Tuberculous Meningitis

Home , Aseptic meningitis, Rich focus, Tuberculous meningitis

J Neurol Neurosurg Psychiatry 2000;68:289–299 289 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

NEUROLOGICAL ASPECTS OF TROPICAL DISEASE

Tuberculous meningitis

G Thwaites, T T H Chau, N T H Mai, F Drobniewski, K McAdam, J Farrar

Uncertainty and doubt dominate all aspects of infection is under debate.11 Certain ethnic tuberculous meningitis (TBM). The variable groups seem to be more susceptible to M natural history and accompanying clinical fea- tuberculosis than others. Studies using tubercu- tures of TBM hinders the diagnosis. Ziehl- lin conversion as a surrogate marker suggest Neelsen staining lacks sensitivity and culture that black skinned people are more susceptible results are often insuYciently timely to aid to infection than white people.12 Recently it has clinical judgement. New rapid diagnostic been proposed that certain polymorphisms in methods are incompletely evaluated, and many the human NRAMP1 gene may aVect suscep- are not suitable for laboratories in low income tibility to pulmonary tuberculosis in West countries. The duration of chemotherapy for Africans.13 Whether genetic factors influence TBM is unclear and the benefits of adjuvant prevalence of TBM within a population is corticosteroids remain in doubt. The only unknown. uncomfortable certainties lie in the fatal conse- The extent to which BCG vaccination Department of Microbiology, St quences of missed diagnoses and delayed treat- aVords protection against TBM is still debated. Thomas’s Hospital, ment. A meta-analysis of the published trials on the London UK This review will discuss the current uncer- eYcacy of BCG vaccination suggested a G Thwaites tainties surrounding TBM. More attention will protective eVect of 64% against TBM.14 This be given to diagnosis and management, as figure is higher than that suggested for pulmo- Centre for Tropical these areas have a direct bearing on patient nary TB (50%), but may only reflect more Diseases, 190 Ben Ham Tu, Quan 5 Ho Chi outcome. accurate case ascertainment of TBM given the Minh City Viet Nam universal requirement for admission to hospi- T T H Chau tal. Overall, these and other studies support the NTHMai Epidemiology About 2000 million people in the world today view that BCG vaccination is protective against are infected with tuberculosis,1 but only 10% TBM. PHLS Mycobacterium Close correlation exists between the ob- Reference Unit, King’s develop clinical disease. Why some people College School of develop clinical disease remains unclear. The served incidence of TBM in children aged 0–4 Medicine, London UK reasons are likely to be multifactorial: inherent years, and the population’s annual average risk http://jnnp.bmj.com/ F Drobniewski not just to the individual person, but to their of infection with M tuberculosis. The incidence of TBM has been calculated to represent 1% of Medical Research given population and environment. 15 Before HIV the most important determinant the annual risk of infection. Risk of infection Council Laboratories depends on the prevalence of infectious cases Banjul The Gambia for the development of TBM was age. In K McAdam populations with high TB prevalence TBM in a given community. Prevalence of infectious diVers from pulmonary, and other extra- cases is dependant not only on the risks perti- Wellcome Trust pulmonary tuberculosis, in that the peak age is nent to each person for developing disease, but

Clinical Research 2 to the factors inherent in the community on October 2, 2021 by guest. Protected copyright. Unit, Centre for from 0–4 years. In populations with lower TB prevalence, most cases of TBM are in adults. encouraging spread of infection. The main rea- Tropical Diseases, Ho son for the spread of tuberculosis is poverty, Chi Minh City, Viet Risk factors identified for these people are Nam alcoholism, diabetes mellitus, malignancy, and with resulting homelessness, malnutrition, and G Thwaites recent corticosteroid use.3–5 Coinfection with breakdown of public health infrastructure. J Farrar HIV now dwarfs these risk factors. HIV The total number of tuberculosis cases in the world is increasing.16 17 It is estimated that most Centre for Tropical increases the lifetime risk of developing clinical 16 TB postinfection to 1 in 3.6 HIV also of these new cases will be in south east Asia Medicine, NuYeld fuelled by the rapid spread of HIV. It has been Department of predisposes to the development of extra- Medicine, University 7 predicted that without intervention 200 million pulmonary TB, and in particular TBM, a risk 1 of Oxford, UK which increases as the CD4 count declines.8 people alive today will develop TB. The physi- J Farrar The disease constitutes either reactivation of cian needs to be aware of these changes, as less latent infection, or new infection. Evidence common forms of tuberculosis such as TBM Correspondence to: will be encountered more often. Dr Guy Thwaites, from DNA fingerprinting of strains using Department of Microbiology, restriction fragment length polymorphism St Thomas’s Hospital, London SE1 7EH, UK suggests that in the United States up to 40% of Causative agent email Guy.Thwaites@ new disease in both HIV positive and non- HIV Tuberculous meningitis was first described as a gstt.sthames.nhs.uk patients is due to recent infection.910 distinct pathological entity in 1836,18 and Rob- Received 4 October 1999 The extent to which a person’s genetic con- ert Koch demonstrated that tuberculosis was Accepted 11 October 1999 stitution eVects resistance or susceptibility to caused by Mycobacterium tuberculosis in 1882.19 290 Thwaites, Chau, Mai, et al J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

M tuberculosis is an aerobic gram positive rod tuberculous bacteraemia. Dissemination to the that stains poorly due to its thick cell wall con- CNS is more likely, particularly if miliary TB taining lipids, peptidoglycans, and arabi- develops. nomannans. The Ziehl-Neelsen stain uses the The second step in the development of TBM properties of the cell wall to form a complex is rupture of a Rich focus into the subarachnoid that prevents decolourisation by acid or space. This heralds the onset of meningitis, alcohol.20 which if left untreated, will result in severe and The characteristics of M tuberculosis enabling irreversible neurological pathology. In 75% of it to cause disease are complex and incom- children the onset of TBM is less than 12 pletely understood. It is accepted that those months after the primary infection.29 with active pulmonary infection vary consider- Three general processes produce the subse- ably in their ability to transmit the disease to quent neurological pathology: adhesion forma- others.21 Part of the variability is explained by tion, an obliterative vasculitis, and an encepha- diVerences in environment, infectious burden, litis or myelitis.30 Adhesions result from a dense and host immunity. Experimental evidence basal meningeal exudate that develops after suggests that the virulence of individual strains inoculation of bacilli into the subarachnoid is also significant, and selected gene mutations space. The exudate contains lymphocytes, have been shown to aVect virulence.22 23 plasma cells, and macrophages, with increasing Whether there are stains which cause more quantities of fibrin. Blockage, through adhe- disease of the CNS is not known. The sion formation, of the basal subarachnoid predominance of one strain typed using cisterns can result in obstruction of the CSF restriction fragment length polymorphism has and hydrocephalus. Adhesions around the been reported from a series of patients with M interpendicular fossa and related structures tuberculosis,24 however the mechanisms by can compromise cranial nerves, particularly II, which neurovirulence may occur is unknown. IV, and VI, and the internal carotid artery. An As the contribution of strain variation and obliterative vasculitis of both large and small virulence becomes more apparent, so tech- vessels develops that can result in infarction niques are developing to determine the genetic and stroke syndromes. These commonly occur components of mycobacterial virulence. in the territories of the internal carotid, “Molecular” Koch’s postulates have been proximal middle cerebral, and the perforating applied to advance a hypothesis for a single vessels to the basal ganglia.31 Infarction gene basis for a virulent phenotype. The through vasculitis is the mechanism by which phenotype is analogous to the disease, the gene many of the diverse clinical neurological analogous to the organism.25 Only a few abnormalities in TBM occur, and accounts for virulence genes have currently satisfied these an appreciable part of the irreversible neuro- methods.26 The recent determination of the logical sequelae. The intensity of the basal complete genomic sequence for M tuberculosis inflammatory process extends into the paren- should expand our understanding in this area.27 chyma resulting in encephalitis. Oedema occurring as a consequence can be marked throughout both hemispheres. This will con- Pathogenesis tribute to rising intracranial pressure and the A discussion on the pathogenesis of tubercu- global clinical neurological deficit. lous meningitis can be directed on two levels. A rare complication of TBM is tuberculous http://jnnp.bmj.com/ On a macroscopic level there are the mecha- encephalopathy.32 Usually occurring in a young nisms by which the tuberculous bacilli dissemi- child with progressive primary TB, the presen- nate to the CNS. This is discussed alongside tation is of reducing conscious level with few the role of granulomatous inflammation, the focal signs and minimal meningism. DiVuse currency of tuberculous pathology, in causing oedema and white matter pallor with demyeli- gross pathological changes within the CNS. On nation are found pathologically. The pathogen- a microscopic level there are the cellular and esis is uncertain, but is presumed to be immune immune mechanisms that can result in both mediated. Diagnosis is important as anecdotal on October 2, 2021 by guest. Protected copyright. the disease and its control. reports suggest a good response to The development of TBM is a two step corticosteroids.33 process28; M tuberculosis bacilli enter the host by The pathogenesis of TBM at a cellular level droplet inhalation, the initial point of infection is poorly understood. Knowledge regarding the being the alveolar macrophage. Escalating pathogenesis of pulmonary infection is limited, localised infection within the lung with dis- but certain key principles may serve to semination to the regional lymph nodes illuminate some of the processes evident in the produces the primary complex. During this CNS. The formation of caseating granuloma- stage there is a short but significant bacterae- tous inflammation is fundamental. mia that can seed tubercle bacilli to other Current theories of immunopathogenesis organs in the body. In those who develop seek to explain the roles and interactions TBM, bacilli seed to the meninges or brain between the macrophage, the helper T cell, and parenchyma, forming small subpial or sub- the organism. Cell mediated immunity is ependymal foci. These are called Rich foci, central to both the control of infection and the after the original pathological studies of Rich production of tissue damage.34 Lurie’s experi- and McCordick.28 In about 10% of cases, par- ments on tuberculosis in rabbits describe the ticularly in children, the primary complex does fundamental stages of the disease.35 Theories of not heal but progresses. Tuberculous pneumo- immunopathogenesis aim to explain these nia develops with heavier and more prolonged stages. Tuberculous meningitis 291 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

The initial stage of infection is the ingestion prompt institution of chemotherapy.43–49 Delay of the inhaled tubercle bacilli by the alveolar in treatment either results in death, or substan- macrophage. Depending on the ability of the tial neurological morbidity.50 macrophage to resist infection the bacilli In those patients presenting with TBM the multiply and destroy the macrophage. The history will often be unhelpful. Recent contact innate and possibly genetically determined with tuberculosis should be elucidated: several resistance to infection at this stage has been studies have shown that between 70% and 90% discussed earlier. of children have had recent contact with During the second stage bacilli grow TB.44 51 The prodrome is usually non-specific logarithmically within newly recruited with no one symptom predominating: 28% macrophages.34 After about 2 weeks CD4 T report headache, 25% were vomiting, and 13% cells specific for mycobacterial peptides ap- had fever.43 Only 2% reported meningitic pear. Production of ã-interferon activates mac- symptoms. rophages enabling more eYcient intracellular In a review of 205 children only 38% had killing of tubercle bacilli. Activated macro- fever at presentation with 9% reporting phages produce interleukin 1-â, and tumour photophobia.43 14% remained free from necrosis factor (TNF) which promotes granu- meningism throughout the illness. Recent loma formation.36 reviews confirm the wide variety of presenta- The basal inflammatory exudate is central to tions seen with TBM.44–48 An Australian series the pathogenesis of TBM. The primary of 58 patients found that on the day of admis- complex results in the development of cell sion TBM was considered a diagnosis in mediated immunity; therefore, the rupture of 36% of cases, with 6% receiving immediate the Rich focus with release of bacilli into treatment.44 The duration of presenting symp- the subarachnoid space will result in a local toms varied from 1 day to 9 months, although T-cell dependent response. The necrotising 55% presented with less than 2 weeks of symp- granulomatous response is fundamental to the toms. In one quarter of patients diagnosis and subsequent pathology.30 Dannenberg hypoth- treatment were delayed until clinical deteriora- esises that necrosis is the result of a delayed tion confirmed the diagnosis of TBM. More type hypersensitivity reaction to exposed advanced disease, however, may be just as hard tuberculoproteins.34 Others propose that the to diagnose. A review of 48 cases admitted to granuloma architecture dictates the extent of a French intensive care unit disclosed that the central necrosis: a reduced capacity for on admission to the unit only 65% had fever, focusing T lymphocytes within a point of 52% had focal neurology, and 88% had infection may result in failure to deliver meningism.45 adequate cytokine concentrations to the cen- The neurological complications that can tres of large granulomas, resulting in degenera- occur are legion.50 52 Their nature and diversity tion and necrosis.37 can be predicted from an understanding of the Studies in bacterial meningitis have shown site of disease and the pathogenesis of TBM. that CSF concentrations of TNF-á correlate Adhesions can result in cranial nerve palsies with disease severity.38 TNF-á concentrations (particularly II, III, IV, VI, VII, and VIII), con- in TBM are lower than they are in bacterial striction of the internal carotid resulting in meningitis. In TB sensitised animals small stroke, and obstruction of CSF flow leading to concentrations of TNF result in substantial tis- raised intracranial pressure, reduced conscious http://jnnp.bmj.com/ sue necrosis.39 Rabbit models of TBM show level, and hydrocephalus. Infarcts occur in that CSF concentrations of TNF-á correlate about 30% of cases,53 commonly in the internal with clinical progression.40 Intervention with capsule and basal ganglia, causing a range of antibiotics and thalidomide, an anti-TNF disorders from hemiparesis to movement agent, resulted in an improvement in survival disorders. Seizures are common, especially in and neurological outcome.40 The protective children and elderly people. Hydrocephalus, role of TNF should not be forgotten, with the tuberculoma, oedema, and hyponatraemia due promotion of granuloma formation36 and to inappropriate ADH secretion can all cause on October 2, 2021 by guest. Protected copyright. enhanced killing of infected cells in vitro.41 seizures. In those presenting with root pain, in These models provide evidence for the combination with either spastic or flaccid important role of cytokines, in particular TNF, paralysis and early loss of sphincter control, the in the pathogenesis of TBM. They suggest diagnosis of spinal meningitis should not be alternative immunomodulatory therapeutic ap- forgotten. proaches that may supercede the blind use of Over the past 10 years there have been stud- corticosteroids. ies documenting the relation between HIV and TBM.46–49 Although HIV infected patients with Clinical features TB are at increased risk of TBM,46 the clinical In textbooks TBM is described as a subacute features and outcomes of the disease do not lymphocytic meningitis.42 Although this may be seem to be altered.45–49 Those with TBM and true in many cases, it is not helpful here to HIV often have concomitant extrameningeal describe the classic presenting features of disease. In one report 65% had clinical or TBM. This is not to suggest that they do not radiographic evidence of extrameningeal TB occur, but more to emphasise the variety of on admission.46 In another series 77% of those clinical presentations and the requirement for a with HIV had clinical evidence of extramenin- high level of diagnostic suspicion. To date all of geal TB, compared with 9% in those without the series of TBM reported in the literature HIV.49 In more than half there may also be a stress the importance of early diagnosis and the CNS tuberculoma.47 These distinguishing 292 Thwaites, Chau, Mai, et al J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

characteristics may facilitate the diagnosis of children, with 86% having greater than 15 mm TBM in those with HIV. of induration with 5 units purified protein Elderly people with TBM are a significant derivative (PPD).51 group, particularly in the developed world. As Diagnosis is dependent on lumbar puncture with many conditions in elderly people, and CSF examination. Abnormalities in the presentation may be atypical. Signs of mening- CSF depend on a tuberculin reaction within ism may be absent, seizures occur more the subarachnoid space. Those with depressed commonly, and CSF findings may be atypical; cell mediated immunity may have atypical the CSF may even be acellular.49 findings in the CSF. Acellular CSF in elderly In summary, the diagnosis of TBM can nei- and HIV positive patients have been reported.49 ther be made nor excluded on clinical grounds. Lymphocytosis of between 100 and 1000 Coinfection with HIV does not seem to change cells/mm3 is more usual, although in the first the clinical manifestations or the outcome of 10 days polymorphonuclear leucocytes may TBM, although the diagnosis may be suggested predominate.59 A raised CSF protein occurs in by the presence of extrameningeal TB or CNS most, and CSF glucose will be reduced in tuberculoma. Elderly people may elude diag- 70%.45 59 nosis altogether unless carefully investigated. A The search for acid fast bacilli is the careful search for extrameningeal TB is likely most crucial part of the investigation. The to be a useful adjunct in establishing whether limit of detection on microscopy is 100 meningitis is due to TB. mycobacteria/ml.60 The clinician can assist the diagnostic yield in two ways: send a large Prognosis volume of CSF (10 ml is recommended), and Some studies have assessed the clinical and repeat the lumbar puncture if the diagnosis is laboratory indices that might predict outcome. suspected. Acid fast bacilli are seen in CSF The early trials used univariate analysis— smears in about 10% to 20% of those with assessing prognostic variables without adjust- TBM,61 although this figure varies consider- ing for the eVect of covariables.54 55 From these ably. The values in recent reviews were studies, some poor prognostic indicators 12.5%,45 37%,4 and 87%.60 The success of the arose—extremes of age, advanced stage of dis- test depends on the quality and volume of ease, concomitant extrameningeal TB, and sample sent, the skill of the technician, and evidence of raised intracranial pressure. Stud- their persistence in examining for acid fast ies employing multivariate analyses that adjust bacilli. for the influence of other variables are scarce. The culture of M tuberculosis from the CSF is One such study in children found that the age the gold standard for diagnosis, but is insensi- of the patient and stage of disease were two tive and slow. Laboratories employing only independent variables associated with solid media such as Lowenstein-Jensen may prognosis.56 A more recent study looked at take up to 8 weeks to culture M tuberculosis. clinical, laboratory, and CT features in 49 Semiautomated radiometric culture systems adults and children with TBM.57 A multivariate such as the Bactec 460 and automated logistic regression model showed that the most continuously monitored systems have reduced significant variables for predicting outcome in culture times.62 Although such systems do TBM were age, stage of disease, focal weak- reduce the time taken for culture the decision ness, cranial nerve palsy, and hydrocephalus. to treat the patient should not wait for culture http://jnnp.bmj.com/ The message for clinicians is simple: children results. with advanced disease with neurological com- The advent of CT and MRI has provided plications have poor outcomes. The interven- insight into disease progression, and gives tion required is rapid diagnosis and treatment. prognostic and diagnostic information.63 64 Both CT and MRI of the brain will disclose Diagnosis hydrocephalus, basilar meningeal thickening, The rapid diagnosis of TBM is fundamental to infarcts, oedema, and tuberculomas. clinical outcome. Current laboratory methods In a CT study of 60 cases of TBM in adults on October 2, 2021 by guest. Protected copyright. are insensitive and slow. Newer methods such and children only three had normal brain as those involving the amplification of bacterial scans.63 Hydrocephalus was reported in 87% of DNA by the polymerase chain reaction (PCR) children and 12% of adults. The incidence of and comparable systems are incompletely hydrocephalus is greater in the young, and assessed, and are not suitable for widespread increases with duration of the illness. In use in the developing world. The careful and children hydrocephalus is almost always repeated search for acid fast bacilli with Ziehl- present after 6 weeks of illness.63 Infarcts are Neelsen staining is still one of the most seen on CT in 28%, with 83% occurring in the eVective rapid diagnostic tests. middle cerebral artery territory.53 The basal The diagnosis of TBM cannot be made or ganglia are the most commonly aVected region. excluded on clinical grounds. A history of Poor prognosis has been associated with recent TB contact is helpful44 51 as is the enhancing basal exudates and periventricular presence of extrameningeal TB.50 56 Tuberculin lucency.53 64 testing is of limited value. Early studies found Magnetic resonance imaging has increased 22% of those with TBM were negative to 100 sensitivity in detecting the distribution of units PPD.58 A recent study demonstrated meningeal inflammatory exudate.53 Gadolin- cumulative reactivity with 10–100 units PPD ium enhancedT1weighted images highlight to be 75%.44 Some studies suggest that the exudate, and show parenchymal infarcts as tuberculin testing may be more useful in hyperintense areas; MRI may provide more Tuberculous meningitis 293 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

diagnostic information than CT when assess- metastases, and lymphoma. The major role of ing space occupying lesions. Cerebral miliary neuroradiology has been in management and TB, with multiple small intraparenchymal in particular in the diagnosis and follow up of granulomas, produces moderate perilesional those complications requiring neurosurgery. oedema and contrast enhancement. Larger The diagnostic dilemma faced on a daily tuberculomas are initially non-enhancing, but basis in many hospitals in the developing world later demonstrate marked enhancement. is illustrated in the two case histories (box). Both CT and MRI are sensitive to the Case 1 would seem like a straightforward case changes of TBM, particularly hydrocephalus of tuberculous meningitis with a lymphocytic and basal meningeal exudates, but they lack meningitis in a patient with partially treated speciﬁcity. The radiological diVerential diagno- pulmonary tuberculosis. Case 2, with a rela- sis includes cryptococcal meningitis, cytome- tively short history, normal chest radiograph, galovirus encephalitis, sarcoidosis, meningeal neutrophilic CSF, and recent retained placenta

Clinical case histories baby 4 days before admission, the delivery CASE HISTORY 1 being complicated by a retained placenta, for A 57 year old woman was admitted with a 5 which surgical intervention was required. She day history of fever, cough, neck stiVness, received an unknown antibiotic for 3 days after headache, and confusion. She was admitted this procedure. On admission her temperature to a provincial hospital where a clinical diag- was 40°C, and GCS 14 with marked neck nosis of bacterial meningitis was made. She stiVness. The chest radiograph was normal. was treated with an unknown antibiotic and Cerebrospinal fluid was clear with an opening transferred to the Centre for Tropical pressure of 40 cm. There were 320 white cells/ Diseases in Ho Chi Minh City. Her medical mm3 in the CSF with 90% neutrophils and history included partially treated pulmonary 10% lymphocytes. The glucose CSF/blood tuberculosis. On admission her temperature ratio was 0.7/7 mmol/l, CSF lactate was 7.8 was 39.3°C, Glasgow coma score (GCS) 10, mmol/l, and CSF protein was 62 mg/dl. Gram and she had neck stiVness, crackles in both and Zeihl-Neelsen stains were negative. Anti- lung fields, and bleeding from the upper gen detection for Streptococcus pneumoniae, gastrointestinal tract. Cerebrospinal fluid Neisseria meningitides, and haemophilus influen- was clear with an opening pressure of 21cm. zae in CSF was negative. A tuberculin skin test It contained 348/mm3 white cells (47% neu- was negative. The presumptive diagnosis was trophils 53% lymphocytes). The glucose bacterial meningitis possibly related to the CSF/blood ratio was 1.5/7.1 mmol/l, CSF surgical procedure after the delivery. She was lactate 6.4 mmol/l, and CSF protein 160 started on ceftriaxone, metronidazole, and mg/dl. Gram and Zeihl-Neelsen stains were tobramycin. Her clinical state remained static negative. Antigen detection for Streptococcus and a repeat CSF examination on day 4 pneumoniae, Neisseria meningitides, and Hae- showed white cells 560/mm3 (55% neu- mophilus influenzae in CSF was negative. A trophils. 45% lymphocytes). The glucose tuberculin skin test was negative and a chest CSF/blood ratio was 0.87/6 mmol/l, CSF lac- radiograph showed signs of pulmonary tu- tate was 9.4 mmol/l, and CSF protein was 173

berculosis (ﬁgure). Tuberculous meningitis mg/dl. The TB hospital was asked to review http://jnnp.bmj.com/ was considered the most likely diagnosis but her case and after 1 week elected to start she was initially started on ceftriaxone for a antituberculous therapy. Four weeks later possible partially treated bacterial meningi- Mycobacterium tuberculosis was isolated from tis. All therapy for tuberculosis in Ho Chi the admission CSF. This patient is now 5 Minh City is coordinated through the TB months into a 9 month course of TB therapy. hospital and they were asked to see her. Their These two cases represent the typical opinion was that this patient had dual dilemma that confronts much of the world in

pathology with a bacterial meningitis and the diagnosis of tuberculous meningitis. The on October 2, 2021 by guest. Protected copyright. pulmonary tuberculosis. There was no im- prevalence of pulmonary tuberculosis in a provement in the clinical situation at 48 country such as Viet Nam is high and hours. At this time the result of the CSF cul- patients with bacterial meningitis often have ture became available. Enterococcus faecium signs of TB on chest radiography. Pretreat- resistant to ceftriaxone was isolated from the ment with antibiotics is widespread and CSF. The antibiotic was changed to amoxy- levels of resistance to Streptococcus pneumo- cillin. After 3 days of amoxycillin the CSF niae are very high (more than 90% of showed 234 white cells/mm3 (40% neu- community isolates are resistant to penicil- trophils, 60% lymphocytes). The glucose lin). Consequently patients with bacterial CSF/blood ratio was 3.7/6.2 mmol/l, CSF meningitis often have negative gram stains Lactate was 1.87 mmol/l, and CSF protein and negative CSF culture on admission. The was 64 mg/dl. She went on to make an CSF cell counts, and glucose and protein uneventful recovery and was discharged well. results can be very similar in partially treated bacterial meningitis and early TBM, making CASE HISTORY 2 a clear diagnosis often impossible as these A 17 year old woman was admitted with an 8 two cases demonstrate. Making the wrong day history of fever, rigors, headache,and diagnosis, or delay in making the correct neck stiVness. She had delivered a normal diagnosis can have disastrous consequences. 294 Thwaites, Chau, Mai, et al J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

after delivery of a baby would certainly be con- ity and specificity, but may be useful in sistent with bacterial meningitis. However, 4 narrowing the diVerential diagnosis in aseptic weeks later the tubercle bacilli was cultured meningitides. from Lowenstein-Jensen media. There is Serological techniques that detect the in- clearly an urgent need for a sensitive and trathecal synthesis of antimycobacterial anti- specific aVordable diagnostic test in TBM. bodies have been studied. A good test will require an antigen with high species specificity and good immunogenicity to be sensitive. The Alternative diagnostic approaches use of crude antigens such as PPD results in The challenge facing new diagnostic strategies low sensitivity and specificity.69 Basic enzyme in TBM is that they must improve on the sen- linked immunosorbent assays (ELISAs) have sitivity of conventional Ziehl-Neelsen staining lacked sensitivity.69 70 The adaptation of ELISA and culture, but maintain the specificity. In the techniques and the identification of specific M developed world cost is less critical, but in the tuberculosis antigens have improved results. developing world cost considerations mandate Using a solid phase antibody competition assay tests that are cheap, use standard reagents with with mouse monoclonal antibodies to the 38 long shelf lives, and are technically undemand- kDA antigen (also known as antigen 5, or anti- ing. gen 78), a large study was performed in Tuberculosteric acid is a structural compo- pulmonary and extrapulmonary TB.71 In extra- nent of mycobacteria that was first detected in pulmonary TB diagnostic sensitivity was 73%, the CSF of a patient with TBM in 1983.65 Fre- specificity 98%, regardless of organ site. quency pulsed electron capture gas liquid Sensitivity improves when ELISA is used to chromatography has been used to detect detect anti-BCG secreting cells in the CSF of femtomole quantities of tuberculosteric acid in those with TBM72), but the test is technically CSF.66 The technique is unlikely to be adopted demanding. A sensitivity of 96% and specificity as standard diagnostic procedure due to its of 92% is reported with this method. A complexity despite 91% sensitivity and 95% cell-ELISA method allowing quantitative de- specificity being reported.66 tection of CSF anti-PPD IgG produced similar Adenosine deaminase is produced by lym- diagnostic sensitivity and specificity.73 phocytes and monocytes. Its detection in CSF The diVerentiation of acute infection from has been reported with variable success, with previous exposure is problematic in antibody sensitivities and specificities as high as 99% detection tests, and test sensitivity may be being suggested.67 A trial comparing concen- compromised in immunocompromised people. trations of adenosine deaminase in the CSF of Methods to directly detect specific mycobacte- those with aseptic meningitides found in- rial antigens in the CSF have been developed to creased concentrations in 30% of those with tackle these inadequacies. Initial studies used pyogenic meningitis, and almost universally various ELISA techniques70 74–76; most using raised concentrations in TBM and polyclonal antibodies directed against crude neurobrucellosis.68 The test lacks both sensitiv- antigen. Despite an expected lack of sensitivity and specificity, one retrospective study showed a sensitivity of 68% and specificity of 100% using these components.77 Other studies have claimed the identification of specific TB http://jnnp.bmj.com/ antigens and consequently specific serological tests based on them. For example, using a preparation of 35 kDa antigen from M tuberculosis, 100% sensitivity (when compared with culture) and 100% specificity was reported.78 The test was simple to perform and the nitrocellulose strips containing the antigen had a shelf life of 2 years. Unfortunately many on October 2, 2021 by guest. Protected copyright. assays showing early promise in highly controlled studies do not perform with high sensitivity and specificity in clinical practice. The advent of DNA amplification techniques such as PCR has turned attention away from serological techniques. In paucibacillary TB the concept of amplifying specific genetic material to detectable levels is attractive. Use of PCR in the diagnosis of TBM is promising, but still poorly defined. The few studies to date suVer from having small numbers, diVerent primer targets, and diVering diagnostic criteria. Sensitivities vary from 33% to 90%, specificities from 88% to 100%. Much of the variability is dependent on varying definitions of the diagnostic gold standard. Many use diVering clinical criteria for gold standard diagnosis, Chest radiograph. Case history 1. Loss of lung volume right side. DiVuse reticular shadowing, particulary right apex. Calcification right hilar and cavitation. Findings which makes interstudy comparison diYcult. A consistent with pulmonary tuberculosis. set of clinical diagnostic criteria has been Tuberculous meningitis 295 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

assessed against bacterial isolation, PCR, Management response to treatment, and necropsy.79 The Before the introduction of chemotherapy TBM PCR was positive in 75% of those clinically was almost universally fatal. Cases of transient adjudged to have highly probable or probable self limiting TBM are reported in the TBM, and whom improved on treatment. literature,87 but these are exceptional. The cur- Future studies need to use universal and evalu- rent United Kingdom guidelines for the 88 ated clinical diagnostic criteria. management of TBM reflect both the ad- Studies have assessed PCR against both bac- vances achieved by modern chemotherapy and terial culture, and ELISA for mycobacterial continuing areas of uncertainty. antibodies in TBM.80–82 One study reported Streptomycin was first used to treat pulmonary TB in 1944 and in 1946 the United King- culture alone as having a sensitivity and dom Medical Research Council (MRC) began specificity of 39% and 100%, whereas PCR studies using streptomycin. In 1948 they pub- had a sensitivity of 48% and a specificity of 81 83 lished data that demonstrated a marked 100%. A Vietnamese study compared 104 improvement in prognosis for those with TBM patients treated for TBM on clinical grounds treated with streptomycin.55 Mortality fell to and the results of initial CSF microscopy, cul- 46% in those presenting with stage 1 (con- ture, and PCR. They report the sensitivities of scious, no neurological deficit), 66% in stage 2 PCR to be 32%, culture 17% and microscopy (disturbed consciousness, with or without focal 1%. The importance of analysing adequate neurology), and 86% in stage 3 (comatose, volumes of CSF is emphasised. Of 17 patients with or without focal signs). The introduction with culture positive TBM only 10 were PCR of isoniazid and para-aminosalicylic acid led to positive. The authors explain the result by sug- further improvements in prognosis. A study gesting that the small quantities of CSF used documenting the changes in available chemo- for the PCR process eVectively reduces the therapy between 1947 and 1958 shows the available mycobacterial DNA to undetectable extent of the improvement.89 Mortality fell concentrations. Other authors emphasise this from 64% using streptomyin alone to 27% with “aliquot phenomenon” as being an important streptomyin and para-aminosalicylic acid, and cause of false negative PCR results.82 then to 17% with the addition of isoniazid. The use of PCR to monitor successful treat- The addition of rifampicin and pyrazina- ment in TBM is not yet defined. Studies have mide produced a further improvement in prognosis with a less toxic, orally administered suggested that PCR could detect M tuberculosis regime. The prognostic benefits of rifampicin up to 6 weeks after starting treatment.81 A small have been questioned,90 91 and uncertainty sur- study that performed PCR on sequential CSF rounds its penetration into the CSF. Ri- samples from seven patients with TBM, found fampicin is 80% protein bound in plasma, ena- that five were negative by day 14, and only one 84 bling a maximum of 20% to penetrate the CSF was positive at day 28. in those with an intact blood-brain barrier. Experience of PCR in respiratory samples Studies have shown slow penetration of ri- has led to the development of commercially fampicin into the CSF of patients with TBM, available tests.85 A trial has evaluated the use of with levels just above the minimum inhibitory Roche AMPLICOR PCR in CSF samples.82 concentrations for M tuberculosis.92 Meningeal Unfortunately only 37% of smear positive inflammation enhances CSF penetration of http://jnnp.bmj.com/ cases were culture positive. The reasons for this antitubercular drugs; however, there is limited include obtaining CSF after treatment was evidence to suggest that rifampicin penetration started, and limited culture techniques. The occurs independently of inflammation.93 result reduces their study to only eight cases There is no conclusive evidence to demon- with a gold standard diagnosis. Given these strate improvement in outcome with the use of numbers, sensitivity and specificity were 87% pyrazinamide. It is well absorbed orally, and 94 and 100% respectively. Other commercial kits has excellent penetration into the CSF. These have been tested, and have been shown to pro- factors, and the sterilising eVect on tubercle on October 2, 2021 by guest. Protected copyright. duce comparable results.86 bacilli, have resulted in pyrazinamide being In summary, PCR has an accepted role in the considered mandatory at the beginning of TBM treatment.88 95 It has been suggested that detection of M tuberculosis in pulmonary given the uncertain benefit and penetration of specimens,85 but is not yet fully evaluated for rifampicin, pyrazinamide should be given for the diagnosis of TBM. The sensitivity of PCR the duration of the treatment.51 on CSF samples seems to be only a moderate The current United Kingdom guidelines88 improvement on that of culture. The specificity suggest treatment for the first two months with is comparable, but depends on scrupulous rifampicin, isoniazid, pyrazinamide, and a laboratory technique to avoid DNA contami- fourth agent. This agent can be streptomycin, nation. As with Ziehl-Neelsen staining and ethambutol, or prothionamide. Streptomycin culture, yield is improved with greater volumes only penetrates the CSF to therapeutic concen- of CSF. The PCR is acknowledged to be inap- trations in the presence of inflammation.96 propriate for the developing world82 and Intrathecal administration is no longer recom- current studies suggest that PCR does not mended, given the availability of drugs that solve the global diagnostic challenge set by penetrate the CSF well. The renal and ototox- TBM. There is a need for novel diagnostic icity of streptomycin, and the necessity for drug approaches if sensitive and specific assay meth- concentration monitoring, has limited its use. ods are to become a reality. Ethambutol has two disadvantages that prompt 296 Thwaites, Chau, Mai, et al J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

some to suggest that, “there is little to the CSF of those with TBM.102 Although the recommend its use as a first line agent” in mechanism remains obscure, clinical trials TBM.51 Firstly, it penetrates CSF poorly, and suggest that corticosteroids have a beneficial secondly the adverse eVect of optic neuritis eVect in some groups of patients and a consen- which, although rare, limits the drug’s use for sus has emerged that adjuvant corticosteroids those in coma. Prothionamide is strongly should be used in those presenting with MRC favoured by some, particularly in South stage II or III TBM.88 95 103 Africa.51 95 Good concentrations in CSF are The evidence for this view is as follows. The achieved at a dose of 20 mg/kg.97 The drawback first controlled trial to suggest benefit in using is a foul metallic taste, commonly occurring corticosteroids for TBM was published in with nausea and vomiting. 1955.104 Of 12 patients with TBM six received In summary, a consensus exists that isoni- steroids in addition to streptomycin and isoniazid, rifampicin, and pyrazinamide constitute azid. The white count in CSF fell faster in the the best start to treatment. The addition of the steroid group, recovery from the acute phase fourth drug is left to local choice and was quicker, and none of the patients given experience, with little evidence to support the steroids had any long term sequelae. Four of use of one over the other. the six who did not receive corticosteroids had There is conflicting evidence for the duration chronic neurological sequelae. Trials confirm- of treatment. The current United Kingdom ing these results with larger numbers were not guidelines recommend 12 months in uncom- performed until the mid-1970s when a pro- plicated cases of TBM (including cerebral spective, randomised, double blind trial was tuberculoma without meningitis), extending to performed using 72 patients.105 A reduction in 18 months should pyrazinamide be omitted.88 mortality in the steroid group was shown, but No guidelines exist as to the components and the eVect on neurological morbidity could not duration of treatment in the case of multidrug be assessed. The largest prospective, ran- resistant TBM. domised, controlled trial to date enrolled 160 Treatment for 12 months is probably a con- patients with TBM.106 Overall mortality and servative estimate of the time required for bac- long term neurological sequelae were reduced terial cure. DiVerent regimes, incomparable in those treated with corticosteroids. The patient groups, and the variable use of adjuvant group that benefited the most were those with steroid therapy, makes meta-analysis from the disease of intermediate severity. Those present- trials impossible. Some suggest that TBM ing either in a coma or with mild disease (stage should be treated for a minimum of 2 years.98 I) received minimal benefit. Evidence from 781 cases of TBM treated for 2 Raised intracranial pressure has long been years showed that 35 had a recrudescence,98 considered important in the prognosis of but nearly all patients with relapse had received TBM.107 Reduction of intracranial pressure by less than 6 months of therapy, indicating that steroids was thought to be one of the means by therapy should be in excess of this period. Evi- which corticosteroids exerted their beneficial dence that 6 months of treatment can be eVect. A recent trial assessed the eYcacy of successful was first postulated in 196099 and steroids with regard to CT evidence of has been supported by more recent work.51 100 increased intracranial pressure, parenchymal Evidence from South Africa51 reported on 95 brain involvement, direct intracranial pressure children treated for 6 months with a combina- measurements, and clinical outcome.108 The http://jnnp.bmj.com/ tion of 20 mg/kg isoniazid, 20 mg/kg ri- trial showed no diVerence in intracranial pres- fampicin, 40 mg/kg pyrazinamide, and 20 sure, ventricular size, or extent of infarction mg/kg ethionamide; 96% of these cases pre- between those treated with or without steroids. sented in either stage II or III TBM. Pred- The benefit to mortality was again found in the nisolone at a dose of 4 mg/kg was randomly steroid group, and improved intellectual out- allocated to 40 of the children. The overall come was suggested. mortality was low at 16%, with only one case of It has been suggested that steroids may recrudescence. Prednisolone made no statisti- reduce the penetration of antituberculous on October 2, 2021 by guest. Protected copyright. cal diVerence to morbidity or mortality. The drugs into the CSF by reducing inflammation. doses of both isoniazid and rifampicin used There is little evidence for this occurring. One were considerably higher than that recom- study found no statistical diVerence between mended in the United Kingdom, but signifi- the plasma/CSF concentration ratios of isoni- cant adverse reactions were not reported. The azid, pyrazinamide, rifampicin, or streptomy- study provides good evidence for the adequacy cin, in those on or oV corticosteroids.109 of short course intensive chemotherapy, but the lack of a control group does not allow conclu- Role of neurosurgery sions as to optimal dosages. Studies using 9 Neurological deterioration occurring in a months chemotherapy (2 months of isoniazid, patient under treatment for TBM may have rifampicin, pyrazinamide, streptomycin, fol- various causes, and requires urgent radiological lowed by 7 months of rifampicin and isoniazid) assessment. Rising intracranial pressure re- at lower doses produced comparable quires active management. Hydrocephalus is a outcomes.101 common complication that may lead to perma- The rationale behind the use of adjuvant nent neurological damage or death if left corticosteroids lies in reducing the harmful untreated. Prompt assessment by CT is of eVects of inflammation as the antibiotics kill value in both diagnosis and management.64 the organisms. Corticosteroids do not seem to Repeated lumbar puncture or external ven- reduce the proinflammatory cytokines found in tricular drainage has been advocated in both Tuberculous meningitis 297 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

preventing and predicting the benefit of shunt considers this problem.118 119 Specific M tuber- surgery.110 111 Studies suggest that prompt ven- culosis phages carrying the firefly luciferase triculoatrial or ventriculoperitoneal shunting gene are able to infect viable M tuberculosis improves outcome, particularly in those who within a culture, thereby labelling them with present with minimal neurological deficit.111 the ability to produce light. This mechanism Indications for neurosurgical review include amplifies the detection of M tuberculosis allow- the presence of tuberculoma. These can ing rapid assessment of viability on a sensitivity develop or enlarge after the start of plate. Sensitivities of M tuberculosis are shown chemotherapy,112 provoking much speculation within days rather than weeks. The technique as to the immunological mechanism behind has also been used to detect M tuberculosis in this phenomenon.113 In practice, surgical inter- early cultures.119–123 vention is rarely required unless the tuberculoma is compromising a vital structure. Man- 1 World Health Organisation. The world health report. Geneva: agement is conservative, with good resolution WHO, 1998. 2 LS Farer, LM Lowell, MP Meador. Extrapulmonary tuber- reported with antituberculous chemotherapy culosis in the United States. Am J Epidemiology 1979;109: in combination with steroids.114 205–17. 3 LE Davis, KR Rastogi, LC Lambert, et al. Tuberculous meningitis in the Southwest United States: a community New developments or research avenues based study. Neurology 1993;43:1775–8. 4 H Pablos-Mendez, J Blustein, CA Knirsch. The role of The complete genome sequence of M tubercu- diabetes mellitus in the higher prevalence of tuberculosis losis strain H37Rv has recently been among Hispanics. Am J Public Health 1997;87:574–9. 27 5 MA Mori, G Leonardson, TK Welty. The benefits of isoni- determined. The sequence will enhance azid chemoprophylaxis and risk factors for tuberculosis research into vaccine design, mechanisms of among Ogala Sioux Indians. Arch Intern Med 1992;152: 547–50. drug resistance, and virulence determinants. 6 PA Selwyn, D Haitel, VA Lewis, et al. A prospective study of As yet specific mechanisms of neurovirulence the risk of tuberculosis among intravenous drug users with human immunodeficiency virus infection. N Engl J Med are unknown. It is conceivable that certain iso- 1989;320:345–50. lates of M tuberculosis may target specific recep- 7 E Bishburg, G Sunderam, LB Reichman, et al Central nervous system tuberculosis with the acquired immuno- tors facilitating meningeal involvement result- deficiency syndrome and its related complex. Ann Intern ing in a neurotropism analagous to that of Med 1986;105:210–13. 8 KM De Cock, B Soro, IM Coulibaly, et al. Tuberculosis and Mycobacterium leprae for peripheral nerves. HIV infection in sub-Saharan Africa. JAMA 1992;268: Skin testing with PPD has long been the 1581–7. 9 PM Small, PC Hopewell, SP Singh, et al. The epidemiology cornerstone for assessing exposure to M tuber- of tuberculosis in San Francisco: a population-based study culosis. The limitations to this test have led to using conventional and molecular methods. N Engl J Med 1994;330:1703–9. the development of in vitro cytokine assays 10 D Alland, GE Kalkut, AR Moss, et al. Transmission of assessing cell mediated immunity.115 Assays tuberculosis in New York City: an analysis by DNA finger- 116 printing and conventional epidemiological methods. N first developed in cattle, have been extended Engl J Med 1994;330:1710–6. to humans, and hold many potential advan- 11 BR Bloom, PM Small, The evolving relation between humans and Mycobacterium tuberculosis. N Engl J Med tages over skin testing. The production of 1998;338:677–8. interferon-ã in whole blood in response to spe- 12 W Stead, J Senner, W Reddick, et al. Racial diVerences in susceptibility to infection by mycobacterium tuberculosis. cific M tuberculosis or specific M tuberculosis N Engl J Med 1990;322:422–7. antigens shows promise. Potential applications 13 R Bellamy, C Ruwende, T Corrah, et al. Variations in the NRAMP1 gene and susceptibility to tuberculosis in West include the detection of infection with M tuber- Africans. N Engl J Med 1998;338:640–4. culosis,diVerentiating exposure to M tuberculo- 14 GA Colditz, TF Brewer, CS Berkley, et al.EYcacy of BCG

vaccine in the prevention of tuberculosis. JAMA 1994;271: http://jnnp.bmj.com/ sis from BCG vaccine or environmental myco- 698–702. bacteria, and assessing response to new 15 P De March-Ayuela. Trend in tuberculous meningitis in Barcelona in children 0–4 years: correlation with the vaccines. annual risk of tuberculosis infection. Tuber Lung Dis 1994; As the role of cytokines is understood, so 75:423–8. 16 MC Raviglione, D Snider, A Kochi. Global epidemiology of their use in diagnosis and therapeutics is tuberculosis. Morbidity and mortality of a woldwide expanding. Similar substances may also exist in epidemic. JAMA 1995;273:220–6. 17 FA Drobniewski, A Pablos-Mendes, M Raviglione. Epide- prokaryotic organisms. A factor has recently miology of tuberculosis in the world. Seminars in Respira- been characterised that promotes the resuscita- tory Critical Care Medicine 1997;18:419–29.

18 PH Green. Tubercular meningitis. Lancet 1836;ii:232–5. on October 2, 2021 by guest. Protected copyright. tion and growth of a group of gram positive 19 R Koch. Die aetiologie der Tuberculosos. Ber Klin Wochen- bacteria including M tuberculosis.117 In picomo- schr 1882;19:2211. 20 PA Jenkins. The microbiology of tuberculosis. In: PDO lar concentrations “resuscitation promoting Davies, ed. Clinical tuberculosis. London: Chapman and factor (Rpf)” dramatically increases the viable Hall Medical. 1994. 21 BR Bloom, PM Small. The evolving relationship between cell count of dormant cultures in organisms humans and Mycobacterium tuberculosis. N Engl J Med such as M tuberculosis. Acting as a “bacterial 1998;10:677–8. 22 DM Collins, RP Kawakami, GW De Lisle, et al. Mutation of cytokine” the ability of Rpf to resuscitate M the principal sigma factor causes loss of virulence in a strain tuberculosis has important implications for of mycobacterium tuberculosis complex. Proc Natl Acad Sci USA 1995;92:8036–40. future research. A rapid increase in bacterial 23 SE Valway, MPC Sanchez, TF Shinnick, et al. An outbreak yield from paucibacillary disease sites such as involving extensive transmission of a virulent strain of Mycobacterium tuberculosis. N Engl J Med 1988;338:633– CSF might be predicted, and may assist TBM 9. diagnosis. If growth and latency of M tuberculo- 24 Z Arvanitakis, RL Long, ES Hershﬁeld, et al. Mycobacte- rium tuberculosis moleculer variation in CNS infection. sis is controlled in vivo by such substances, Evidence for strain-dependent neurovirulence. Neurology novel therapeutic strategies or vaccines might 1998;50:1827–32. 25 S Falcow. Molecular Koch’s postulates applied to microbial be developed. pathogencity. Reviews of Infectious Diseases 1988;10:S274–6. As drug resistance becomes more prevalent 26 DM Collins, In search of tuberculosis virulence genes. Trends Microbiol 1996;4:426–430. so the requirement for rapid sensitivity testing 27 ST Cole, R Brosch, J Parkhill, et al. Deciphering the biology becomes more urgent, particularly in TBM of Mycobacterium tuberculosis from the complete genome sequence. Nature 1998;393:537–44. where inappropriate treatment can be fatal. 28 AR Rich, HA McCordick. The pathogenesis of tuberculous Rapid sensitivity testing using bacteriophages meningitis. Bulletin of John Hopkins Hospital 1933;52:5–37. 298 Thwaites, Chau, Mai, et al J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

29 EM Lincoln, SUR Sordillo, PA Davies. Tuberculous menin- 66 JB Brooks, MI Daneshvar, RL Haberberger, et al. Rapid gitis in children. J Paediatr 1960;57:807–23. diagnosis of tuberculous meningitis by frequency-pulsed 30 DK Dastur, DK Manghani, PM Udani. Pathology and electron-capture gas-liquid chromatography detection of pathogenetic mechanisms in neurotuberculosis. Radiol Clin carboxylic acids in cerebrospinal fluid. J Clin Microbiol North Am 1995;33:733–52. 1990;28:989–97. 31 FY Hsieh, LG Chia, WC Shen. Locations of cerebral 67 E Ribera, JM Martinez-Vazquez, I Ocana, et al. Activity of infarctions in tuberculous meningitis. Neuroradiology 1992; adenosine deaminase in cerebrospinal fluid for the diagno- 34:197–9. sis and follow-up of tuberculous meningitis in adults. J 32 DK Dastur, PM Udani. Pathology and pathogenesis of Infect Dis 1987;155:603–7. tuberculous encephalopathy. Acta Neuropathol 1966;6:311– 68 LF Lopez-Cortes, M Cruz-Ruiz, J Gomez-Mateos, et al. 26. Adenosine deaminase activity in the CSF of patients with 33 PM Udani, PK Dastur. Tuberculous encephalopathy with aseptic meningitis: utility in the diagnosis of tuberculous and without meningitis: clinical features and pathological meningitis or neurobrucellosis. Clin Infect Dis 1995;20: correlations. J Neurol Sci 1970;10:541–61. 525–30. 34 AM Dannenberg. Delayed-type hypersensitivity and cell- 69 SB Kalish, RC Radin, D Levitz, et al. The enzyme-linked mediated immunity in the pathogenesis of tuberculosis. immunosorbent assay method for IgG antibody to purified Immunol Today 1991;12:228–33. protein derivative in cerebrospinal fluid of patients with 35 MB Lurie. Resistance to tuberculosis: experimental studies in tuberculous meningitis. Ann Intern Med 1983;99:630–3. native and acquired defensive mechanisms. Cambridge, MA: 70 G Watt, G Zaraspe, S Bautista, et al. Rapid diagnosis of Harvard University Press, 1964. tuberculous meningitis by using an enzyme-linked 36 V Kinder, AP Sappino, GE Grau, et al. The inducing role of immunosorbent assay to detect mycobacterial antigen and tumour necrosis factor in the development of bacteriacidal antibody in cerebrospinal fluid. J Infect Dis 1988;158:681– granulomas during BCG infection. Cell 1989;56:731–40. 6. 37 IM Orme. The immunopathogenesis of tuberculosis: a new 71 EGL Wilkins, J Ivanyi. Potential value of serology for diag- working hypothesis. Trends Microbiol 1998;6:94–7. nosis of extrapulmonary tuberculosis. Lancet 1990;336: 38 MK Sharief, M Ciardi, EJ Thompson. Blood brain barrier 641–4. damage in patients with bacterial meningitis: association 72 C-Z Lu, J Qiao, T Shen, et al. Early detection of tuberculous with tumour necrosis factor á but not interleukin-1 â. J meningitis by detection of anti-BCG secreting cells in cer- Infect Dis 1992;166:350–8. ebrospinal fluid. Lancet 1990;336:10–13. 39 EA Filley, GA Rook. EVect of mycobacteria on sensitivity to 73 SM Baig. Anti-purified protein derivative cell-enzyme- the cytotoxic eVects of tumour necrosis factor. Infect Immu- linked immunosorbent assay, a sensitive method for the nol 1991;59:2567–72. early diagnosis of tuberculous meningitis. J Clin Microbiol 40 L Tsenova, K Sokol, V Freedman, et al. A combination of 1995;33:3040–1. thalidomide plus antibiotics protects rabbits from myco- 74 E Sada, GM Ruiz-Palacios, Y Lopez-Vidal, et al. Detection bacterial meningitis-associated death. J Infect Dis 1998; of mycobacterial antigens in cerebrospinal fluid of patients 177:1563–72. with tuberculous meningitis by immunosorbent assay. Lan- 41 GAW Rook, R Al Attiyah. Cytokines and the Koch cet 1983;ii:651–2. phenomena. Tubercle 1991;72:13–20. 75 GV Kadival, TB Mazarelo, SD Chaparas. Sensitivity and 42 M Parsons. Tuberculous meningitis. A handbook for clinicians. specificity of enzyme-linked immunosorbent assay in the Oxford: Oxford University Press, 1979. detection of antigen in tuberculous meningitis cerebrospi- 43 RS Illingworth. Miliary and meningeal tuberculosis. Lancet nal fluids. J Clin Microbiol 1986;23:901–4. 1956;ii:646–9. 76 E Krambovitis, MB McIlmurray, PE Lock, et al. Rapid 44 SJ Kent, SM Crowe, A Yung, et al. Tuberculous meningitis: diagnosis of tuberculous meningitis by latex particle agglu- a 30 year review. Clin Infect Dis 1993;17:987–94. tination. Lancet 1984;ii:1229–31. 45 R Verdon, S Chevret, JP Laissy, et al. Tuberculous meningi- 77 VV Radhakrishnan, S Sehgal, A Mathai. Correlation tis in adults: review of 48 cases. Clin Infect Dis between culture of Mycobacterium tuberculosis and detec- 1996;22:982–8. tion of mycobacterial antigens in cerbrospinal fluid of 46 J Berenguer, S Moreno, F Laguna, et al. Tuberculous men- patients with tuberculous meningitis. J Med Microbiology ingitis in patients with human immunodeficiency virus. N 1990;33:223–6. Engl J Med 1992;326:668–72. 78 A Mathai, V.V Radhakrishnan, S Shobha. Diagnosis of 47 MP Dube, PD Holtom, RA Larsen, Tuberculous meningitis tuberculous meningitis confirmed by means of an immu- in patients with and without human immunodeficiency noblot method. J Infect 1994;29:33–9. virus infection. Am J Med 1992;93:520–4. 79 GK Ahuja, KK Mohan, K Prasad, et al. Diagnostic criteria 48 VK Yechoor,WX Shandera, P Rodriguez, et al. Tuberculous for tuberculous meningitis and their validation. Tuber Lung meningitis among adults with and without HIV infection. Dis 1994;75:149–52. Arch Int Med 1996;156:1710–16. 80 P Shankar, N Manjunath, KK Mohan, et al. Rapid diagnosis 49 AS Karstaedt, S Valtchanova, R Barriere, et al. Tuberculous of tuberculous meningitis by polymerase chain reaction. meningitis in South African urban adults. QJMed Lancet 1991;337:5–7. 1988;91:743–7. 81 LFF Kox, S Kuijper, AHJ Kolk. Early diagnosis of tubercu- 50 J Leonard, RM Des Prez. Tuberculous meningitis. Infect Dis lous meningtitis by polymerase chain reaction. Neurology Clin North Am 1990;4:769–87. 1995;45:2228–32. 51 PR Donald, JF Schoeman, LE Van Zyl, et al. Intensive short 82 A Bonington, JI Strang, PE Klapper, et al. Use of Roche course chemotherapy in the mangement of tuberculous AMPLICOR Mycobacterium tuberculosis PCR in early http://jnnp.bmj.com/ meningitis. International Journal of Tuberculosis and Lung diagnosis of tuberculous meningitis. J Clin Microbiol 1998; Diseases 1998;2:704–11. 36:1251–4. 52 MJ Humphries, WK Lam, R Teoh. Non-respiratory 83 LN Nguyen, LFF Kox, D Linh, et al. The potential contri- tuberculosis. In: PDO Davies, ed. Clinical tuberculosis. bution of polymerase chain reaction to the diagnosis of London: Chapman and Hall Medical 1994. tuberculous meningitis. Arch Neurol 1996;53:771–6. 53 T Tartaglione, GM Di Lella, A Cesare, et al. Diagnostic 84 JJ Lin, HJ Harn. Application of polymerase chain reaction to imaging of neurotuberculosis. Rays 1998;23:164–80. monitor Mycobacterium tuberculosis DNA in the CSF of 54 PD Gulati, GP Mathur, H Vaishnavah. Prognosis and patients with tuberculous meningitis after antibiotic sequelae of tuberculous meningitis in adults. J Assoc Physi- treatment. J Neurol Neurosurg Psychiatry 1995;242:147–52. cians India 1970;18:281–6. 85 RF D’Amato, AA Wallman, LH Hochstein, et al. Rapid

55 Medical Research Council. Streptomycin in tuberculous diagnosis of pulmonary tuberculosis by using Roch on October 2, 2021 by guest. Protected copyright. trials committee. Streptomycin treatment of tuberculous AMPLICOR Mycobacterium tuberculosis PCR test. J Clin meningitis. Lancet 1948;i:582–97. Microbiol 1995;33:1832–4. 56 MJ Humphries, R Teoh, J Lau, et al. Factors of prognostic 86 AM Lang, J Feris-Iglesias, C Pena, et al. Clinical evaluation significance in Chinese children with tuberculous meningi- of the Gen-Probe amplified direct test for detection of tis. Tubercle 1990;71;161–8. Mycobacterium tuberculosis complex organisms in cer- 57 UK Misra, J Kalita, M Srivastava, et al. Prognosis of tuber- ebrospinal fluid. J Clin Microbiol 1998;36:2191–4. culous meningitis: a multivariate analysis. J Neurol Sci 87 RTD Emond, GSW McKendrick. Tuberculosis as a cause 1996;137:57–61. of transient aseptic meningitis. Lancet 1973;ii:234–6. 58 HV Smith, RL Vollum. The diagnosis of tuberculous men- 88 Joint Tuberculosis Committee of the British Thoracic Soci- ingitis. Br Med Bull 1954;10:140–4. ety. Chemotherapy and management of tuberculosis in the 59 T Jeren, I Beus. Characteristics of cerebrospinal fluid in United Kingdom: recommendations 1998. Thorax 1998; tuberculous meningitis. Acta Cytol 1982;26:678–80. 53:536–48. 60 DH Kennedy, RJ Fallon. Tuberculous meningitis. JAMA 89 J Lorber. Treatment of tuberculous meningitis. BMJ 1960;i: 1979;241:264–8. 1309–12. 61 PC Hopewell. Overview of clinical tuberculosis. In: BR 90 P Ramachandran, M Duraipandian, M Nagarajan, et al. Bloom, Tuberculosis: pathogenesis, protection, and control. Three chemotherapy studies of tuberculous meningitis in Washington, DC: ASM, 1994. children. Tubercle 1986;67:17–29. 62 SH Gillespie, TD McHugh. The genus Mycobacterium. In, 91 P Ramachandran, M Duraipandian, AM Reetha, et al. AM Emmerson, PM Hawkey, SH Gillespie, eds. Principles Long-term status of children treated for tuberculous men- and practice of clinical bacteriology. Chichester: John Wiley, ingitis in South India. Tubercle 1989;70:17–29. 1997. 92 GA Ellard, MJ Humphries, BW Allen. Cerbrospinal fluid 63 S Bhargava, AK Gupta, PN Tandon. Tuberculous drug concentrations and treatment of tuberculous menin- meningitis: a CT study. Br J Radiol 1982;55:189–96. gitis. Am Rev Respir Dis 1993;148:650–5. 64 MRR Bullock, JM Welchman. Diagnostic and prognostic 93 R Nau, HW Prange, S Menck, et al. Penetration of features of tuberculous meningitis on CT scanning. J Neu- rifampicin into CSF of adults with uninflamed meninges. J rol Neurosurg Psychiatry 1982;45:1098–101. Antimicrob Chemother 1992;29:719–24. 65 PA Mardh, L Larsson, N Holby, et al. Tuberculostearic acid 94 R Forgan-Smith, G Ellard, D Newton, et al. Pyrazinamide as a diagnositic marker in tuberculous meningitis. Lancet and other drugs in tuberculous meningitis. Lancet 1973;ii: 1983;i:367. 374. Tuberculous meningitis 299 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.68.3.289 on 1 March 2000. Downloaded from

95 M Humphries. The management of tuberculous meningitis. 111 R Palur, V Rajshekhar, MJ Chandy, et al. Shunt surgery for Thorax 1992;47:577–81. hydrocephalus in tuberculous meningitis: a long term 96 DH Heilman, FR Heilman, HK Hinshaw, et al. follow up study. J Neurosurg 1991;74:64–9. Streptomycin: absorption, diVusion, excretion, and toxic- 112 B Afghani, JM Lieberman. Paradoxical enlargement or ity. Am J Med Sci 1945;210:567–82. development of intracranial tuberculomas during therapy: 97 PR Donald, HI Seifart. Cerebrospinal fluid concentrations case report and review. Clin Infect Dis 1994;19:1092–9. of ethionamide in children with tuberculous meningitis. J 113 BG Marshall, MA Chambers. Central nervous system Paeds 1989;115:483–6. 98 A Goel, S Pandya, A Satoskar. Whither short-course tuberculosis: the paradox of the host immune response. J chemotherapy for tubercuous meningitis. Neurosurgery Infect 1998;36:3–4. 1990;27:418–21. 114 PH Tandin, S Bhargaga. EVect of medical treatment in 99 EL Kendig, CD Burch. Short-term antimicrobial therapy of intracranial tuberculoma: a CT study. Tubercle 1985;66:85. tuberculous meningitis. Am Rev Respir Dis 1960;82:672– 115 AD Lein, CF Von Reyn. In vitro cellular and cytokine 81. responses to mycobacterial antigens: application to diagno- 100 RF Jacobs, P Sunakorn, T Chotpitayasunonah, et al. Inten- sis of tuberculous infection and assessment of response to sive short course chemotherapy for tuberculous meningitis. mycobacterial vaccines. Am J Med Sci 1997;313:364–71. Paediatric Infectious Disease Journal 1992;11:194–8. 116 PR Wood, JS Rothel. In vitro immunodiagnostic assays for 101 P Phuapradit, A Vejjajiva. Treatment of tuberculous bovine tuberculosis. Vet Microbiol 1994;40:125–35. meningitis: role of short-course chemotherapy. QJMed 117 GL Mukamolova, AS Kaprelyants, DI Young, et al. A bac- 1987;62:249–58. terial cytokine. Proc Natl Acad Sci USA 1998;95:8916–21. 102 PR Donald, JF Schoeman, N Beyers. Concentrations of 118 WR Jacobs, RG Barletta, R Udani. Rapid assessment of interferon ã, tumour necrosis factor á, and interleukin-1 â drug susceptibilities of MTB by means of luciferase in the cerebrospinal fluid of children treated for tuberculous meningitis. Clin Infect Dis 1995;21:924–9. reporter phages. Science 1993;260:819–22. 103 D Dooley, JL Carpenter, S Rademacher. Adjunctive 119 FA Drobniewski, SM Wilson. The rapid diagnosis of isoni- corticosteroid therapy for tuberculosis: a critical reappraisal azid and rifampicin drug resistance in Mycobacterium of the literature. Clin Infect Dis 1997;25:872–87. tuberculosis: a molecular story. J Med Microbiol 1997;47: 104 M Ashby, H Grant. Tuberculous meningitis treated with 189–96. cortisone. Lancet 1955;i:65–6. 120 C Carriere, PF Riska, O Zimhony, et al. Conditionally rep- 105 J Escobar, MA Belsey, A Duenas, et al. Mortality from licating reporter phages: improved sensitivity for rapid tuberculous meningitis reduced by steroid therapy. Paediat- detection and assessment of drug susceptibility of myco- rics 1975;56:1050–5. bacterium tuberculosis. J Clin Microbiol 1997;35:3232–9. 106 NI Girgis, Z Farid, ME Kilpatrick, et al. Dexamethasone 121 SM Wilson, Z Al-Suwaidi, R McNerney, et al. Evaluation adjunctive treatment for tuberculous meningitis. Paediatric of a new rapid bacteriophage-based method for the drug Infectious Disease Journal 1991;10:179–83. susceptibility testing of Mycobacterium tuberculosis.Nat 107 JM Leonard, RM Prez. Tuberculous meningitis. Infect Dis Med 1997;3:465–8. Clin North Am 1990;4:769–87. 122 IJ Eltringham, FA Drobniewski, JA Mangan, et al. Evalua- 108 JF Schoeman, LE Van Zyl, JA Laubscher, et al.EVect of tion of RT-PCR and a bacteriphage based assay for the corticosteroids on intracranial pressure, computed tomo- graphic findings and clinical outcome in young children rapid phenotypic detection of rifampicin resistance in clini- with tuberculous meningitis. Paediatrics 1997;99:226–31. cal isolates of Mycobacterium tuberculosis. J Clin Microbiol 109 S Kaojarern, K Supmonchai, P Phuapradit, et al.EVect of 1999;37:3524–7. steroids on cerebrospinal fluid penetration of anti- 123 IJ Eltringham, SM Wilson, FA Drobniewski. Evaluation of tuberculous drugs in tuberculous meningitis. Clin Pharma- a bacteriophage based assay (Phab assay) as a rapid screen col Ther 1991;49:6–12. for resistance to isoniazid, ethambutol, streptomycin, 110 PK Newman, WJK Cumming, JB Foster. Hydrocephalus pyrazinamide, and ciprofloxacin among clinical isolates of and tuberculous meningitis in adults. J Neurol Neurosurg Mycobacterium tuberculosis. J Clin Microbiol 1999;37: Psychiatry 1980;43:188–90. 3528–32. http://jnnp.bmj.com/ on October 2, 2021 by guest. Protected copyright.