinical findings. Tuberculin testing is of limited value. Variable natural history and nonspecific clinical features make the diagnosis of TBM difficult. [21, 22, 23, 24, 25]

Spinal tap carries some risk of brain herniation in any instance when intracranial pressure (ICP) is increased (which may occur in TBM), but if meningitis is suspected, the procedure should generally be performed, using suitable precautions and obtaining informed consent before the procedure. A low-volume tap may be performed to rule out other more common types of meningitis (eg, bacterial), although this may limit diagnostic sensitivity of TBM.

Computed tomography (CT) scanning and magnetic resonance imaging (MRI) lack specificity but may aid in identifying complications, such as hydrocephalus, ischemic or hemorrhagic strokes, and tuberculomas, some of which may require neurosurgical evalauation. (See Imaging in Bacterial Meningitis).

Ziehl-Neelsen staining lacks sensitivity, and culture results are often too late to aid clinical judgment. Semiautomated radiometric culture systems, such as the Bactec 460, and automated continuously monitored systems have reduced culture times. Newer methods involving amplification of bacterial DNA by polymerase chain reaction (PCR) and comparable systems have not been assessed completely and may not be suitable for laboratories in developing countries with limited resources. Interferon-gamma assays may have promising sensitivity and specificity for extra-pulmonary TB (including TBM). [24, 57]

A complete blood count should be performed, and the erythrocyte sedimentation rate should be determined.

A hepatic function panel should be obtained before starting antituberculous medications, many of which are hepatotoxic.

The serum glucose level should be measured; this value is a useful comparison with the glucose level measured in the cerebrospinal fluid (CSF).

HIV testing should be performed, and include CD4 count and viral load if positive. 

Serologic testing for syphilis should be performed. Complementation testing or its equivalent for fungal infections should also be performed.


Serum and Urine Chemistry Studies

Electrolyte concentrations should be assessed. Mild-to-moderate hyponatremia is present in roughly 45% of patients, which may be due to SIADH or CSWS. Blood urea nitrogen (BUN) and creatinine levels should be measured as well. Urinalysis should be performed.


Tuberculin Skin Testing

Despite its many limitations, tuberculin skin testing, by necessity, remains in widespread use. The Centers for Disease Control and Prevention (CDC), the American Thoracic Society, and the Infectious Disease Society of America have updated the guidelines, and they are quite useful in practice. [26]

These guidelines stress that in general, one should not obtain a tuberculin skin test unless treatment would be offered in the event of a positive test result. Cutoff points for induration (5, 10, or 15 mm) for determining a positive test result vary based on the pretest category into which the patient falls. While this approach might decrease the specificity of the test, it increases the sensitivity for capturing those at highest risk for developing the disease in the short term.

Negative results from the purified protein derivative (PPD) test do not rule out tuberculosis (TB); if the 5-tuberculin test skin test result is negative, repeat the test with 250-tuberculin test. Note that this test is often nonreactive in persons with TBM.


Lumbar Puncture

A lumbar puncture should be performed in most cases of suspected tuberculous meningitis (TBM), unless there is significant concern for precipitating herniation (eg, in the case of hydrocephalus, in which case a ventriculostomy may be performed to divert CSF, from which CSF could be obtained). A high-volume LP should be performed (if there is no concern for herniation), as it improves the sensitivity of detecting TB. 

As with any standard lumbar puncture, manometry should be performed to check CSF pressure. The patient may have a normal opening pressure (OP), but in 50% of patients, the opening pressure is > 25cm H2O. [57]

Inspect the CSF visually and note its gross appearance. It typically is clear or slightly turbid. If the CSF is left to stand, a fine clot resembling a pellicle or cobweb may form. This faintly visible "spider's web clot" is due to the very high level of protein in the CSF (1–8 g/L, or 1000–8000 mg/dL) typical of this condition. Hemorrhagic CSF also has been recorded in proven cases of TBM; this is attributed to fibrinoid degeneration of vessels resulting in hemorrhage (Smith, 1947).

CSF analysis

Tests that may be performed on CSF specimens obtained by lumbar puncture include:

  • Cell counts, differential count, cytology

  • Glucose level, with a simultaneous blood glucose level

  • Protein level

  • Acid-fast stain, Gram stain, appropriate bacteriologic culture and sensitivity, India ink stain

  • Cryptococcal antigen and herpes antigen testing

  • Culture for Mycobacterium tuberculosis (50-80% of known cases of TBM yield positive results)

  • PCR: Results imply that PCR can provide a rapid and reliable diagnosis of TBM, although false-negative results potentially occur in samples containing very few organisms (< 2 colony-forming units per mL).

  • Syphilis serology

CSF typically has a pleocytosis, an elevated protein level, and marked hypoglycorrhachia.

In adults, the mean white blood cell (WBC) count averages around 223 cells/µL (range, 0–4000 cells/µL), while the proportion with neutrophilic pleocytosis (> 50% neutrophils) averages 27% (range, 15%–55%) and the proportion with normal cell count averages 6% (range, 5%–15%). In children, these numbers are 200 cells/µL (range, 5–950 cells/µL), 21% (range, 15%–30%), and 3% (range, 1%–5%), respectively.

The mean protein level in adults averages 224 mg/dL (range, 20–1000 mg/dL), and in children it is 219 mg/dL (range, 50–1300 mg/dL). The proportion with a normal protein content averages 6% (range, 0%–15%) for adults and 16% (range, 10%–30%) for children.

The proportion with depressed glucose levels (< 45 mg/dL or 40% of serum glucose) averages 72% (range, 50%–85%) for adults and 77% (range, 65%–85%) for children.

A positive smear result is present in an average of 25% (range, 5%–85%) of adults and only 3% (range, 0%–6%) of children, whereas the numbers with a positive CSF culture average 61% (range, 40%–85%) and 58% (range, 35%–85%) for adults and children, respectively. Failure to respond to treatment should prompt a search for fungal infections or malignancy.

For patients with HIV and/or immunosuppression, while the mean WBC count in the CSF is 230 cells/µL, as many as 16% of HIV-infected patients may have acellular CSF, compared with 3%–6% of HIV-negative patients. Patients whose CSF samples are acellular may show pleocytosis if a spinal tap is repeated 24–48 hours later. The proportion who have neutrophilic pleocytosis of the CSF (>50% neutrophils) is 42% (range, 30%–55%).

While HIV-infected patients generally have a mean protein level of 125 mg/dL (range, 50–200 mg/dL), as many as 43% of these patients may have a normal CSF protein content. The proportion who have depressed CSF glucose levels (< 45 mg/dL or 40% of serum glucose) averages around 69% (range, 50%–85%). The number who have a positive CSF culture results averages 23%.

Within a few days after commencement of anti-TB therapy, the initial mononuclear pleocytosis may change briefly in some patients to one of polymorphonuclear predominance, which may be associated with clinical deterioration, coma, or even death. This therapeutic paradox has been regarded by some authors as virtually pathognomonic of TBM. This syndrome is probably the result of an uncommon hypersensitivity reaction to the massive release of tuberculoproteins into the subarachnoid space.

In patients with tuberculous radiculomyelitis (TBRM), CSF evaluation usually demonstrates an active inflammatory response with a very high protein level.

When CSF analysis offers no clues and the diagnosis remains elusive, a brain biopsy may be warranted under appropriate circumstances. Depending on the location of the brain to be biopsied, there may be significant risks, including hemorrhage, surgical site infection, stroke, and seizure. 


Dot-Immunobinding Assay

A dot-immunobinding assay (Dot-Iba) has been standardized to measure circulating antimycobacterial antibodies in CSF specimens for the rapid laboratory diagnosis of tuberculous meningitis (TBM). [27]  Specific CSF immunoglobulin G antibody to M. tuberculosis from a patient with culture-proven TBM was isolated and coupled with activated cyanogen bromide-Sepharose 4B. A 14-kd antigen present in the culture filtrates of M. tuberculosis was isolated by immunosorbent affinity chromatography and used in the Dot-Iba to quantitate specific antimycobacterial antibodies. The Dot-Iba gave positive results in all 5 patients with culture-proven TBM; no false-positive results were obtained from CSF specimens from patients with partially treated pyogenic meningitis. In the opinion of Sumi et al, the Dot-Iba developed in their laboratory is a simple, rapid, and specific method and, more importantly, is suited for routine application in laboratories with limited resources. [27] This is not yet available for routine use, and proof of its utility requires further studies.

Researchers evaluated previous multicenter/multinational studies to determine the frequency of the absence of CSF pleocytosis in patients with central nervous system infections, as well as the clinical impact of this condition. It was found that 3% of TBM cases did not display CSF pleocytosis. Most patients were not immunosuppressed, and patients without pleocytosis had a high rate of unfavorable outcomes. [28]

Sun et al describe a novel, highly sensitive molecular diagnostic method, one-tube nested PCR-lateral flow strip test (OTNPCR-LFST), for detecting Mtuberculosis. This one-tube nested PCR offers improved sensitivity compared with traditional PCR; the limit of detection was up to 1 fg DNA isolated from M. tuberculosis. Since this assay is specific for Mtuberculosis, it reduces both the chance of cross-contamination and the time required for analysis. The PCR product was detected by a lateral flow strip assay, which provided a basis for migration of the test to a point-of-care (POC) microfluidic format. It has shown 89% overall sensitivity and 100% specificity for TBM patients. In TBM, where rapid and sensitive detection of Mtuberculosis in CSF is crucial in diagnosis and treatment, this one-tube nested PCR-lateral flow strip assay offers hope due to its rapidity, high sensitivity, and simple manipulation. [29]


Chest Radiography

Chest radiography posteroanterior and lateral views may reveal hilar lymphadenopathy, simple pneumonia, infiltrate, fibronodular infiltrate/cavitation, and/or pleural effusion/pleural scar.

Go to Imaging in Bacterial Meningitis for more complete information on this topic.


Brain and Spinal Imaging

Neuroimaging should be obtained in all patients with suspected tuberculous meningitis (TBM). This includes a CT head with and without contrast, and MRI brain with and without gadolinium contrast (remember to check renal function). CT scanning and MRI of the brain reveal hydrocephalus, basilar meningeal thickening, infarcts, edema, and tuberculomas (see the image below). Although they lack specificity, they may suggest the diagnosis, and help in monitoring complications that require neurosurgical intervention.

MRI of the brain in a patient with TBM and concurr MRI of the brain in a patient with TBM and concurrent AIDS (with 8 CD4 cells/mL). The patient's history includes previous interstitial pneumonia, pericarditis, adnexitis, and a positive result on the Mantoux test. His recent history includes fever, headache, strabismus, diplopia, and cough. Laboratory studies revealed hyponatremia. CSF findings strongly suggested a diagnosis of tuberculous meningitis, and culture results were positive for Mycobacterium tuberculosis. The MRI shows the presence of exudates, in and over the sellar seat, with parasellar extension to the left, with irregular margins, marked heterogenous enhancement, and compression of the optic chiasm and third ventricle. Presence of nodular areas with marked enhancement of basal cisterns is an expression of basilar leptomeningeal involvement. This patient died after 2 months of inadequate antituberculosis therapy (due to poor compliance). Courtesy of Salvatore Marra, AIDS Imaging (

Hydrocephalus may be present (on non-contrast and contrasted studies). 

Unlike bacterial meningitis, which tends to occur at the convexities, TBM tends to affect the skull base. As a result, in TBM, the basal cisterns often enhance strikingly, corresponding to the thick tubercular exudate that is observed on pathological examination. In particular, the quadrigeminal cistern, interpeduncular fossa, ambient cistern, and chiasmatic region are frequently involved, with associated arachnoiditis. Meningeal enhancement may occur, although it is more common among HIV-infected patients.

Contrast administration may show focal enhancement of parenchymal and space-occupying lesions.

Intracranial tuberculomas may also be viasualized. The characteristic CT finding is a nodular, enhancing lesion with a central hypodense lesion. [30]  Early stages are characterized by low-density or isodense lesions, often with edema out of proportion to the mass effect, and with little encapsulation. At a later stage, well-encapsulated tuberculomas appear as isodense or hyperdense lesions with peripheral ring enhancement. On MRI, characteristics depend on the maturity of the tuberculoma. In general, non-caseating tuberculomas as T1 hypointense and T2 hyperintense, and homogenously enhance. Caseating tuberculomas are generally hypo- to isointense on T1 and T2, and have rim enhancement. Tuberculous abscesses, which are larger than tuberculomas, may also be seen. On MR spectroscopy, tuberculous lesions generally have raised lipid peaks, which helps distinguish them from non-tuberculous lesions. [57]

As TBM may result in vasculitis, vascular imaging (eg, CT angiogram, formal angiogram), may be considered to assess patency of intracranial large vessels, particularly if a TBM-related stroke is suspected. MRI would show diffusion restriction on DWI sequence in the setting of acute stroke. 

Srikanth et al concluded that CT features of TBM in elderly patients were few, atypical, and noncontributory for diagnosis, probably because of age-related immune senescence. [31] Hence, strong clinical suspicion and correlation with laboratory findings is necessary for early diagnosis.

Skull radiography may reveal evidence of increased intracranial pressure in children, in the form of sutural diastasis. During follow-up of patients with TBM, intracranial calcifications may be evident.

Calcifications may occur in two main sites, (1) more commonly in the basal meninges and, (2) to a lesser extent, within brain parenchyma. Calcifications are generally in the sellar region, either as a single lesion or as a cluster of small calcifications. These calcifications sometimes harbor tubercle bacilli, which may be responsible for a relapse of the disease.

For tuberculous spinal meningitis, MRI shows that the subarachnoid space is obliterated, with focal or diffusely increased intramedullary signal on T2-weighted images and variable degrees of edema and mass effect. Most spinal cord lesions appear hyperintense on T2-weighted images and isointense or hypointense on T1-weighted images. MRI findings in patients with spinal cord TB have both diagnostic and prognostic significance. Cord atrophy or cavitation and the presence of syrinx on MRIs may be associated with a poor outcome. [1]  With gadolinium administration, contrast enhancement is often seen surrounding the spinal cord and the nerve roots. The nerve roots may appear clumped and show contrast enhancement, secondary to inflammation and edema, depending on the degree of involvement.

Rarely, tuberculomas occur in the spinal cord, and they may occur on the surface of the cord, as dural lesions, or deep inside in an intramedullary location. Less frequently, intramedullary tuberculous abscesses have been reported.

On neuroimaging, tuberculous spondylitis (Pott's disease) generally reveals bony destruction with relative sparing of the disc space, unlike pyogenic infections. In addition, the involvement of the posterior elements is more common, and the paraspinal abscess, if present, tends to be much larger relative to bony involvement. An associated psoas abscess may be present. The infection tends to begin anteriorly, extend under the anterior longitudinal ligament, and spread hematogenously via the venous plexus of Bateson. MRI has a sensitivity of 94% in vertebral osteomyelitis. It reveals hypointense T1-weighted areas in the vertebral bodies, alternating with areas of hyperintense T2-weighted signal in the disk spaces and the paravertebral soft tissue. Infected bone and disk often reveal contrast enhancement. A psoas abscess with calcifications, which is better detected on CT rather than MRI, strongly raises suspicion for a tuberculous etiology. Epidural deposits are best shown by MRI, which reveals a soft-tissue mass that is isointense to hypointense compared with the spinal cord on T1-weighted images and hyperintense on proton-density and T2-weighted images and has variable degrees of contrast enhancement.

Tuberculous myelitis and tuberculous radiculomyelitis (TBRM) are predominantly diseases of the thoracic spinal cord. MRI and CT scanning are critical for the diagnosis of TBRM, revealing loculation and obliteration of the subarachnoid space along with linear intradural enhancement.

Diffusion tensor imaging (DTI)-derived anisotropy has been shown to demonstrate meningeal inflammation and this could be a valuable tool to assess the response to antituberculous therapy, in addition to the standard neuroimaging techniques. [32]

Go to Imaging in Bacterial Meningitis for more complete information on this topic.

Additional imaging may be seen at 



As tuberculous meningitis (TBM) may result in vasculopathy, vascular imaging (eg, CT angiogram, MR angiogram, formal catheter angiogram) may be considered to assess patency of large vessels, particularly if a TBM-related stroke is suspected. TBM vasculitis affects up to one third of cases, and is more common in pediatric cases. Vascular imaging may reveal arterial narrowing or occlusion. 



Seizures affect 16.3% to 31.5% of tuberculous meningitis (TBM) patients, may be focal or generalized, and occur more frequently in children and in those with HIV, and worsen mortality and disability. [60]  Surviving patients who had TBM-associated seizures may develop chronic epilepsy. Unlike many other forms of meningitis, there are more electroencephalography (EEG) abnormalities in TBM. [60]  EEG changes vary depending on the site of the ongoing inflammatory process. There may be diffuse slowing with or without focal changes and epileptiform discharged. [60]  In one study, findings from EEG were abnormal in 24 patients. EEG abnormalities included diffuse theta-to-delta slowing in 22 patients, intermittent rhythmic delta activity in the frontal region in 15 patients, right-to-left asymmetry in 5 patients, and epileptiform discharges in 4 patients. At the end of 3 months, 5 patients had died, while recovery was poor in 13 patients, partial in 3, and complete in 11. EEG findings correlated with severity of meningitis and degree of coma; outcome at 3 months was assessed using the Barthel index score.


Brainstem Auditory Evoked Response Testing

Brainstem auditory evoked responses (BAERs) have been observed in more than 50% of patients with tuberculous meningitis (TBM). Motor and somatosensory evoked potentials may be helpful in objective documentation of respective motor and sensory functions in patients with TBM and altered sensorium.


Use of Neurochemical Markers

Use of neurochemical markers has been investigated in patients with aseptic meningitis or tuberculous meningitis (TBM). CSF levels of amino acids, nitrite (a metabolite of nitric oxide), vitamin B-12, and homocysteine were quantitated in both groups of patients. Levels of excitatory amino acids aspartic acid and glutamic acid, gamma-aminobutyric acid (GABA), glycine, and tryptophan all were increased significantly in both groups, whereas levels of taurine were decreased and levels of phenylalanine were increased only in patients with TBM. Levels of nitrite and its precursor arginine were significantly higher in patients with TBM, whereas they were unchanged in patients with aseptic meningitis. Levels of homocysteine were increased significantly, and levels of vitamin B-12 decreased only in patients with TBM, whereas these levels were unchanged in patients with aseptic meningitis. This indicates that patients with TBM are particularly prone to vitamin B-12 deficiency, resulting in increased levels of homocysteine and free radicals, showing the potential importance of these biological markers in the development and design of therapeutic approaches.

Janvier et al report that once purulent bacterial meningitis and cryptococcosis have been ruled out, adenosine deaminase activity measurement could be an inexpensive, valuable tool in the diagnosis of early tuberculous meningitis. [33]


Histologic Findings

The Ziehl-Neelsen stain uses the properties of the cell wall to form a complex that prevents decolorization by acid or alcohol. Fluorochrome tissue stains also can be helpful in the diagnosis of TBM (see the image below).

Fluorochrome for tuberculosis. Fluorescent stainin Fluorochrome for tuberculosis. Fluorescent staining procedures are used with auramine O or auramine-rhodamine as the primary fluorochrome dye. After decolorization with an acid-alcohol preparation, the smear is counterstained with acridine orange or thiazine red and scanned at a lower magnification with a 25X dry objective fluorescent microscope. Acid-fast bacilli appear as yellow-green fluorescent thin rods against a dark background. Courtesy of Robert Schelper, MD, Associate Professor of Pathology, State University of New York Upstate Medical University.
Hematoxylin and eosin stain showing caseation in t Hematoxylin and eosin stain showing caseation in tuberculosis. Courtesy of Robert Schelper, MD, Associate Professor of Pathology, State University of New York Upstate Medical University.

Clinically silent single or multiple enhancing granulomata are observed in a significant minority of cases of TBM and in some cases of miliary TB without meningitis. [34]