MAJOR ARTICLE

Uniform Research Case Definition Criteria

Differentiate Tuberculous and Bacterial
Meningitis in Children
Regan S. Solomons,1,a Marie Wessels,1,a Douwe H. Visser,2 Peter R. Donald,1 Ben J. Marais,3 Johan F. Schoeman,1

Downloaded from https://academic.oup.com/cid/article-abstract/59/11/1574/1746435 by guest on 11 April 2019

and Anne M. van Furth2
1
Department of Paediatrics and Child Health, Faculty of Medicine and Health Sciences, Stellenbosch University, Cape Town, South Africa; 2Department of
Pediatric Infectious Diseases and Immunology, VU University Medical Center, Amsterdam, The Netherlands; and 3Marie Bashir Institute for Infectious
Diseases and Biosecurity Institute and The Children’s Hospital at Westmead, The University of Sydney, Australia

Background. Tuberculous meningitis (TBM) research is hampered by low numbers of microbiologically con-
firmed TBM cases and the fact that they may represent a select part of the disease spectrum. A uniform TBM research
case definition was developed to address these limitations, but its ability to differentiate TBM from bacterial men-
ingitis has not been evaluated.
Methods. We assessed all children treated for TBM from 1985 to 2005 at Tygerberg Children’s Hospital, Cape
Town, South Africa. For comparative purposes, a group of children with culture-confirmed bacterial meningitis,
diagnosed between 2003 and 2009, was identified from the National Health Laboratory Service database. The per-
formance of the proposed case definition was evaluated in culture-confirmed TBM and bacterial meningitis cases.
Results. Of 554 children treated for TBM, 66 (11.9%) were classified as “definite TBM,” 408 (73.6%) as “probable
TBM,” and 72 (13.0%) as “possible TBM.” “Probable TBM” criteria identified culture-confirmed TBM with a sen-
sitivity of 86% and specificity of 100%; sensitivity was increased but specificity reduced when using “possible TBM”
criteria (sensitivity 100%, specificity 56%).
Conclusions. “Probable TBM” criteria accurately differentiated TBM from bacterial meningitis and could be
considered for use in clinical trials; reduced sensitivity in children with early TBM (stage 1 disease) remains a
concern.
Keywords. tuberculous meningitis; case definition; diagnostic value.

Tuberculosis continues to pose a major global health symptoms such as cough, weight loss, fever, vomiting,
challenge with a high morbidity and mortality, despite and lethargy. With disease progression, a more definitive
effective antituberculosis medication, in tuberculosis- clinical picture becomes manifest [3]; however, early
endemic areas [1, 2]. Tuberculous meningitis (TBM) is treatment initiation is critical to optimize outcome [4, 5].
the most devastating manifestation of tuberculosis and Identification of Mycobacterium tuberculosis in cere-
is generally associated with poor outcomes. Diagnosis is brospinal fluid (CSF) provides bacteriological confirma-
often delayed due to the nonspecific nature of early tion of TBM. Unfortunately, due to the paucibacillary
nature of TBM, CSF microscopy for acid-fast bacilli
(AFB) [6] and CSF culture for M. tuberculosis has low
Received 2 June 2014; accepted 12 August 2014; electronically published 19 Au- sensitivity (<20% and <50%, respectively) [3, 7, 8] com-
gust 2014.
a
R. S. S. and M. W. contributed equally to this work. pared with clinical criteria. Although culture provides
Correspondence: Regan S. Solomons, MMed, Department of Paediatrics and Child the accepted “gold standard,” it has little clinical utility
Health, Faculty of Medicine and Health Sciences, Stellenbosch University, PO Box
19063, Tygerberg, Cape Town 7505, South Africa (regan@sun.ac.za).
as it takes up to 42 days to confirm a positive result. A
Clinical Infectious Diseases® 2014;59(11):1574–8 meta-analysis of the accuracy of commercial nucleic
© The Author 2014. Published by Oxford University Press on behalf of the Infectious acid amplification tests (NAATs) for the diagnosis of
Diseases Society of America. All rights reserved. For Permissions, please e-mail:
journals.permissions@oup.com.
TBM, mostly adult studies, showed a sensitivity of
DOI: 10.1093/cid/ciu665 64% and specificity of 98% compared with culture [9].

1574 • CID 2014:59 (1 December) • Solomons et al

Thus it may serve as a rapid “rule-in” test, but does not provide Table 1. Diagnostic Criteria in the Uniform Tuberculous
a reliable “rule out” test. Meningitis Research Case Definition
TBM research is greatly hampered by the absence of reliable
diagnostic criteria and standardized approaches [10]. A recent re- Diagnostic
Criteria Score
view found that existing trials assessing antituberculosis treatment
for TBM had limited power and poor methodology and used Clinical criteria (maximum category score = 6)
Symptom duration of >5 d 4
varying treatment regimens with conflicting results [11]. Answer-
Systemic symptoms suggestive of tuberculosis (≥1): 2
ing critical research questions will require multicenter random- weight loss/(poor weight gain in children), night
ized controlled trials using standardized diagnostic and staging sweats, or persistent cough >2 wk
criteria, with sufficient patient numbers to demonstrate differenc- History of recent close contact with an individual with 2
pulmonary tuberculosis or a positive TST/IGRA in a
es in outcome [11]. Because it is difficult to obtain adequate sam- child aged <10 y
ple sizes using only culture-confirmed TBM cases, inclusion of Focal neurological deficit (excluding cranial nerve 1
probable and possible TBM cases requires consideration. palsies)

Downloaded from https://academic.oup.com/cid/article-abstract/59/11/1574/1746435 by guest on 11 April 2019

A uniform research case definition for TBM in adults and Cranial nerve palsy 1
children was developed to improve standardization of diagnosis CSF criteria (maximum category score = 4)
Clear appearance 1
for research purposes [12]. Classification as definite, probable,
Cells: 10–500/µL 1
or possible TBM is based on a composite score of clinical find-
Lymphocytic predominance (>50%) 1
ings and special investigations (see Table 1). Individual criteria
Protein concentration >1 g/L 1
and their relative weights were assigned based on an extensive CSF: plasma glucose ratio of <50% or an absolute 1
literature review together with international expert consensus CSF glucose concentration <2.2 mmol/L
[12–17]. The accuracy of proposed criteria for probable and Cerebral imaging criteria (maximum category score = 6)
possible TBM has not been assessed. Our study aimed to eval- Hydrocephalus (CT and/or MRI) 1
uate the diagnostic performance of probable and possible TBM Basal meningeal enhancement (CT and/or MRI) 2
criteria in children with culture-confirmed TBM and culture- Tuberculoma (CT and/or MRI) 2
Infarct (CT and/or MRI) 1
confirmed bacterial meningitis.
Precontrast basal hyperdensity (CT) 2
Evidence of tuberculosis elsewhere (maximum category score = 4)
PATIENTS AND METHODS Chest radiograph suggestive of active tuberculosis 2
(excludes miliary tuberculosis)
We conducted a retrospective review of all children (<13 years Chest radiograph suggestive of miliary tuberculosis 4
of age) treated for TBM between January 1985 and April 2005 at CT/MRI/US evidence for tuberculosis outside the 2
CNS
Tygerberg Children’s Hospital, Cape Town, South Africa. Clin-
AFB identified or Mycobacterium tuberculosis 4
ical data were prospectively collected in all TBM patients as part cultured from another source, ie, sputum, lymph
of ongoing studies conducted at the time. For comparative pur- node, gastric washing, urine, blood culture
poses, we identified a group of children (<13 years of age) with Exclusion of alternative diagnoses: an alternative diagnosis must be
confirmed microbiologically, serologically, or histopathologically
culture-confirmed bacterial meningitis, diagnosed between Jan-
Definite TBM = AFB seen on CSF microscopy, positive CSF
uary 2003 and November 2009. Bacterial meningitis cases were M. tuberculosis culture, or positive CSF M. tuberculosis commercial
identified from the National Health Laboratory Service database NAAT in the setting of symptoms/signs suggestive of meningitis; or
AFB seen in the context of histological changes consistent with
at Tygerberg Children’s Hospital; relevant clinical data were re- tuberculosis brain or spinal cord together with suggestive symptoms/
trieved from patient files. The study was approved by the signs and CSF changes, or visible meningitis (on autopsy).
Human Research Ethics Committee of Stellenbosch University Probable TBM = total score of ≥12 when neuroimaging available =
total score of ≥10 when neuroimaging unavailable
(reference number N10/11/367).
Possible TBM = total score of 6–11 when neuroimaging available =
total score of 6–9 when neuroimaging unavailable
CASE DEFINITIONS Source: Marais et al [12].
Abbreviations: AFB, acid-fast bacilli; CNS, central nervous system; CSF,
Tuberculous Meningitis cerebrospinal fluid; CT, computed tomography; IGRA, interferon-γ release
assay; MRI, magnetic resonance imaging; NAAT, nucleic acid amplification
Patients were classified as “definite TBM” if the CSF culture was
test; TBM, tuberculous meningitis; TST, tuberculin skin test; US, ultrasound.
positive for M. tuberculosis; smear microscopy for AFB and/or
commercial NAAT testing was not performed at the time.
Identification of probable and possible TBM was based on (2) CSF results, (3) neuroimaging findings, (4) evidence for tu-
predefined diagnostic scoring criteria (Table 1) [12]. Points berculosis outside the central nervous system, and (5) additional
were allocated in the following categories: (1) clinical findings, laboratory criteria. A total score of at least 12 was compatible

Value of Uniform Case Definition in TBM • CID 2014:59 (1 December) • 1575

Table 2. Overview of Scoring Criteria in Children With Definite, Probable, and Possible Tuberculous Meningitis Compared With Bacterial
Meningitis

Criteria Definite TBM Probable TBM Possible TBM Definite BM

Total number 66 408 72 32
Stage I 2 (3) 5 (1) 4 (6) NA
Stage IIa 35 (53) 214 (52) 47 (65) NA
Stage IIb 2 (3) 16 (4) 1 (1) NA
Stage III 27 (41) 172 (42) 18 (25) NA
Unknown stage 0 1 2 NA
Symptom duration >5 d 47/66 (71) 366/408 (90) 40/72 (56) 6/32 (19)
≥1 of weight loss/poor weight gain, night sweats, or 31/66 (47) 204/408 (50) 25/72 (35) 1/32 (3)
persistent cough >2 wk

Downloaded from https://academic.oup.com/cid/article-abstract/59/11/1574/1746435 by guest on 11 April 2019

History of recent tuberculosis contact or positive TST or IGRA 33/65 (52) 291/374 (78) 36/66 (55) 2/32 (6)
Focal neurological deficit 33/66 (50) 286/408 (70) 27/72 (38) 2/32 (6)
Cranial nerve palsy 18/66 (27) 110/408 (27) 17/72 (24) 2/32 (6)
Altered consciousness 46/66 (70) 398/408 (98) 65/72 (90) 19/32 (59)
CSF clear 66/66 (100) 408/408 (100) 72/72 (100) 18/32 (56)
CSF cells: 10–500/ µL 55/63 (87) 361/396 (91) 53/54 (98) 20/32 (63)
CSF lymphocyte predominance (>50%) 51/63 (81) 326/396 (82) 53/54 (98) 19/32 (59)
CSF protein concentration >1 g/L 49/61 (80) 292/387 (72) 42/52 (81) 15/32 (47)
CSF: serum glucose ratio <50% and/or CSF glucose 47/59 (80) 263/361 (73) 34/48 (71) 17/32 (53)
concentration <2.2 mmol/L
Hydrocephalus (CT and/or MRI) 60/62 (97) 353/391 (90) 51/62 (82) 9/21 (43)
Basal meningeal enhancement (CT and/or MRI) 48/62 (77) 318/391 (81) 19/62 (31) 1/21 (5)
Tuberculoma (CT and/or MRI) 5/62 (8) 55/391 (14) 5/62 (8) 1/21 (5)
Infarct (CT and/or MRI) 15/62 (24) 142/391 (36) 3/62 (5) 3/21 (14)
Precontrast basal hyperdensity (CT) No data No data No data 0
CXR suggestive of active tuberculosis 26/66 (39) 206/398 (52) 17/66 (26) 3/32 (9)
CXR suggestive miliary tuberculosis 11/66 (17) 53/398 (13) 1/66 (2) 0
Extraneural radiological tuberculosis 0 0 0 0
Extraneural Mycobacterium tuberculosis confirmation 19/66 (29) 100/408 (25) 2 (3) 0

Data are presented as no/No. (%).

Abbreviations: BM, bacterial meningitis; CSF, cerebrospinal fluid; CT, computed tomography; CXR, chest radiograph; IGRA, interferon-γ release assay; MRI,
magnetic resonance imaging; NA, not applicable; TBM, tuberculous meningitis; TST, tuberculin skin test.

with probable TBM (when neuroimaging was unavailable, the tomography (CT) and imaging of tuberculosis outside the cen-
total score was reduced to at least 10), whereas a total score of tral nervous system (apart from standard posteroanterior and
6–11 equated to a possible TBM diagnosis (when neuroimaging lateral chest radiographs). TBM disease staging was determined
was unavailable, the total score required was reduced to 6–9), as follows: stage I, Glasgow Coma Scale (GCS) of 15 and no
with a minimum number of points coming from CSF or neuro- focal neurological signs; stage IIa, GCS of 15 plus focal neuro-
imaging criteria. logical signs; stage IIb, GCS of 11–14 plus focal neurological
Potential alternative causes had to be excluded by performing signs; and stage III, GCS <11 [17, 18].
at least (1) CSF Gram stain and bacterial culture and (2) CSF
India ink stain, cryptococcal antigen latex agglutination test, Bacterial Meningitis
and/or cryptococcal culture. These investigations were routinely A diagnosis of definite bacterial meningitis was made if bacte-
performed. Additional investigations such as (3) viral culture rial pathogens were identified by Gram stain and/or culture
and/or polymerase chain reaction (PCR) and tests to exclude from CSF, and patients were treated for bacterial meningitis.
(4) neurosyphilis and (5) cerebral malaria should be conducted
if clinically or epidemiologically indicated; these investigations Statistical Analysis
were not routinely performed in our cohort. Due to the retro- Statistical analysis was done using SPSS version 20 (SPSS Inc,
spective nature of the study, the following imaging criteria could Chicago, IL) and SAS version 9.1 (SAS Institute Inc, Cary,
not be evaluated: precontrast basal hyperdensity on computed NC). The Fisher exact test and odds ratios (ORs) were used in

1576 • CID 2014:59 (1 December) • Solomons et al

univariable analyses comparing “definite TBM” and “definite Table 3. Findings on Univariate Analysis Comparing Definite
bacterial meningitis,” using a 2-tailed test for statistical Tuberculous Meningitis to Definite Bacterial Meningitis
significance.
P
Finding OR (95% CI) Value
RESULTS
Symptom duration >5 d 10.7 (3.8–30.2) <.01
Weight loss or persistent cough >2 wk 27.5 (3.5–213.1) <.01
In total, 554 children were treated for TBM during the study pe-
Recent close tuberculosis contact or 15.0 (3.3–67.9) <.01
riod; 66 (11.9%) were classified as definite TBM, 408 (73.6%) as positive TST/IGRA
probable TBM, and 72 (13.0%) as possible TBM; 8 had insuffi- Focal neurological deficit 15.0 (3.3–67.9) <.01
cient information for reliable classification [12]. Table 2 pro- Cranial nerve palsy 5.6 (1.2–26.0) .03
vides an overview of TBM cases, including staging and Altered level of consciousness 1.6 (.7–3.8) .31
diagnostic criteria. Diagnostic criteria are also reflected for the CSF clear 50.6 (6.2–410.7) <.01
32 cases with definite bacterial meningitis identified. When ap- CSF cells 10–500/µL 3.0 (1.1–7.9) .26

Downloaded from https://academic.oup.com/cid/article-abstract/59/11/1574/1746435 by guest on 11 April 2019

plying the uniform research case definition for TBM to bacterial CSF lymphocyte predominance >50% 1.6 (.6–4.0) .37
meningitis cases, 14 scored as possible TBM and 18 as “no CSF raised protein >1 g/L 3.3 (1.4–7.9) <.01
CSF:serum glucose ratio <50% 2.2 (.9–5.2) .08
TBM.” None of the bacterial meningitis cases tested positive
Hydrocephalus (CT and/or MRI) 53.7 (13.4–215.7) <.01
for human immunodeficiency virus (HIV). Among TBM pa-
Basal meningovascular enhancement 82.7 (10.5–651.0) <.01
tients, 8 were HIV infected, 205 were HIV uninfected, and (CT and/or MRI)
the HIV status was unknown in 341 patients. HIV prevalence Tuberculoma (CT and/or MRI) 2.5 (.3–22.7) .40
was low during the earlier part of the 20-year study period Infarct (CT and/or MRI) 2.8 (.8–10.7) .12
when HIV testing was not routinely done. In the possible CXR suggestive of active tuberculosis 6.3 (1.7–22.8) <.01
TBM group, 25% of patients had stage 3 disease, compared CXR suggestive miliary tuberculosis 6.2 (.8–50.3) .09
with 42% and 41% in the probable and definite TBM groups, Extraneural Mycobacterium 12.5 (1.6–98.5) .02
tuberculosis confirmation
respectively.
When scoring children with definite TBM independent of Abbreviations: CI, confidence interval; CSF, cerebrospinal fluid; CT, computed
tomography; CXR, chest radiograph; IGRA, interferon-γ release assay; MRI,
their M. tuberculosis culture results, 57 of 66 (86%) scored as magnetic resonance imaging; OR, odds ratio; TST, tuberculin skin test.
probable TBM and 9 of 66 (14%) as possible TBM. Of the 9 cul-
ture-confirmed TBM cases that scored as “possible” TBM, 3 had
limited clinical symptoms (1 was TBM stage I and the other 2
were TBM stage IIa) and 6 had minimal signs on neuroimaging. DISCUSSION
Seven culture-confirmed TBM cases were within 2 points of the
cutoff for a “probable” TBM score; the remaining 2 patients had In this retrospective assessment, the proposed uniform research
low scores for both clinical and neuroimaging criteria. case definition [12] provided excellent diagnostic accuracy com-
Among the definite bacterial meningitis group, 18 of 32 pared with culture confirmation of TBM, if a “probable” TBM
(56%) scored as unlikely TBM and 14 of 32 (44%) as possible score was used. With a “possible” TBM score, not a single TBM
TBM. Comparing patients with definite TBM and definite bac- case would have been missed, but clinical utility was minimal
terial meningitis, a probable TBM score provided good diagnos- given the low specificity achieved. It is important to try and
tic accuracy (sensitivity 86% and specificity 100%). A possible strike an optimal balance between improved sensitivity and ad-
TBM score was more sensitive but very nonspecific (sensitivity equate specificity given the need to initiate antituberculous ther-
100% and specificity 56%). apy as early as possible in all TBM cases.
Table 3 summarizes findings for specific criteria on univari- TBM often presents with nonspecific symptoms in stage I
ate analysis. Comparing definite TBM to definite bacterial men- disease, which complicates early diagnosis. Morbidity and mor-
ingitis, the most informative clinical criteria were either weight tality increases exponentially with disease progression. There-
loss or persistent cough >2 weeks (OR, 27.5; 95% confidence fore, rapid diagnosis is needed to initiate treatment at the
interval [CI], 3.5–213.1), recent close tuberculosis contact or earliest opportunity. Unfortunately, it is difficult to achieve a
positive tuberculin skin test (OR, 15.0; 95% CI, 3.3–67.9), and bacteriologically confirmed diagnosis at this early stage.
focal neurological deficit (OR, 15.0; 95% CI, 3.3–67.9). Macro- Maintaining a high index of suspicion and using a combina-
scopically clear CSF was highly indicative (OR, 50.6; 95% CI, tion of clinical, laboratory, and neuroimaging criteria for early
6.2–410.7), as were hydrocephalus (OR, 53.7; 95% CI, 13.4– identification of suspected TBM remains essential for early
215.7) and basal meningovascular enhancement (OR, 82.7; treatment initiation. In 2 previous studies (1 comprising both
95% CI, 10.5–651.0) on CT scan. adults and children and the other comprising only adult

Value of Uniform Case Definition in TBM • CID 2014:59 (1 December) • 1577

patients) comparing TBM and bacterial meningitis, features All authors have submitted the ICMJE Form for Disclosure of Potential
Conflicts of Interest. Conflicts that the editors consider relevant to the con-
suggestive of TBM included young age, subacute onset, head-
tent of the manuscript have been disclosed.
ache, peripheral white blood cell count, clear CSF appearance,
total CSF white cell count, low CSF neutrophil proportion, and
elevated CSF protein [13, 16]. Good sensitivities and specificities References
were obtained both using a composite clinical reference stan-
1. World Health Organization. Global tuberculosis report. WHO report
dard and when applied to a microbiologically proven M. tuber-
2013. Geneva, Switzerland: WHO.
culosis reference standard [14]. 2. World Health Organization. Global tuberculosis control; epidemiology,
Our study was limited by absence of data regarding precontrast strategy and financing. WHO report 2011. Geneva, Switzerland: WHO.
basal hyperdensity, suggestive CT, magnetic resonance imaging, 3. van Well GT, Paes BF, Terwee CB, et al. Twenty years of pediatric tu-
berculous meningitis: a retrospective cohort study in the Western Cape
or ultrasound imaging of tuberculosis outside the central nervous of South Africa. Pediatrics 2009; 123:1–8.
system and positive PCR evidence of extraneural tuberculosis. 4. Schoeman JF, Wait J, Burger M, et al. Long-term follow-up of childhood
Precontrast basal hyperdensity was described only in 2004 and tuberculous meningitis. Dev Med Child Neurol 2002; 44:522–6.

Downloaded from https://academic.oup.com/cid/article-abstract/59/11/1574/1746435 by guest on 11 April 2019

5. Garg RK. Tuberculosis of the central nervous system. Postgrad Med J
was therefore not available [19]. Univariate analysis showed 1999; 75:133–40.
that the majority of criteria were associated with a diagnosis of 6. Thwaites G, Chau TTH, Mai NTH, et al. Tuberculous meningitis. J
TBM, which warrants inclusion in the proposed research algo- Neurol Neurosurg Psychiatry 2000; 68:289–99.
7. Jönsson B, Ridell M. The Cobas Amplicor MTB Test for detection of
rithm [12]. The uniform research TBM case definition needs to
Mycobacterium tuberculosis complex from respiratory and non-respira-
be tested robustly in future prospective studies including bacterial tory clinical specimens. Scand J Infect Dis 2003; 35:372–7.
meningitis, viral meningitis, and normal CSF control groups, and 8. Hosoglu S, Geyik MF, Balik I, et al. Predictors of outcome in patients
in both adult and pediatric populations. It should be noted that with tuberculous meningitis. Int J Tuberc Lung Dis 2002; 6:64–70.
9. Solomons RS, van Elsland SL, Visser DH, et al. Commercial nucleic acid
although some of the TBM data were collected prospectively, all amplification tests in tuberculous meningitis—a meta-analysis. Diagn
bacterial meningitis data were collected retrospectively. Microbiol Infect Dis 2014; 78:398–403.
It is important to appreciate that the proposed research case 10. Thwaites GE, Tran TH. Tuberculous meningitis: many questions, too
few answers. Lancet Neurol 2005; 4:160–70.
definition criteria were not designed for use in clinical practice. 11. Wolbers M, Heemskerk D, Chau TT, et al. Sample size requirements for
The normal clinical examination findings in stage I TBM are a separating out the effects of combination treatments: randomised con-
limitation of the diagnostic criteria. It would therefore be less trolled trials of combination therapy vs. standard treatment compared
to factorial designs for patients with tuberculous meningitis. Trials
likely to score as probable TBM at an earlier stage of disease.
2011; 12:26.
The uniform case definition also holds promise for research pur- 12. Marais S, Thwaites GE, Schoeman JF, et al. Tuberculous meningitis: a
poses in resource-constrained settings with limited availability of uniform case definition for use in clinical research. Lancet Infect Dis
more sophisticated diagnostic methods until better rapid diag- 2010; 10:803–12.
13. Youssef FG, Afifi SA, Azab AM, et al. Differentiation of tuber-
nostic testing becomes available. Larger prospective studies are culous meningitis from acute bacterial meningitis using simple clini-
required where the uniform case definition is robustly tested. cal and laboratory parameters. Diagn Microbiol Infect Dis 2006;
55:275–8.
14. Moghtaderi A, Alavi-Naini R, Izadi S, et al. Diagnostic risk factors to
CONCLUSIONS differentiate tuberculous and acute bacterial meningitis. Scand J Infect
Dis 2009; 41:188–94.
The proposed uniform TBM research case definition performed 15. Sunbul M, Atilla A, Esen S, et al. Thwaites’ diagnostic scoring and
the prediction of tuberculous meningitis. Med Princ Pract 2005;
well when using a “probable” TBM score to differentiate childhood
14:151–4.
TBM and bacterial meningitis. Rigorous data collection is essential 16. Török ME, Nghia HD, Chau TT, et al. Validation of a diagnostic algo-
to obtain the necessary information in all criterion categories and rithm for adult tuberculous meningitis. Am J Trop Med Hyg 2007;
should assist comparability of future prospective studies. However, 77:555–9.
17. Medical Research Council Streptomycin in Tuberculous Trials Commit-
caution is needed when applying the uniform research case defini- tee. Streptomycin treatment of tuberculous meningitis. Lancet 1948;
tion to clinical diagnosis of children with early TBM. 1:582–96.
18. van Toorn R, Springer P, Laubscher JA, et al. Value of different staging
Notes systems for predicting neurological outcome in childhood tuberculous
meningitis. Int J Tuberc Lung Dis 2012; 16:628–32.
Financial support. This work was supported by a Vrije University– 19. Andronikou S, Smith B, Hatherhill M, et al. Definitive neuroradiological
National Research Foundation Desmond Tutu PhD Scholarship. diagnostic features of tuberculous meningitis in children. Pediatr Radiol
Potential conflicts of interest. All authors: No reported conflicts. 2004; 34:876–85.

1578 • CID 2014:59 (1 December) • Solomons et al