Approach to the patient with chronic meningitis

INTRODUCTION — Meningitis can be classified on the basis of its underlying cause, type of inflammatory response, or the time course of the illness. Based upon the time course, meningitis is defined as acute or chronic. The onset of symptoms of acute meningitis typically is abrupt with progression over hours, but some patients become ill over several days, especially when early symptoms are not appreciated. (See "Clinical features and diagnosis of acute bacterial meningitis in adults".)

Acute meningitis due to infectious causes usually does not recur. However, a small number of patients with acute meningitis may develop recurrent attacks between intervals of good health. (See "Approach to the adult with recurrent infections", section on 'Meningitis'.)

The differential diagnosis, clinical evaluation, prognosis, and management strategies for patients with chronic meningitis, including those in whom a diagnosis cannot be established with routine tests, will be reviewed here. Individual conditions associated with chronic meningitis are discussed separately. (See related topics.)

DEFINITION — Chronic meningitis is arbitrarily defined as meningitis lasting for four weeks or more and is a complex entity with both infectious and noninfectious causes [1-7]. Although some authors attempting to define the syndrome of chronic meningitis have excluded patients with meningitis occurring concurrently with mass lesions of the central nervous system (CNS), meningitis associated with previously diagnosed systemic diseases known to cause meningitis, and meningitis following neurosurgical procedures, this definition may be too restrictive for clinical purposes [8]. Patients with chronic meningitis usually have a subacute onset of symptoms including fever, headache, and vomiting. The symptoms can remain static, fluctuate, and/or slowly worsen. The symptoms and clinical course of chronic meningitis vary widely from patient to patient.

LIMITATIONS OF THE LITERATURE — Because of its rarity and the diversity of causes, the literature on chronic meningitis consists largely of case reports and a small number of retrospective case series from single centers. Selection, publication, and ascertainment biases greatly limit the utility of this literature for assessing the relative proportions of patients with any given etiologic diagnosis and for assessing the natural history of individual patients with chronic meningitis.

Most of these case series describe patients from referral centers or geographic areas where selection bias was likely because of unique patient populations or because of local referral patterns. For example, one case series summarizing the clinical features of 83 patients with chronic meningitis collected over a 16-year period in a referral hospital in New Zealand reported that 40 percent of patients had tuberculous (TB) meningitis and that an additional 17 percent of patients responded to empiric antituberculous therapy [9]. However, most of the patients with TB described in this report were Maori people or Pacific Islanders.

In another report describing 114 consecutive patients with chronic meningitis admitted to a Bangkok hospital, TB was responsible for 37 percent of cases, and Cryptococcus neoformans was the etiologic agent in 54 percent [10]. A high proportion of these patients had HIV infection and the presence or absence of HIV infection had a major impact on the likelihood that a specific pathogen would be found. Similar problems related to selection or ascertainment bias make it difficult to interpret reports concerning the utility of various diagnostic tests in patients with chronic meningitis [11].

ETIOLOGIES — An array of infectious agents can present as chronic meningitis, but a nearly identical syndrome can result from a number of inflammatory, malignant, or other noninfectious diseases (table 1) [5,6,12,13]. Despite extensive testing, an etiologic diagnosis may not be determined in up to one-third of all patients.

The initial evaluation of patients with chronic meningitis is often complex and difficult. In some cases, the diagnosis is first suspected and later confirmed because of subtle historical or epidemiologic clues or because of associated clinical findings that at first glance may seem to be incidental. For example, the presence of iritis or uveitis or the finding of localized cutaneous lesions may provide a crucial clue to the presence of diseases, such as sarcoidosis, Behçet syndrome, granulomatosis with polyangiitis, uveo-meningitis syndromes, or a systemic fungal infection that initially manifested as chronic meningitis. Similarly, the simultaneous presence of deafness and optic neuritis and uveitis are strong clues to a diagnosis of Vogt-Koyanagi-Harada syndrome.

Patients with Mollaret's meningitis, a form of benign recurrent aseptic meningitis, may be mistakenly thought of having chronic meningitis when recurrent attacks occur frequently. (See "Aseptic meningitis in adults", section on 'Recurrent (Mollaret) meningitis'.)

Similarly, patients with viral or postinfectious encephalitis (also known as acute disseminated encephalomyelitis) and a small percentage of patients with partially treated bacterial meningitis may have symptoms lasting for more than a month and thus may be erroneously considered to have chronic meningitis. (See "Viral encephalitis in adults", section on 'Viral versus postinfectious encephalitis'.)

CLINICAL FEATURES — The clinical symptoms of patients with chronic meningitis rarely point to a specific etiologic diagnosis. However, occasionally, a historical or epidemiologic clue can lead to the discovery of an otherwise obscure diagnosis (table 2).

History — It is important to elicit a history of possible exposure to a pathogen associated with chronic meningitis. As an example, a history of travel to the southwestern United States or southern California can be a crucial clue to the presence of coccidioidomycosis in a patient with chronic meningitis residing in a nonendemic area. Sometimes the history of travel may be distant, forgotten, or deemed to be inconsequential. Patients with coccidioidomycosis may acquire their infection after short stays in airports in endemic areas, and such information may be missed unless a meticulous history is obtained. Thus, all patients with chronic meningitis should be questioned about travel or residence in geographic areas known to be endemic for coccidioidomycosis, histoplasmosis, paracoccidioidomycosis, blastomycosis, schistosomiasis, trypanosomiasis, Angiostrongylus cantonensis infection, or cysticercosis. (See "Coccidioidal meningitis" and "Pathogenesis and clinical manifestations of disseminated histoplasmosis".)

A past history of a positive tuberculin skin test or an interferon-gamma release assay, current or prior treatment with a tumor necrosis factor-alpha inhibitor, or of a known exposure to tuberculosis (TB) may be enough justification to institute empiric therapy for tuberculous meningitis or to pursue this diagnosis further with special cultures or testing based upon polymerase chain reaction (PCR) technology, especially if preexisting infection with HIV is present. (See "Central nervous system tuberculosis: An overview".)

The immune status of all patients with chronic meningitis should be assessed by both history and laboratory testing. Patients should be questioned about risk factors for both HIV and human T-lymphotropic virus I and II (eg, a history of injection drug use, sexual promiscuity) and about possible or known exposure to syphilis. Lyme disease is a rare cause of chronic meningitis, but patients should be asked about a past history of skin lesions typical or suggestive of erythema migrans or travel to or contact with ticks in an endemic area for Borrelia burgdorferi infection.

A careful review of all recent medications including the use of immunosuppressive drugs, immunoglobulin therapy, and nonsteroidal anti-inflammatory agents should be undertaken. Patients should be questioned about contact with animals, including cats and wild game, or meat processing. The presence or absence of systemic symptoms suggesting a vasculitic disorder, such as granulomatosis with polyangiitis, systemic lupus erythematosus, or Behçet syndrome, should be sought. For example, a history of iritis, recurrent genital or oral ulcers, or of nasal inflammation may lead to a diagnosis of Behçet syndrome.

The presence of systemic symptoms such as weight loss, unexplained cough, or night sweats may be difficult to assess, since such symptoms can occur as a consequence of chronic meningitis or be related to an underlying malignancy that is the primary cause of the inflammatory reaction in the central nervous system (CNS).

A history of invasive spinal procedures, such as glucocorticoid injections for spinal pain, can be an important diagnostic clue to the presence of fungal meningitis due to contaminated glucocorticoids. (See "Central nervous system infections due to dematiaceous fungi (cerebral phaeohyphomycosis)", section on 'Outbreaks'.)

Careful questioning about the onset of symptoms is important, as it can help to determine if symptoms are intermittent (ie, more compatible with recurrent meningitis) or continuous. This determination occasionally can be difficult if symptoms in patients with chronic meningitis fluctuate over time.

DIAGNOSIS — Patients with chronic meningitis typically undergo an array of complex diagnostic investigations including serologic assays, multiple imaging tests, biopsy of skin lesions and lymph nodes, and repeated lumbar punctures (table 3). The decision regarding which laboratory tests to perform should be based upon the clinical features of each individual patient and the probability that a specific disease is present. For example, routine evaluation of patients with chronic meningitis usually includes a tuberculin skin test, a chest radiograph, and serologic testing for syphilis as well as testing for the presence of HIV infection. However, tests for other infectious diseases such as Lyme disease, brucellosis, cysticercosis, trypanosomiasis, and/or schistosomiasis are only necessary in patients in whom there is reasonable pretest probability that the disease or condition is present.

The complexity of the diagnostic workup is illustrated by a report summarizing the diagnostic testing of 37 consecutive patients with chronic idiopathic meningitis evaluated at the Mayo Clinic [14]. Despite a total of 2295 tests of 44 different types performed on the cerebrospinal fluid (CSF) of these 37 patients, a diagnosis was rarely made.

CSF examination and other laboratory testing — Analysis of CSF reveals abnormalities in patients with chronic meningitis, but these abnormalities are rarely diagnostic with some notable exceptions. The presence of eosinophilia can provide an important clue to the presence of a parasitic etiology or coccidioidomycosis (see "Eosinophilic meningitis" and "Coccidioidal meningitis"). Similarly, stained smears of a centrifuged sample of the CSF may occasionally reveal infectious agents, such as fungi or bacteria, and thus lead to a specific etiologic diagnosis.

Antigen testing of the CSF for the presence of C. neoformans and a Venereal Disease Research Laboratory test for syphilis should be performed on all patients with chronic meningitis. The sensitivity and specificity of testing for cryptococcal meningitis is extremely high [5]; however, capsule-deficient strains of C. neoformans can rarely cause a false-negative antigen test [15]. More detailed discussions of how to diagnose cryptococcal meningitis and neurosyphilis are presented in separate topic reviews. (See "Clinical manifestations and diagnosis of Cryptococcus neoformans meningoencephalitis in patients without HIV" and "Neurosyphilis", section on 'Diagnosis'.)

Most patients with chronic meningitis have a predominance of lymphocytes, although a small percentage of patients have a predominance of neutrophils [16]. The differential diagnosis of chronic meningitis with a neutrophil predominance includes nocardiosis, brucellosis, and the endemic mycoses (table 4). Nontuberculous mycobacteria can also cause chronic meningitis associated with a predominantly neutrophilic response in some cases and with a lymphocytic response in others. Acid-fast bacilli are usually not detectable in stains of CSF in such patients. Occasionally, gram-positive rods may be visualized in the CSF of patients with chronic meningitis due to nontuberculous mycobacteria [17]. Most of the case reports describing chronic meningitis due to nontuberculous mycobacteria involved patients with a prior history of trauma or neurosurgery but occasionally central nervous system (CNS) infections have arisen secondary to otitis media or secondary to disseminated infections such as endocarditis [17]. (See "Central nervous system tuberculosis: An overview".)

A sample of CSF should be submitted for cultures using media appropriate for growth of aerobic bacteria, mycobacteria, and fungi. If routine cultures are negative, it is sometimes useful to send 10 to 20 mL of CSF to the microbiology laboratory with instructions to culture the entire sample on appropriate media. This may be particularly useful if fungal meningitis is suspected (eg, in neurosurgical or immunocompromised patients), since the organism may be present in low numbers and the yield of standard cultures of CSF is poor. Polymerase chain reaction (PCR) testing of a sample of CSF using primers specific for Mycobacterium tuberculosis may be useful in some cases, and is discussed in detail elsewhere. (See "Tuberculous meningitis: Clinical manifestations and diagnosis", section on 'Nucleic acid amplification tests'.)

If a fungal pathogen is suspected, cultures of blood should be submitted to the laboratory along with notification that fungi are possible pathogens so that appropriate media and techniques can be utilized. For example, isolator blood cultures that utilize lysis centrifugation may be useful on rare occasions in patients with chronic meningitis due to an underlying fungal or mycobacterial infection. If brucellosis is suspected on the basis of epidemiologic clues such as exposure to unpasteurized milk or milk products, blood and CSF should be cultured using media incubated in a high concentration of carbon dioxide for three weeks, and serum should be tested for the presence of anti-Brucella antibodies. Whipple's disease has been rarely associated with chronic meningitis and, in such cases, PCR testing of CSF has been diagnostic [18]. (See "Whipple's disease", section on 'Evaluation for CNS involvement'.)

If routine tests fail to disclose a likely pathogen or a likely cause for the meningitis, at least 20 to 30 mL of CSF should also be sent for cytologic examination, flow cytometry, and immunoglobulin heavy-chain rearrangement studies by PCR. (See "Primary central nervous system lymphoma: Clinical features, diagnosis, and extent of disease evaluation", section on 'CSF analysis'.)

Serum antibody tests for fungal pathogens may not accurately detect the presence of a fungal pathogen in the CSF. Some patients with meningitis due to Histoplasma capsulatum, for example, do not have detectable antibody titers in the serum, yet antibodies against H. capsulatum or antigen can be detected in the CSF [19,20]. Similarly, detection of meningitis due to Sporothrix may not be possible unless specific serologic testing for sporotrichosis is performed on the CSF [21].

1,3-Beta-D-glucan, a cell wall component of many fungi, is detected by the beta-D-glucan assay. The serum beta-D-glucan assay may be positive in patients with a variety of invasive fungal infections, including invasive candidiasis. The beta-D-glucan assay using CSF might be a useful adjunct to CSF cultures in patients with chronic Candida meningitis [22]. The beta-D-glucan assay using CSF also appeared to be useful both for diagnosing and monitoring response to treatment in the large outbreak of fungal meningitis due to contaminated glucocorticoids; Exserohilum rostratum was the predominant pathogen in the outbreak [23,24]. (See "Clinical manifestations and diagnosis of candidemia and invasive candidiasis in adults", section on 'Beta-D-glucan assay' and "Candida infections of the central nervous system", section on 'CSF analysis' and "Central nervous system infections due to dematiaceous fungi (cerebral phaeohyphomycosis)", section on 'Outbreak of fungal meningitis and osteoarticular infections'.)

Testing of serum for antibodies to Brucella, Ehrlichia/Anaplasma, Mycoplasma pneumoniae [25], and Toxoplasma gondii can rarely yield a diagnosis in a patient with chronic meningitis, but a single positive antibody test may not be diagnostic of active infection.

Metagenomic next-generation sequencing, where available, may identify the microbial nucleotide sequences after extracting DNA from CSF samples [26]. The use of 16s rRNA PCR technology on CSF samples may also be useful, particularly in patients who have received empiric antibiotic therapy [27]. However, such testing may also produce false-positive results due to contaminants. As an example, a study evaluating the utility of metagenomic sequencing of CSF specimens from 80 patients with chronic meningitis found that only 15 percent of these samples yielded any organisms, and in half of these cases the detected organism was considered to be inconsistent with the clinical presentation [28].

Although CSF obtained from the lumbar space is usually adequate to diagnose most causes of meningitis, obtaining spinal fluid from the ventricular space for the examination may be useful in rare cases (eg, patients with ventriculitis) [15,29].

Imaging studies — Magnetic resonance imaging (MRI) and computed tomography (CT) are useful in the evaluation of patients with chronic meningitis, but these imaging techniques lead to a specific diagnosis in a minority of patients. However, such imaging is important since it can exclude important abnormalities such as a parameningeal focus of infection, abscess, or a tumor. Rarely, such imaging may disclose one of these processes or the cystic changes typical of cysticercosis. Some patients with chronic meningitis have focal or diffuse areas of meningeal enhancement, a finding not normally diagnostic of any specific entity. However, such focal abnormalities, when present, can be useful in selecting the site for brain biopsy if the usual diagnostic testing methods are not positive and if deterioration occurs during follow-up.

Brain imaging can also detect hydrocephalus that may require shunts or other neurosurgical procedures, particularly if patients are symptomatic from increased intracranial pressure and if symptoms persist after empiric or directed medical therapy. MRI of the spine may occasionally be useful in patients with signs or symptoms of spinal cord dysfunction or back pain. Cerebral angiography may be diagnostic in up to two-thirds of patients with granulomatous angiitis of the CNS; however, neither a normal angiogram nor a negative cerebral biopsy can reliably rule out this rare condition [30]. (See "Primary angiitis of the central nervous system in adults".)

Chest radiographs should be obtained in all patients with chronic meningitis. Such radiographs can lead to a diagnosis such as tuberculosis (TB), sarcoidosis, a systemic fungal infection, or a malignancy. Further imaging with a chest CT should be pursued if chest radiographs are negative and suspicion for one of these diagnoses is high.

Role of meningeal or brain biopsy — Brain and meningeal biopsy may be useful in patients who have a progressive deteriorating course despite empiric therapy, particularly if focal findings are detected on brain imaging. Several studies have examined the yield of meningeal or cortical biopsy in patients with chronic meningitis who are without an etiologic diagnosis after a careful history, clinical examination, and cultures and serologic testing of the CSF. In one case series involving 37 patients with chronic meningitis of unknown cause who eventually underwent a biopsy, a definitive diagnosis was made in 16 of 41 biopsies (39 percent) [11]. However, when a biopsy was performed in an area of meningeal enhancement on either MRI or CT, a diagnosis was obtained in 12 of 15 cases (80 percent). In contrast, surgical biopsy was diagnostic in only 2 of 22 cases (9 percent) in which the MRI showed no enhancing regions.

The findings in the preceding study are in contrast with earlier studies that failed to show a similarly high yield from meningeal biopsy. For example, meningeal biopsy was diagnostic in only 5 of 25 patients undergoing meningeal or cortical biopsy at a hospital in Auckland, New Zealand, during the period from 1967 to 1990 [31]. Similar disappointing results were reported from an earlier case series from the same group involving 83 patients [9].

MANAGEMENT OF PATIENTS WITH CHRONIC MENINGITIS OF UNKNOWN CAUSE

Role of empiric therapy

Antituberculous therapy — If a diagnosis is not established by the above diagnostic modalities and if symptoms are severe or fail to improve after a period of observation, empiric therapy with antituberculous therapy may be useful (even if the tuberculin skin test is negative). Empiric antituberculous therapy may also be warranted for patients with less severe symptoms if epidemiologic factors or clinical findings suggest a high risk for tuberculosis (TB; eg, in patients with a past history of direct contact with others with TB or a prior positive tuberculin skin test).

In one study including 28 patients with chronic meningitis for whom no cause could be established after a thorough workup, almost half of the undiagnosed cases responded to empiric antituberculous medications and 11 additional patients had a favorable or complete resolution of symptoms while on antituberculous therapy [9]. Because this study was done in a location where TB was relatively common, this practice may not be applicable in areas with relatively low TB prevalence.

Concurrent steroid therapy should ideally be avoided during an empiric trial of antituberculous medication because it may obscure the evaluation of the clinical response. However, in patients with severe symptoms, concurrent steroid therapy is reasonable and appropriate. If no improvement occurs after four to six weeks of empiric antituberculous therapy and if all mycobacterial cultures remain negative, antituberculosis therapy can reasonably be discontinued.

Glucocorticoids — Empiric glucocorticoid therapy may be useful in selected patients who fail to improve during follow-up, despite the absence of carefully controlled studies demonstrating benefit in patients with chronic meningitis. Some patients with chronic meningitis in whom an infectious etiology was not detected have responded dramatically to empiric glucocorticoids as illustrated by the following findings:

●In one study of 39 patients from the Mayo Clinic with chronic meningitis of unknown cause, symptoms resolved after the institution of glucocorticoid therapy in 19 of the 39 patients (49 percent) despite the fact that a specific cause was not found in any of these patients [14]. The mean duration of symptoms prior to initiating glucocorticoids in this series was 17 months; in some cases, symptoms had been present for up to nine years.

Fourteen of 39 patients (36 percent) remained symptomatic during the follow-up period; symptoms were worse at the end of follow-up in four patients (10 percent). There were no significant differences in the clinical features at the time of initial evaluation between patients who had a poor outcome (ie, death or worsened symptoms during follow-up) compared with those whose symptoms resolved, subsided, or remained stable during follow-up.

A cause for chronic meningitis was eventually established in 10 of 49 patients during follow-up. In 4 of these 10 patients, a diagnosis was obtained by repeated cerebrospinal fluid (CSF) examination (in which CSF cytology revealed a neoplasm in four patients). A total of 21 patients underwent a meningeal or brain biopsy; a diagnosis was obtained in 5 of these 21 patients (neoplasm was present in four of the five positive biopsies).

●Seven of 17 patients with idiopathic chronic meningitis treated with glucocorticoids maintained their response to treatment after glucocorticoids were withdrawn; treatment durations ranged from six weeks to six years [32]. A later diagnosis of granulomatosis with polyangiitis or multiple sclerosis was made in 2 of these 17 patients. Four patients had a transient response to glucocorticoids but later died, and an additional four patients had recurrence of symptoms upon reducing the dose of glucocorticoids.

Glucocorticoids may have to be slowly tapered and transiently reinstituted since some patients relapse rapidly when glucocorticoid therapy is discontinued. Glucocorticoid therapy must be individualized, titrated, or reinstituted if relapse occurs.

Antifungal therapy — Empiric antifungal therapy in patients with chronic meningitis should be used with caution, particularly if there are no compelling clinical or epidemiologic clues pointing to a specific fungal pathogen. Response to such empiric therapy may be hard to assess because patients with fungal meningitis often respond slowly to appropriate therapy and because such therapy can have significant side effects and/or can obscure the true diagnosis.

Antibacterial or antiviral therapy — Empiric antibacterial therapy is rarely indicated unless there are specific clues such as a neutrophilic pleocytosis or there is other evidence of prior clinically suspected but partially treated central nervous system (CNS) or parameningeal bacterial infection. Antiviral therapy is usually not useful unless the diagnosis of Mollaret's meningitis is thought to be possible or likely.

Prognosis — Despite numerous difficulties in establishing reliable data on prognosis, most patients with chronic meningitis in whom a diagnosis cannot be established after a thorough workup have a relatively good outcome. Symptoms either resolve or remain stable over follow-up periods, ranging from one to three years. Repeated diagnostic evaluations in this subgroup of patients will reveal a diagnosis in a minority of cases; the most likely diagnosis to be uncovered during follow-up is neoplasm.

SUMMARY AND RECOMMENDATIONS

●Chronic meningitis is arbitrarily defined as meningitis lasting for four weeks or more and is a complex entity with both infectious and noninfectious causes. Patients with chronic meningitis usually have a subacute onset of symptoms including fever, headache, and vomiting. The symptoms can remain static, fluctuate, and/or slowly worsen. (See 'Introduction' above.)

●An array of infectious agents can present as chronic meningitis, but a nearly identical syndrome can result from a number of inflammatory, malignant, or other noninfectious diseases (table 1). Despite extensive testing, an etiologic diagnosis may not be determined in up to one-third of patients. (See 'Etiologies' above.)

●The clinical symptoms of patients with chronic meningitis rarely point to a specific etiologic diagnosis. However, occasionally, a historical or epidemiologic clue can lead to the discovery of an otherwise obscure diagnosis (table 2). (See 'Clinical features' above.)

●Patients with chronic meningitis typically undergo an array of complex diagnostic investigations including serologic assays, multiple imaging tests, and repeated lumbar punctures (table 3). The type of laboratory testing should be based upon the clinical features of an individual case and the subsequent probability that a specific disease is present. For example, routine evaluation of patients with chronic meningitis usually includes tuberculin skin tests, a chest radiograph, and serologic testing for syphilis as well as testing for the presence of HIV infection. However, tests for other infectious diseases, such as Lyme disease, cysticercosis, trypanosomiasis, and/or schistosomiasis, are only necessary in patients in whom there is reasonable pretest probability that the disease or condition is present. (See 'Diagnosis' above.)

●Analysis of cerebrospinal fluid (CSF) reveals abnormalities in patients with chronic meningitis, but these abnormalities are rarely diagnostic with some notable exceptions. The presence of eosinophilia can provide an important clue to the presence of a parasitic etiology or coccidioidomycosis. Similarly, stained smears of a centrifuged sample of the CSF may occasionally reveal infectious agents, such as fungi or bacteria, and thus lead to a specific etiologic diagnosis. Antigen testing of the CSF for the presence of Cryptococcus neoformans and a Venereal Disease Research Laboratory test for syphilis should be performed on all patients with chronic meningitis. A sample of CSF should be submitted for cultures using media appropriate for growth of aerobic bacteria, mycobacteria, and fungi. Other studies that should be obtained are discussed above. (See 'CSF examination and other laboratory testing' above.)

●Magnetic resonance imaging and computed tomography are useful in the evaluation of patients with chronic meningitis, but these imaging techniques lead to a specific diagnosis in a minority of patients. However, such imaging is important since it can exclude important abnormalities, such as a parameningeal focus of infection, abscess, or a tumor. Rarely, such imaging may disclose one of these processes or the cystic changes typical of cysticercosis. (See 'Imaging studies' above.)

●Brain and meningeal biopsy may be useful in patients who have a progressive deteriorating course despite empiric therapy, particularly if focal findings are detected on brain imaging. (See 'Role of meningeal or brain biopsy' above.)

●If a diagnosis is not established despite a thorough search and if symptoms are severe or fail to improve after a period of observation, empiric therapy with antituberculous therapy may be useful. Empiric antituberculous therapy may also be warranted for patients with less severe symptoms if epidemiologic factors or clinical findings suggest a high risk for TB (eg, in patients with a past history of direct contact with others with TB or a prior positive tuberculin skin test). (See 'Antituberculous therapy' above.)

●Empiric glucocorticoid therapy may be useful in selected patients who fail to improve during follow-up, despite the absence of carefully controlled studies demonstrating benefit in patients with chronic meningitis. Some patients with chronic meningitis in whom an infectious etiology was not detected have responded dramatically to empiric glucocorticoids. (See 'Glucocorticoids' above.)