Journal of Neurology

, Volume 258, Issue 1, pp 3–13

Tuberculous meningitis in patients infected with human immunodeficiency virus


    • Department of NeurologyChhatrapati Shahuji Maharaj Medical University
  • Manish Kumar Sinha
    • Department of NeurologyChhatrapati Shahuji Maharaj Medical University

DOI: 10.1007/s00415-010-5744-8

Cite this article as:
Garg, R.K. & Sinha, M.K. J Neurol (2011) 258: 3. doi:10.1007/s00415-010-5744-8


Tuberculosis is the most common opportunistic infection in human immunodeficiency virus (HIV) infected persons. HIV-infected patients have a high incidence of tuberculous meningitis as well. The exact incidence and prevalence of tuberculous meningitis in HIV-infected patients are not known. HIV infection does not significantly alter the clinical manifestations, laboratory, radiographic findings, or the response to therapy. Still, some differences have been noted. For example, the histopathological examination of exudates in HIV-infected patients shows fewer lymphocytes, epithelioid cells, and Langhan’s type of giant cells. Larger numbers of acid-fast bacilli may be seen in the cerebral parenchyma and meninges. The chest radiograph is abnormal in up to 46% of patients with tuberculous meningitis. Tuberculous meningitis is likely to present with cerebral infarcts and mass lesions. Cryptococcal meningitis is important in differential diagnosis. The recommended duration of treatment in HIV-infected patients is 9–12 months. The benefit of adjunctive corticosteroids is uncertain. Antiretroviral therapy and antituberculosis treatment should be initiated at the same time, regardless of CD4 cell counts. Tuberculous meningitis may be a manifestation of paradoxical tuberculosis-associated immune reconstitution inflammatory syndrome. Some studies have demonstrated a significant impact of HIV co-infection on mortality from tuberculous meningitis. HIV-infected patients with multidrug-resistant tuberculous meningitis have significantly higher mortality. The best way to prevent HIV-associated tuberculous meningitis is to diagnose and isolate infectious cases of tuberculosis promptly and administer appropriate treatment.


BCG vaccinationExtrapulmonary tuberculosisHuman immunodeficiency virusDexamethasoneMycobacterium tuberculosis


In endemic regions of tuberculosis, tuberculous meningitis is a frequently encountered neurological disorder. Despite adequate chemotherapy, tuberculous meningitis is fatal in up to 50% of the cases. A high frequency of disabling morbidity is observed among survivors [1]. Even in advanced countries, such as the United States, tuberculous meningitis is associated with a high mortality. It was observed that even after a long follow-up of several years, only 40% (total 135 patients) of confirmed cases of tuberculous meningitis were still alive as compared to 85% (total 75 patients) of patients with unconfirmed tuberculous meningitis [2].

Human immunodeficiency virus (HIV)-infected patients have a high incidence of all forms of tuberculosis, including tuberculous meningitis. HIV infection influences the pathological, clinical, and laboratory findings in patients with tuberculous meningitis in various ways and may be associated with poorer outcome. HIV tuberculosis co-infection contributes to HIV-related pathogenesis and often increases the viral load in HIV-infected people. In this review, we will be discussing the impact of HIV infection on epidemiology, pathogenesis, clinical features, neuroimaging, and management of tuberculous meningitis. An extensive review of the literature was performed using the PubMed and Google Scholar databases. The search terms used included HIV and tuberculosis; HIV and tuberculous meningitis; AIDS and tuberculosis; AIDS and tuberculous meningitis; tuberculous meningitis; meningeal tuberculosis; and central nervous system tuberculosis.


Tuberculosis is a leading cause of death among people infected with HIV. According to the latest World Health Organization estimate, in 2008 there were 33.4 million HIV-infected cases [3]. In the same year, there were approximately 1.4 million new cases of tuberculosis among persons with HIV infection, and tuberculosis accounted for 23% of AIDS-related deaths. Worldwide, 14 million people are currently co-infected with tuberculosis and HIV. In some countries with high HIV prevalence, up to 80% of the people with tuberculosis test positive for HIV [4, 5]. In advanced stages of HIV infection, tuberculosis may have atypical presentations, including extrapulmonary tuberculosis. Up to 25% of tuberculosis cases in HIV-infected persons may present with extrapulmonary tuberculosis. Extrapulmonary tuberculosis is more common with lower CD4 cell counts [6].

Data from developed countries indicate an increasing trend for extrapulmonary tuberculosis (including tuberculous meningitis) in the population because of prevalent HIV infection [7, 8]. According to the latest tuberculosis report from the United States of America, among 253,299 cases (from 1993 to 2006) 73.6% had pulmonary tuberculosis and 18.7% had extrapulmonary tuberculosis. Approximately 5% of the cases had tuberculous meningitis. The risk factors for extrapulmonary tuberculosis were female sex and foreign birth of the patient, positive HIV status, homelessness and excessive consumption of alcohol [7]. The national tuberculosis surveillance data from 1999 to 2006 for England and Wales (a total of 55,607 cases) also suggested an increasing trend in the proportion of extrapulmonary tuberculosis. Among all the cases of tuberculosis, the proportion with extrapulmonary disease increased from 48% in 1999 to 53% in 2006. The largest increase was seen in miliary tuberculosis, where the proportion rose threefold. The proportion of tuberculous meningitis cases also significantly increased from 1.5% (86) to 2% (165). Miliary tuberculosis and tuberculous meningitis were associated with age over 60 years, foreign birth of Indian, Pakistani, or Bangladeshi ethnic origin and co-infection with HIV [8].

The epidemiological data of tuberculous meningitis from resource poor countries with a very high incidence of pulmonary tuberculosis are not readily available. In one of the studies from a large teaching hospital in India, 375 patients (all patients admitted between 2001 and 2003) with HIV infection were evaluated for opportunistic disease. Tuberculosis was the most common opportunistic disease, seen in 163 patients, and 25 (7%) patients had tuberculous meningitis [9]. In Thailand, among 114 consecutive patients with chronic meningitis, the most common causative agents were Cryptococcus neoformans (54%) and Mycobacterium tuberculosis (37%). Out of the 43 patients with tuberculous meningitis, 3 patients were HIV-positive [10].


Tuberculous meningitis is caused by Mycobacterium tuberculosis, which is an acid-fast bacterium. An extensive heterogeneity in the genetic composition of M. tuberculosis has been demonstrated.

There are several strains of M. tuberculosis, which have a distinct geographical distribution, interactions with the host, and may even differ in their transmission potential. The Beijing genotype of M. tuberculosis is considered a more virulent strain which frequently affects HIV-infected patients [11]. The Beijing M. tuberculosis genotype is globally present with the highest prevalence found in Asia and the territory of the former Soviet Union. In Thailand 58% of patients with tuberculous meningitis were found to be infected with this genotype. The proportion of tuberculous meningitis caused by the Beijing strain was found significantly higher than previously reported figure for pulmonary tuberculosis caused by the Beijing strain [12]. In a large cohort of Vietnamese adults with tuberculous meningitis, authors have recently demonstrated a strong association between the Beijing genotype, drug resistance, and HIV infection [13].

Tuberculosis and HIV infection pathogenesis

HIV co-infection results in an increased risk of all forms of tuberculosis. In patients harboring M. tuberculosis who do not have HIV infection, the lifetime risk of developing tuberculosis is between 10 and 20%. However, in persons co-infected with M. tuberculosis and HIV, the annual risk of developing active tuberculosis may exceed 10% [14]. The pathogenesis of tuberculosis in HIV-1 infected persons includes both reactivation of prior infection and aggravation of existing primary infection. In advanced stages of HIV infection, a disseminated, miliary, or extrapulmonary form of tuberculosis is much more frequent [15].

Active tuberculosis has a major impact on the course of HIV infection. Tuberculosis can accelerate the course of HIV disease by enhancing viral replication. Increased HIV replication, in turn, leads to enhanced CD4 T cell destruction and higher mortality in co-infected patients. The levels of plasma viremia are reduced after successful treatment of the active tuberculosis [16].

Several mechanisms have been proposed to explain the tuberculosis-HIV association. For example, HIV has been shown to impair tumor necrosis factor-alpha (TNF-α) mediated macrophage apoptosis. Apoptosis of macrophages, in response to M. tuberculosis infection, is a critical host defense response, and decreased apoptosis may be responsible for increased susceptibility to M. tuberculosis in HIV-infected persons [17, 18]. In an in vitro study, in response to HIV and tuberculosis co-infection, a variety of genes of human macrophages were found to be up-regulated. However, genetic changes in response to HIV infection alone were fewer in number and significantly lower in magnitude. Normally, these genes encode for pro-inflammatory chemokines and cytokines, their receptors, signaling associated genes, type-I interferon signaling genes and genes of the tryptophan degradation pathway [19]. According to another proposed mechanism, the HIV infection may produce a rapid loss of M. tuberculosis-specific T helper-1 cells in the peripheral blood [20].

Pathogenesis of tuberculous meningitis

Mycobacterium tuberculosis infection is acquired by the inhalation of bacilli as aerosols, which reach and multiply in alveolar macrophages. Through the hematogenous route, the bacilli reach into the central nervous system. M. tuberculosis breaches the blood–brain barrier (composed of tightly apposed brain microvascular endothelial cells). The mechanisms involved in the process of breach of the blood–brain barrier are poorly understood. Inside the central nervous system the bacilli produce small granulomas in the meninges and adjacent brain parenchyma. These tuberculomas may remain dormant for several months or years. Tuberculous meningitis develops when a caseating Rich focus ruptures and discharges its contents into the subarachnoid space [21]. What triggers the rupture of Rich foci is not exactly known. Decreased immunity of the host may be a factor. Following rupture of a Rich focus into the cerebrospinal spaces, the content containing mycobacteria induces an intense immune response and, subsequently, exudate formation.

Mycobacterium tuberculosis is capable of entering and replicating within macrophages. The microglial cells (the resident macrophages of the brain) are the principal targets of M. tuberculosis. Tumor necrosis factor-alpha released from microglial cells has been shown to play a critical role in the containment of infection, granuloma formation, alteration of blood–brain barrier permeability, and cerebrospinal fluid leukocytosis [1]. Several other cytokines present in the microglia, such as β2-integrin (CD-18), interleukin-6, interleukin-1β, chemokine (C–C motif) ligand 2 and chemokine (C–C motif) ligand-5, and chemokine (C–X–C motif) ligand-10, are also involved in the host’s defense mechanisms [22]. However, in patients with tuberculous meningitis, variable cerebrospinal fluid inflammatory responses were observed. For example, a South African study could not demonstrate any significant difference in the cerebrospinal fluid cytokine concentrations and CD4 counts between HIV seropositive and HIV seronegative patients of tuberculous meningitis [23]. HIV-infected patients of tuberculous meningitis had lower cerebrospinal fluid interleukin-10 and interferon-gamma concentrations [24].


The hallmark pathological feature of tuberculous meningitis is the presence of thick gelatinous exudates which are prominent in the basilar regions of the brain. The exudates may block the cerebrospinal fluid pathways resulting in the development of hydrocephalus. The entrapment of intracranial vessels within exudates manifests as cerebral infarcts. The entrapment of cranial nerves manifests as cranial nerve palsies. The basal inflammatory process may also affect the brain parenchyma, resulting in encephalopathy. Frequently, there is formation of cerebral tuberculoma.

HIV infection may influence the pathological features of tuberculous meningitis in several ways. Tuberculous exudates in the HIV-infected patients are minimal, thinner, and of a serous type. The exudates in the HIV-positive patients contain fewer lymphocytes, epithelioid cells, and Langhan’s type of giant cells as compared to HIV-negative patients. Larger numbers of acid-fast bacilli may be seen in the cerebral parenchyma and meninges of HIV-infected patients. Hydrocephalus is not common. Mild ventricular dilatation may be observed secondary to cerebral atrophy [25]. Patients with HIV-associated tuberculous meningitis may present with lower leukocyte counts in peripheral blood and cerebrospinal fluid and may be more likely than HIV-uninfected patients to have concomitant active extrapulmonary extrameningeal tuberculosis [26].

Clinical features—impact of HIV infection

In immunocompetent patients, headache, vomiting, meningeal signs, focal deficits, vision loss, cranial nerve palsies, and raised intracranial pressure are the characteristic clinical features of tuberculous meningitis. The sixth cranial nerve is the most frequently affected cranial nerve. Vision loss, secondary to optic nerve involvement, is a disabling complication. The possible reasons for optic nerve involvement include optochiasmatic arachnoiditis, a large hydrocephalus, optic nerve granulomas, or ethambutol toxicity. Changes in cerebral vessels are characterized by inflammation, spasms, constriction, and eventually thrombosis of cerebral vessels. Infarcts are located at internal capsule, basal ganglion, and thalamic regions and frequently manifest as focal neurological deficits. Tuberculous radiculomyelopathy is characterized by the subacute paraparesis [1].

Usually, human immunodeficiency virus infection does not significantly alter the clinical manifestations, laboratory, or neuroimaging findings in patients with tuberculous meningitis. However, some authors have suggested that some clinical differences exist between immunodeficiency virus-infected and immunodeficiency virus-negative patients [27]. Overall, HIV-positive patients with tuberculous meningitis with higher CD4 cell counts often present in the ‘classic’ form, whereas patients with low CD4 cell counts are more likely to present atypically [28]. Manifestations of tuberculous meningitis are subtle and less specific in patients with low CD4 cell counts. These patients present late in the course of the disease with a prolonged duration of illness and a severe grade of tuberculous meningitis [28].

The most common clinical manifestations observed in HIV-infected patients with tuberculous meningitis were fever and an abnormal mental status [29]. The classical clinical manifestations of tuberculous meningitis such as fever, headache, vomiting, and weight loss occurred in equal frequency in patients with and without HIV infection [30]. Even in a pediatric study, both HIV-infected patients and HIV-uninfected patients with tuberculous meningitis had almost similar clinical manifestations [31]. A Vietnamese study observed, in a comparison to patients with HIV‐negative tuberculous meningitis, that HIV-infected patients with tuberculous meningitis were younger in age and were more commonly male. HIV-infected patients with tuberculous meningitis weighed less but had a higher incidence of other types of extrapulmonary tuberculosis [26].

Differences in several hematological and blood biochemical parameters have also been noted. Concentrations of aspartate transaminase and alanine aminotransferase were significantly higher in HIV-infected patients. A greater proportion of HIV-infected patients with tuberculous meningitis had hepatitis B surface antigenemia. The authors suggested that these differences relate partly to the epidemiological pattern of HIV infection (young male drug users with a high prevalence of viral hepatitis) and partly to the effects of systemic immunodeficiency (low weight, low hematocrit level, and high prevalence of extrapulmonary/meningeal tuberculosis) [26]. Other studies have also reported frequent liver function abnormalities in these patients [2]. HIV-infected children with tuberculous meningitis may have a lower hemoglobin level (<8 gm/dL) [32]. These patients have more frequent concurrent pulmonary infection, even in the absence of respiratory symptoms. Lymphadenopathy and hepatosplenomegaly were also frequent findings in HIV-infected patients with tuberculous meningitis [33].


Cerebrospinal fluid findings

Cerebrospinal fluid examination is the cornerstone of the diagnosis of tuberculous meningitis. The ‘gold standard’ for the diagnosis is the demonstration of M. tuberculosis bacilli in the cerebrospinal fluid. In human immunodeficiency virus-associated tuberculous meningitis, a relatively higher 69% positivity for smear and 87.9% positivity for bacterial culture have been demonstrated [34].

The values of routinely measured cerebrospinal fluid parameters are almost similar in HIV-positive and negative patients with tuberculous meningitis [26, 31, 35]. Some studies, however, noted a lower cerebrospinal fluid leukocyte count and a lower protein level in HIV-positive patients [25, 33]. Patients with advanced HIV disease usually have low numbers of lymphocytes in the peripheral blood, which may reflect in a low lymphocyte count in the cerebrospinal fluid. Moreover, tuberculous meningitis may stimulate increased HIV replication in the central nervous system, resulting in the destruction of cerebrospinal fluid lymphocytes [34]. M. tuberculosis may be isolated from the cerebrospinal fluid in a higher proportion of HIV-infected patients than in HIV-uninfected patients, perhaps due to greater mycobacterial dissemination within the central nervous system [30]. In a Vietnamese study, microbiological confirmation of tuberculous meningitis was obtained in 45% of HIV-positive patients, in contrast to 33% of HIV-negative patients [26]. The quantity of acid-fast bacilli seen in the cerebrospinal fluid smear appeared to be higher, with a shorter time for detection of acid fast bacilli in HIV-associated tuberculous meningitis than in HIV-negative tuberculous meningitis [34]. M. tuberculosis can be isolated from significantly smaller cerebrospinal fluid volumes from HIV-infected individuals compared to uninfected ones [36]. Cerebrospinal fluid examination findings may be normal in 5% of HIV-positive patients with tuberculous meningitis. The percentages of HIV-positive tuberculous meningitis patients with normal cerebrospinal fluid parameters are as follows: glucose 15%, protein 40% and leukocyte count 10% [37].

Immunological tests such as tuberculin skin testing and interferon gamma release assays should not be relied upon, as HIV-related immunosuppression might be associated with false-negative results. The frequency of false-negative and indeterminate interferon gamma release assay results increases with advancing immunodeficiency [38]. However, recently the quantitative region of difference (RD)-1 interferon-gamma-T-cell ELISPOT assay (an immunological test), using CSF mononuclear cells, has demonstrated an accurate rapid test in HIV-infected patients with tuberculous meningitis [39]. Lipoarabinomannan is a glycolipid forming part of the M. tuberculosis cell wall. The lipoarabinomannan antigen-detection test in serum or cerebrospinal fluid is a rapid and relatively simple assay. A recent South African study evaluating the cerebrospinal fluid lipoarabinomannan antigen has reported sensitivity and specificity of 64 and 69%, respectively, for serum and cerebrospinal fluid [40].

The microscopic observation drug susceptibility (MODS) assay is a low-cost liquid mycobacterial culture technique. In HIV-infected patients, the MODS assay detected M. tuberculosis with greater sensitivity and speed and ruled out tuberculosis more quickly and with fewer indeterminate culture results in comparison to that of the Lowenstein–Jensen culture [41].

Chest radiography

The presence of pulmonary tuberculosis in a chest radiograph often helps in diagnosing tuberculous meningitis. The chest radiograph is abnormal in up to 46% of HIV-positive patients with tuberculous meningitis [26]. Patients with HIV infection and pulmonary tuberculosis may present with an atypical chest radiograph. In patients with low CD4 cell counts, a primary tuberculosis-like pattern, with diffuse interstitial or miliary infiltrates, little or no cavitation, and intrathoracic lymphadenopathy, is more common. Lobar infiltrates with or without hilar adenopathy or diffuse infiltrates resembling the interstitial pattern of Pneumocystis jirovecii pneumonia may also be seen.


The dominant neuroradiologic findings in tuberculous meningitis include basal meningeal enhancement, hydrocephalus, tuberculoma, and infarctions in the brain parenchyma. The influence of HIV infection on intracranial imaging of tuberculous meningitis has been extensively investigated, and the findings suggested that basal meningeal enhancement and hydrocephalus on computed tomography of the brain were less common in HIV-infected patients [25, 31]. HIV-infected individuals were also more likely to present with cerebral infarcts and mass lesions [35]. Infarcts were more commonly located in the cortex in HIV-infected patients and basal ganglia in HIV-uninfected patients [25]. (Figs. 1, 2, and 3).
Fig. 1

Contrast enhanced computed tomography showing basal exudates, meningeal enhancement and ventricular dilatation in a HIV-infected patient with tuberculous meningitis
Fig. 2

Gadolinium enhanced cranial magnetic resonance imaging showing a tuberculoma in the pontine region of the brain
Fig. 3

Cranial magnetic resonance imaging (T2-weighted, FLAIR, and diffusion weighted images) shows an infarct in the left perisylvian region

Differential diagnosis

In HIV-infected patients, a variety of central nervous system opportunistic infections and malignancies need to be considered as a differential diagnosis of tuberculous meningitis. Six features (duration of illness more than 5 days, presence of headache, cerebrospinal fluid white blood cell count of <1,000/mm [3], clear appearance, lymphocyte count >30% and protein content of >100 mg/dL) favor the diagnosis of tuberculous meningitis [1].

Cryptococcal meningitis is the most important differential diagnosis. It generally occurs in patients with very low CD4 T cell counts (<100/μL). In cryptococcal meningitis, headache is often the most dominant and sometimes may be the sole manifestation. In cryptococcal meningitis, meningeal signs may not be demonstrable. Neuroimaging evaluation is often normal. The cerebrospinal fluid examination may be normal in 16% of patients [37]. The diagnosis of cryptococcal meningitis is made by identification of fungus in cerebrospinal fluid by India ink preparation. Other fungi that may rarely cause meningitis in patients with HIV infection are Coccidiodes immitis and Histoplasma capsulatum. Acute aseptic meningitis may develop at the time of seroconversion. The clinical manifestations are similar to other viral meningitis, with fever, headache, stiff neck, photophobia, and cranial nerve palsies.

Patients with toxoplasmosis can also present with diffuse meningoencephalitis. Toxoplasmosis is a common late complication of HIV infection, usually occuring in patients with CD4 T cell counts <200/μL. Progressive multifocal leukoencephalopathy and primary central nervous system lymphoma are other conditions which may present with headache, confusion, and focal deficits mimicking chronic meningitis. All these conditions produce focal lesions of the brain and can be diagnosed on the basis of characteristic neuroimaging findings.


In patients with tuberculous meningitis, antituberculosis treatment should be started as quickly as possible. The basic principles, which are applicable for the treatment of pulmonary tuberculosis, remain the same for the treatment of tuberculous meningitis [42].

The standard antituberculosis treatment regimens are equally efficacious in HIV-negative and HIV-positive patients with tuberculous meningitis. Hence, in HIV-infected patients there is no need to alter the choice or duration of anti-tuberculosis treatment [42]. The recommended duration of treatment for tuberculous meningitis is at least 9–12 months. The usual treatment consists of an initial phase of isoniazid, a rifamycin, pyrazinamide, and ethambutol for the first 2 months. This is followed by a continuation phase of isoniazid and a rifamycin for 7–9 months. World Health Organization guidelines suggest that in patients with tuberculous meningitis, ethambutol should preferably be replaced by streptomycin [42], as ethambutol has the potential to cause vision impairment. Tuberculosis patients with positive HIV status and all tuberculosis patients living in HIV-prevalent settings should receive daily antituberculosis treatment [43]. The incidence of relapse and failure among HIV-positive pulmonary tuberculosis patients who are treated with intermittent antituberculosis treatment may be 2–3 times higher than that in patients who received a daily intensive phase [42].

Role of corticosteroids

The exact benefit of corticosteroids in HIV-infected patients with tuberculous meningitis is uncertain. A study conducted on 545 Vietnamese adults (which also included 98 HIV-infected patients) found a non-significant reduction in death and severe disability in dexamethasone-treated HIV-infected patients with tuberculous meningitis [44]. Still, the British Infectious Diseases Society guidelines suggest that concomitant corticosteroids should be given [45]. Corticosteroids may also be of possible value in the management of tuberculous meningitis secondary to tuberculosis associated immune reconstitution inflammatory syndrome [46].

Co-administration of antituberculosis and antiretroviral therapy

The World Health Organization treatment guidelines recommend early antiretroviral treatment for all HIV-infected individuals with active tuberculosis irrespective of CD4 cell count [41]. The first-line anti-retroviral therapy regimen should contain two nucleoside reverse transcriptase inhibitors plus one non-nucleoside reverse transcriptase inhibitor [42]. Efavirenz is a preferred non-nucleoside reverse transcriptase inhibitor for tuberculosis HIV co-infected patients [47].

Antiretroviral therapy has been reported to reduce tuberculosis rates by up to 90% at an individual level, by 60% at a population level, and to reduce tuberculosis recurrence rates by 50% [42, 48]. Initiation of antiretroviral treatment in patients with HIV/tuberculosis co-infection, if accompanied by high levels of coverage and drug compliance, reduces the number of tuberculosis cases, mortality rates, and tuberculosis transmission [49].

Four important considerations are relevant for antituberculosis treatment in HIV-infected patients: timing of antiretroviral therapy initiation, drug interactions between antiretroviral therapy and rifamycins, an increased frequency of paradoxical reactions, and development of drug-resistant tuberculosis.

Timing of anti-retroviral therapy initiation

It is uncertain whether antiretroviral therapy should be started with antituberculosis therapy or after a delay. Simultaneous initiation of antituberculosis therapy and anti-retroviral therapy may lead to unwanted drug interactions and toxicities. Some authors suggest that anti-retroviral treatment may be delayed for those with higher CD4 counts, but should not be delayed in those with severe immune suppression (CD4 count <100 cells/μL) [50]. The Center for Disease Control and Prevention recommends that for patients with a CD4 count <100 cells/μL, antiretroviral therapy should be started after more than 2 weeks of antituberculosis treatment [51]. Delay in initiating antiretroviral therapy is associated with serious risk of other opportunistic infections. The recent SAPiT trial (starting antiretroviral therapy in tuberculosis) from South Africa found that mortality among people co-infected with HIV and tuberculosis could be halved if antiretroviral therapy was initiated either within 4 weeks of starting antituberculosis treatment or within 4 weeks of completing the intensive phase of antituberculosis therapy [52].

According to the World Health Organization recommendations, antituberculosis treatment should be started first, followed by antiretroviral therapy as soon as possible after starting antituberculosis treatment, preferably within the first 8 weeks of starting tuberculosis treatment [42].

Co-trimoxazole preventive therapy

In all HIV-positive tuberculosis patients, co-trimoxazole preventive therapy should be initiated as soon as possible and given throughout the course of antituberculosis treatment. Co-trimoxazole therapy substantially reduces mortality in HIV-positive tuberculosis patients. The exact mode of activity is not clear but co-trimoxazole is known to have preventive impact on Pneumocystis jirovecii, malaria, toxoplasmosis, and on several other bacterial infections [42].

Drug interactions between antiretroviral therapy and rifamycins

Concomitant use of rifampicin and antiretroviral drugs is likely to be complicated by drug-to-drug interactions. These drug interactions can result in subtherapeutic antiretroviral drug concentrations, loss of antiviral efficacy, and the development of viral resistance [43].

Rifampicin, a potent enzyme inducer of the cytochrome P450 system, may lower serum levels of many HIV protease inhibitors and some nonnucleoside reverse transcriptase inhibitors. The World Health Organization recommends that first-line antiretroviral therapy regimens for tuberculosis patients are those that contain efavirenz, since interactions of efavirenz with antituberculosis drugs are minimal [42].

Rifabutin is a rifamycin with significantly less induction of P450 enzymes. Therefore, rifabutin has less effect on the serum concentrations of antiretroviral agents. In individuals who need antituberculosis treatment and who require an antiretroviral therapy containing a boosted protease inhibitor, a rifabutin-based antituberculosis treatment is recommended [42].

Immune reconstitution inflammatory syndrome and tuberculous meningitis

The immune reconstitution inflammatory syndrome is an important complication of antiretroviral therapy, especially in patients with tuberculosis. There are two forms (paradoxical and unmasking) of tuberculous immune reconstitution inflammatory syndrome. The ‘paradoxical’ type is characterized by clinical worsening of a patient on tuberculosis treatment, and the ‘unmasking’ type is characterized by undiagnosed tuberculosis becoming apparent after starting antiretroviral therapy [53]. Immune reconstitution inflammatory syndrome associated with M. tuberculosis is common in high tuberculosis-prevalent areas, occurring in approximately 11–36% cases. Risk factors for immune reconstitution inflammatory syndrome include a high pathogen load and very low CD4 T-cell count (<50 cells/μL) when anti-retroviral therapy is initiated [54].

Paradoxical neurologic tuberculosis-associated immune reconstitution inflammatory syndrome accounts for approximately 12% of all paradoxical tuberculosis-associated immune reconstitution inflammatory syndrome cases [45]. Dominant manifestations, in addition to tuberculous meningitis, were intracranial tuberculoma and tuberculous radiculo-myelopathy [45, 55]. Neuroimaging revealed that in patients with meningitis, meningeal enhancement and hydrocephalus were infrequent [55].

Paradoxical tuberculous reactions should not be labeled as a new or resistant infection. Differential diagnoses include failure of antituberculosis treatment because of drug resistance or suboptimal antituberculosis drug concentrations, drug reactions, and alternative opportunistic conditions such as toxoplasma and cryptococcal meningitis.

Drug-resistant tuberculous meningitis in HIV-positive patients

Multidrug-resistant tuberculosis is caused by bacteria that are resistant to at least isoniazid and rifampicin. It has been estimated that 440,000 people had multidrug-resistant tuberculosis worldwide in 2008 and that one-third of them died [56]. Drug-resistant tuberculosis is a major public health concern in European countries as well. The estimated number of multidrug-resistant tuberculosis cases in Europe in 2008 is approximately 81,000. Eastern European countries have the highest rates of multidrug-resistant tuberculosis in the world. HIV-positive tuberculosis patients are at higher risk of harboring multidrug-resistant tuberculosis strains. Tuberculosis patients living with HIV in Eastern European countries are at a high risk of harboring multidrug-resistant tuberculosis strains [57].

Drug-resistant tuberculous meningitis has frequently been reported in patients with HIV infection. For example, out of 90 HIV-infected Brazilian patients with tuberculous meningitis, 7% had primary resistance to isoniazid and 9% to multidrug-resistant strains [27]. Authors from South Africa reported drug resistance to at least isoniazid and rifampicin in 8.6% of their patients with tuberculous meningitis. Sixty percent of them were HIV-positive. In this study, during the period of 1999–2002, 350 patients with tuberculous meningitis were identified by cerebrospinal fluid culture for M. tuberculosis [58]. In the Vietnamese study, drug resistance to one or more first-line drugs was found in 54.3% and multi-drug resistance in 8.7% of HIV-positive tuberculous meningitis patients. All patients with multidrug-resistant tuberculous meningitis died, but streptomycin and/or isoniazid resistance were not associated with mortality [34]. In Argentina, multidrug-resistance was observed in 41.6% isolates from HIV-infected patients with tuberculous meningitis. In this investigation, 42 out of 101 isolates were multidrug-resistant strains. Ten isolates had isolated resistance to single antituberculosis drugs. Because of multidrug-resistant strains, tuberculous meningitis was more frequent in patients who received irregular antituberculosis treatment [59, 60]. Patients with multidrug-resistant tuberculous meningitis and HIV-infection have lower cure rates and higher mortality rates than patients with drug-susceptible tuberculous meningitis, and most patients die within 3 months [26, 61].

In addition to drug resistance, several other mechanisms may also be responsible for treatment failure in HIV-infected patients with tuberculosis. It has been observed that HIV-infected individuals had significantly low peak serum rifampicin and isoniazid concentration compared to HIV-uninfected individuals [62]. The percent of rifampicin dose excreted in the urine positively correlated with CD4 count, indicating greater malabsorption in patients with more advanced HIV disease [63]. The patients with advanced HIV disease (CD4 T cell counts <100/μL) are more prone to treatment failure and relapse with rifampicin-resistant organisms when treated with “highly intermittent” (for example, once- or twice-weekly) rifampin or rifabutin-containing regimens [64].

Current treatment guidelines recommend that an antituberculosis treatment regimen for multidrug-resistant tuberculosis should include at least five drugs during the intensive phase. Treatment regimen should include drugs that a patient has not received before and to which the bacilli are susceptible. The regimen should also include an injectable medication. Appropriate second-line drugs, those that produce significant concentrations in the cerebrospinal fluid (ethionamide, cycloserine, and fluoroquinolones), should be included. In five drug regimens, one of the antituberculosis drugs should be fluoroquinolones. The initial phase of 6 months should be followed by a continuation phase of 12–18 months [58].

Adverse drug reactions

Adverse drug reactions are more common among HIV‐infected patients than among HIV‐uninfected patients being treated for tuberculosis. Risk of drug reaction increases with declining CD4 cell counts. The antituberculosis drugs and first-line antiretroviral drugs have many common side effects, such as skin rashes, gastrointestinal intolerance, hepatoxicity, central nervous system symptoms, peripheral neuropathy, and blood dyscrasias [65]. Most reactions occur in the first two months of treatment. Skin rash is the most common reaction, and fever often precedes and accompanies a rash. Mucous membrane involvement is common. Severe skin reactions, which may be fatal, include exfoliative dermatitis, Stevens–Johnson syndrome, and toxic epidermal necrolysis. Rifampicin-associated anaphylactic shock and thrombocytopenia have also been reported [46].


There are conflicting reports available about the effect of HIV infection on the outcome of tuberculous meningitis. Some authors observed no significant impact of HIV infection on the mortality due to tuberculous meningitis [35, 66], whereas others have reported higher mortality rates in HIV-infected tuberculous meningitis patients [25, 26]. Two Vietnamese studies reported mortality rates of 65 and 67% in HIV-infected patients with tuberculous meningitis, in contrast to approximately 28% deaths in HIV-uninfected patients [26, 34]. Advanced stage of tuberculous meningitis, low serum sodium, and decreased cerebrospinal fluid lymphocyte percentage were associated with increased risk of death [34]. A CD4 T-cell count less than 50 cells/μL, infection caused by multidrug-resistant strains, altered sensorium and hemiplegia were also found to be associated with poor prognosis [25, 61]. Asignificantly higher number of treatment failures in the HIV-infected group suggests that HIV infection may influence the response to treatment [67].


Bacillus Calmette-Guérin (BCG) vaccination is effective in preventing childhood tuberculous meningitis and miliary tuberculosis. Unfortunately, BCG vaccination in HIV-uninfected children is associated with disseminated BCG infection and deaths. The risk of disseminated BCG disease increased several hundred times in HIV-infected infants compared to HIV-uninfected infants. The World Health Organization does not recommend BCG vaccination for children with symptomatic HIV infection [68].

According to a recent systematic review, treatment of latent tuberculosis infection reduces the risk of active tuberculosis in HIV-positive individuals [69]. HIV-infected patients who have been exposed to an infectious tuberculosis patient should also receive isoniazid preventive therapy regardless of the Mantoux test result. A 9-month course of isoniazid at a daily dose of 5 mg/kg (up to 300 mg/day) reduces the risk of active tuberculosis in infected people by up to 90%. The protective effect is believed to be life-long in the absence of re-infection. The World Health Organization recommends a 3Is policy (intensified tuberculosis case finding, infection control, and isoniazid preventive therapy) for prevention of HIV-associated tuberculosis [5]. Several tuberculosis vaccines are entering into field trials and have shown promise for the future [70].


Tuberculous meningitis is a serious life-threatening disease, especially in HIV-infected persons. Infection by multidrug-resistant strains poses a major challenge for the clinician, as it is an important predictor of mortality. To fight this deadly combination, clinicians should be aware of the pathogenesis of infection and disease, rapid diagnosis and identification of resistant strains, optimal regimens of antituberculosis treatment and adjunctive corticosteroids, and the optimal time to initiate antiretroviral therapy. Currently, the 3Is policy remains the best way to fight this menace.

Copyright information

© Springer-Verlag 2010