Diagnostic Criteria for Tuberculous Meningitis

Rachna Seth 1 and Usha Sharma 2

1Department of Pediatrics, All India Institute of Medical Sciences, New Delhi

2Department of Pediatrics, SMS Medical College, Jaipur, Rajasthan, India

Abstract. Objective: Tuberculous Meningitis is associated with a high morbidity and mortality if there is a delay in diagnosis.
The diagnosis is based on clinical evaluation since the bacteriological diagnosis takes time and has a low yield. This study
attempts to validate these criteria in children with TBM. Methods: Forty-two children clinically suspected to have TBM were
enrolled in the study. History, examination, CT scan and CSF findings were utilized to categorize patients into "definite', "highly
probable", "probable" and "possible" TBM based on the criteria laid down by Ahuja et al. The validity of these criteria was tested
against bacterial isolation and response to treatment. Results : Thirty one children, with complete data, were included for
analysis. Using "improvement on therapy' as a criterion for definite TBM, we analyzed the sensitivity and specificity of the Ahuja
cdtefia in diagnosing TBM. Using the criteria of "highly probable" TBM, the sensitivity was 65% with a specificity of 75%. When
the criteria of "probable" TBM were used, the sensitivity increased to 96% while the specificity dropped to 38%. In an attempt
to make these criteria more appropriate for children, we modified the criteria by including mantoux reaction, and family history
of exposure in the criteria. The modified criteria gave a sensitivity of 83% and a specificity of 63%. Discussion : A sensitivity
of 65% (highly probable group) implies that 35% of TBM patients will be missed, while the probable criteria gave a 63% false
positive rate suggesting that the trade-off for a higher sensitivity makes the criteria very unreliable. Our modification of the
criteria gave us a reasonable sensitivity of 83% with a higher specificity of 63%. The false positive rate was also reduced to
38%. Thus the modified Ahuja criteria worked better for children with TBM. Conclusion : The modified Ahuja criteda are better
applicable for use in pediatric patients with TBM. Since the number of patients was small in this study, the study needs to
be validated with a larger sample size. [Indian J Pedlatr 2002; 69 (4) : 299-303]

Key words : TBM; Diagnostic criteria

Tuberculous meningitis (TBM) is a major health hazard
and is associated with a high morbidity and mortality.
The onset of the disease is generally insidious with a
vague symptomatology that makes the diagnosis of the
disease extremely difficult. The diagnosis of TBM rests on
the isolation of Mycobacterium tuberculosis. The bacilli are
seldom seen by staining and culture, which takes several
weeks and is associated with a very low yield. There is
thus a need to identify this potentially treatable disease in
the very early stages in order to reduce the associated
morbidity and mortality. Alternatives to bacterial
isolation Would therefore be of great help. Ahuja et al
applied a set of diagnostic criteria to 76 adult patients
suspected of having TBM and categorized them into
"definite", "highly probable", "probable", and "possible
TBM" based on the presence or absence of these criteria. 1
The validity of these criteria was tested using information
from bacterial isolation, polymerase chain reaction (PCR)
for tuberculosis, response to treatment and autopsy
wherever possible. PCR was positive in over 75% of
patients in the highly probable and probable groups.
Ninety one percent of patients with highly probable and
65% with probable TBM improved on antituberculosis
therapy. The authors found that these criteria were useful
Reprint requests : Dr. RachnaSeth, F-61,AIIMS,Ansari Nagar, New
Delhi-110 029. E-mail : drsandeepseth@hotmail.com
in the diagnosis of TBM. The present study has been
designed to evaluate Ahuja's criteria to diagnose TBM in
children combining the information from clinical,
laboratory and imaging data.
MATERIALS AND METHODS
Forty-two children clinically suspected to have TBM
admitted in the Pediatric Medicine w a r d of the Sir
P a d a m p a t Mother and Child Health Institute, SMS
Medical College, Jaipur, were enrolled in thestudy. In all
the patients a detailed clinical history was taken and
examination done with particular emphasis on
neurological complaints. Family history of tuberculosis
and immunization history was obtained from all patients
with particular reference to BCG and measles vaccination.
CSF examination was done in all the subjects after a
detailed fundus evaluation. The CSF obtained was
subjected to direct smear examination for acid fast bacilli
(AFB), mycobacterial culture, cell count, biochemistry,
serodiagnostic tests, fungal cells and malignant cells. A
fraction was left overnight at the bedside to see for
cobweb formation. Mantoux test was done irrespective of
the BCG status by giving 0.1 ml PPD containing I TU
intradermally on the ventral aspect of the forearm and
was read on the third day after giving the injection. The

Indian Joumal of Pediatrics, Volume 69--April, 2002 299

 R. Seth and U. Sharma
r e a d i n g was b a s e d on the p r e s e n c e or a b s e n c e of
induration b y the pen or palpation m e t h o d ( Positive
M a n t o u x test was c o n s i d e r e d as g r e a t e r t h a n ten
m i l l i m e t e r i n d u r a t i o n after 48 hours). Radiological
evaluation included chest radiograph, X-ray skull and
c o m p u t e r i z e d t o m o g r a p h i c (CT) scan. L a b o r a t o r y
evaluation included complete blood counts, erythrocyte
sedimentation rate (ESR) (westergren method), detailed
peripheral blood smear examination, serum biochemistry
and immunodiagnostic tests, wherever feasible.
T r e a t m e n t a n d Follow-up
All p a t i e n t s w e r e initially s t a r t e d on a f o u r d r u g
antituberculosis treatment [isoniazid (H) 5 mg per kg
b o d y w e i g h t / d a y , rifampicin (R) 10 mg per kg b o d y
weight/day, streptomycin (S) 20 mg per kg body weight
/ d a y , pyrazinamide (Z) 25-30 mg per kg body weight /
day] and corticosteroids (dexamethasone/prednisolone)
in a dose of I m g per kg body weight). These were given
for a p e r i o d of two m o n t h s a n d s u b s e q u e n t l y the
t r e a t m e n t was m a i n t a i n e d on a two d r u g (isoniazid,
rifampicin) regime for a period of ten more months. The
p a t i e n t s w e r e d i s c h a r g e d once their d i a g n o s i s was
confirmed, specific therapy was started and they showed
signs of improvement. They were followed up in the
Pediatric Medicine Outpatient Department of the J.K.Lon
Hospital, Jaipur.
Diagnosis of TBM
The above information was utilized to categorize patients
into " d e f i n i t e " , " h i g h l y p r o b a b l e " , " p r o b a b l e " and
"possible" TBM based on the criteria laid down by Ahuja
et al (Table 1) 1. The validity of these criteria was tested
against bacterial isolation and response to treatment. The
sensitivity, specificity, positive predictive value, negative
predictive value, percent of false positives and false
negatives were calculated and the statistical analysis was
done using the Microsoft Excel Analysis Pack, Microstat
and SPSS statistical software packages. C o n t i n u o u s
variables were analyzed using the Student's t-test and
proportions were analyzed by the Chi-square test.
RESULTS
Of the forty-two patients who were initially enrolled in
the study, 6 were untracable to follow up and 5 had
incomplete data (no CT scan done). Thus 31 children
treated as TBM were included for analysis. The mean
duration of follow-up was 5 + 4 months (range 3 to 12
months). The mean age of the patients was 4 + 3 years
(range 2 to 10 years). BCG vaccination had been given to
only one patient. All patients presented with symptoms
for more than two weeks and the presenting complaints
included fever (100%), headache (80%) and neurological
complications (including seizures, focal neurological
deficits and coma) in 50%.
Fifty five percent of the patients were diagnosed as
300
TABLE 1. Diagnostic Criteria for TBM (Ahuja et alO
A. Clinical
Fever and headache lasting for more than 14 days (mandatory).
Vomiting, alteration of sensorium or focal deficit (optional).
B. Cerebrospinal fluid
Pleocytosis with more than 20 cells, predominantly lymphocytes
(greater
than 60%),protein greater than 100mg%,sugars less than 60% of the
corresponding blood sugars.
Negative India ink studies and cytology for malignant cells (in
relevant situations)
C. Radiological
CT studies of head showing 2 or more of the following:
1. Exudates in the basal cisterns or in the sylvian fissures
2. Hydrocephalus
3. Infarcts
4. Gyral enhancement
D. Extraneural tuberculosis
Active tuberculosis of the lungs, gastrointestinal tract, urogenital
tract, lymph nodes, skeletal system, or skin as evidenced by
appropriate radiological,microbiologicaltests or by the presence of
caseation necrosis on histopathological examination.
Based on the above criteria, patients are categorized into four
categories: -
1. Definitive TBM
(I) Clinical criteria (A)
(ii) Bacteriologicalisolation from the CSF or diagnosis at autopsy
2. Highly probable TBM
(i) Clinical criteria (A)
(ii) All 3 of (B),and (C)
3. Probable TBM
(i) Clinical criteria (A)
(ii) Any 2 of B, C and D
4. Possible TBM
(i) Clinical criteria (A)
(ii) Any one of (B) (C) and (D)
" h i g h l y p r o b a b l e " TBM and 88% of these p a t i e n t s
improved on treatment. Thirty two percent of the patients
were diagnosed as "probable" TBM and 70% of these
improved on therapy. Thirteen percent of the patients
were diagnosed to have "possible" TBM (Table 2). Three
p a t i e n t s died. The baseline h e m o g r a m a n d s e r u m
chemistry was not different among the 3 groups. IgM was
positive for TB in 6 out of 10 patients and cerebro-spinal
fluid (csf) ELISA for TB was positive in 12 out of 15
patients. Anti-tuberculosis therapy resulted in elevation in
liver enzymes of at least 3 times in 26% patients. Extra-
neural tuberculosis was present in 26 patients, 3 having
lymph node involvement and the rest (n=23) having lung
involvement. CSF was abnormal in 30 of the 31 patients,
majority had a lymphocytic response (83%) with a mean
total cell c o u n t of 210 + 43 c e l l s / m m 3. No AFB was
isolated on direct smear or after culture. The CT scan was
abnormal in 17 patients of the 31 patients, with basal
exudates in all, hydrocephalus in 65%, infarcts in 65% and
Indian Journal of Pediatrics, Volume 69--April, 2002
 Diagnostic Criteria for Tuberculous Meningitis

gyral enhancement in 23%. The individual abnormalities M o d i f i e d Ahuja Criteria for t h e D i a g n o s i s o f T B M in

on CT scan did not show any correlation with diagnosis or Children.
response to therapy, so the analysis was done for the
combined CT scan findings. Mantoux reaction was done MANDATORY
in all the patients and was positive in 6 patients with
"highly probable" TBM (Table 3). 1. Fever for two weeks.
Family history of contact was positive in 5 patients 2. Abnormal CSF finding (Pleocytosis with more than 20
with highly probable TBM (Table 3). Using "improvement cells, predominantly lymphocytes (greater than 60%),
on therapy" as a criterion for definite TBM, we analyzed protein greater than 100mg%, sugars less than 60% of
the sensitivity and specificity of the Ahuja criteria in the corresponding blood sugars.)
diagnosing TBM (Table 4). Using the criteria of "highly
PLUS ANY TWO OF THE FOLLOWING
probable" TBM, the sensitivity was 65% with a specificity
of 75%. When the criteria of "probable" TBM were used, 1. Evidence of extra neural tuberculosis.
the sensitivity increased to 96% while the specificity 2. Positive (family) h i s t o r y of e x p o s u r e to a case of
dropped to 38% (Table 4). Since the Ahuja criteria were tuberculosis.
developed for adults, Mantoux reaction was not included 3. Positive Mantoux reaction (1TU)(> 10mm induration)
as a criterion. In an attempt to make these criteria more 4. Abnormal CT scan findings (as in original criteria)
a p p r o p r i a t e for children, we modified the criteria b y CT studies of head showing 2 or more of the following:
i n c l u d i n g M a n t o u x r e a c t i o n , a n d f a m i l y h i s t o r y of 1. Exudates in the basal cisterns or in the sylvian
exposure in the criteria. fissures
2. Hydrocephalus
3. Infarcts
TABLE2. Response to Treatment in the Different Categories 4. Gyral enhancement
(Ahuja criteria)*
Category of Tuberculous meningitis When these modified criteria were used, 22 patients
w e r e d i a g n o s e d as TBM, as c o m p a r e d to 17 a n d 27
Response Possible Probable Highly probable
"IBM TBM TBM patients for the "highly probable" and "probable" Ahuja
criteria. The modified criteria gave a sensitivity of 83%
Improve 1 (25%0) 7 (70%0) 15 (88%) and a specificity of 63%. (Table 5)
Static 1 (25%) 1 (10%) 0
Worsen 2 (50%) 2 (12%) 2 (12%) DISCUSSION

The clinical diagnosis of TBM w o u l d be easy if all the

*All variables shown as number (%) associated symptoms, signs, characteristic CSF and CT
TAeLE3. Distribution of Positive Mantoux Reaction and Positive Table 5. Evaluation of Modified Ahuja Criteria
Family History in Various Sub-Categories
Parameter Percent
Ahuja Criteria
Sensitivity 83%
Highly probable Probable Possible Specificity 63%
Positive Mantoux 6 2 0 Positive Predictive Value 86%
Negative Predictive Value 56%0
Positive Family History 5 3 0 False Negative % 17%
Mantoux test was done in all patients(n=31) and family history False Positive % 38%
was also available for all patients(n=31).
scan abnormalities are present. The problem arises when
TABLE4. Evaluation of Ahuja Criteria in the Diagnosis of TBM in only some of the classical features are present and one
Children Considering Response to Antituberculosis needs to consider other causes of chronic meningitis like
Treatment as the "Gold Standard" f u n g a l m e n i n g i t i s , m a l i g n a n c y or p a r t i a l l y t r e a t e d
Parameter Highlyprobable TBM Probable TBM b a c t e r i a l m e n i n g i t i s . The p r o b l e m w i t h m a k i n g a
bacteriological diagnosis is all the more difficult because
Sensitivity % 65 96 tubercle bacilli are rarely identified on Ziehl-Neelsen
Specificity % 75 38 staining and culture takes several weeks and is frequently
Positive Predictive Value % 88 81 negative. All this ultimately leads to a delay in starting
Negative Predictive Value % 43 75 treatment for this potentially treatable disease. TBM is the
False Negative % 35 43 c o m m o n e s t type of meningitis in children b e t w e e n 9
False Positive % 25 63 months to 5 years of age in the developing countries and

Indian Joumal of Pediatrics, Volume 69--April, 2002 301

 R. Seth and U. Sharma
as many as 60% of cases of meningitis are tuberculous in
origin. 2In patients with TBM, the prognosis is poor if the
confirmation of the diagnosis is delayed despite the best
available therapy and supportive measures. Therefore
logically, the emphasis should be on early diagnosis of
TBM before it progresses to a more severe form. Pyogenic
m e n i n g i t i s c o n s t i t u t e s a l m o s t 30% of all cases of
m e n i n g i t i s and so is the s t r o n g e s t c o n t e n d e r in the
differential diagnosis of chronic meningitis. A commonly
t r i e d s o l u t i o n is to start w i t h b o t h antibiotics a n d
antituberculosis therapy when in doubt. The slow onset of
r e s p o n s e to a n t i t u b e r c u l o s i s d r u g s , d r u g i n d u c e d
hepatotoxicity in 3-10% of patients and the high incidence
(6-19%) of multidrug resistant tuberculosis argues against
any such simplistic approach. 3,4
Currently molecular diagnostics like the PCR and
i m m u n o d i a g n o s t i c tests such as the e n z y m e linked
immunosorbent assay (ELISA) are being studied but they
are expensive, need trained personnel and not freely
available limiting their utility, s The need thus arises is to
combine the easily available tests into a simple scoring
system that can be applied in all the hospitals. However,
such scoring s y s t e m s n e e d to be v a l i d a t e d for
generalization of results.
The problems of diagnosis and treatment of childhood
tuberculosis have been extensively reviewed and the
focus has been categorically on one single issue that
childhood tuberculosis is a diagnostic dilemma due to the
absence of any gold standard (like sputum positivity for
AFB in adults). Ahuja et al have used PCR, AFB isolation,
autopsy and response to therapy as a gold standard for
diagnosis of TBM. In the present study, the response to
therapy as the indicator of definite TBM has been utilized
and tested the Ahuja criteria in children with chronic
meningitis.
Using response to therapy as a gold standard and then
validating a set of criteria based on this, and finally using
these criteria for diagnosis and predicting a response to
t h e r a p y is o b v i o u s l y circular r e a s o n i n g w i t h all its
limitations. There is also lack of reproducibility of the
d i a g n o s t i c criteria, m o s t of w h i c h are d e v e l o p e d ,
p u b l i s h e d and t h e n f o r g o t t e n . Inspite of all these
limitations, response to therapy is the only way to validate the
diagnostic criteria. Diagnostic criteria for c h i l d h o o d
tuberculosis have previously been used by Stegen et al 6
and Nair et al 7 and the criteria with maximum weightage
have been positive bacteriology and histopathology,
Mantoux test (induration > 10 mm), suggestive radiology,
contact w i t h a s p u t u m AFB p o s i t i v e in the family.
Therefore added a positive Mantoux reaction and history
of contact in our modification of the criteria were added.
The differential diagnosis of TBM includes all the
causes of chronic meningitis. Any patient who develops
meningitis is treated with antibiotics and when he does
not r e s p o n d after a d e q u a t e therapy, other causes of
chronic m e n i n g i t i s are s e a r c h e d for. Two w e e k s of
therapy are sufficient for resolution of bacterial meningitis
302
a n d a n y o t h e r self-limiting m e n i n g o e n c e p h a l i t i s .
Therefore, a criteria of two weeks of fever was kept as one
of the essential requirements for the diagnosis of TBM by
Ahuja et al 1, and retained in our modification of the
criteria.
CSF is usually abnormal in TBM, t h o u g h it can be
n o r m a l in certain situations like tuberculosis serous
meningitis or encephalitis in as m a n y as 20% patients.
Normal CSF can have upto five cells per cubic m m and
therefore the cut off was kept at four times the normal
level. A lymphocytic response was also kept in the criteria
b e c a u s e it w o u l d also e x c l u d e cases of r e c u r r e n t
meningitis that have a neutrophilic response. TBM can
h a v e a n e u t r o p h i l i c r e s p o n s e and p a r t i a l l y t r e a t e d
p y o g e n i c m e n i n g i t i s can also h a v e a l y m p h o c y t i c
response 4 but it was felt that a lymphocytic response
would improve the discriminatory value of the criteria.
A cut off of 100 rag% for the protein level was taken to
exclude patients with acute demyelinating disorders with
a meningeal response and a reduction in CSF sugar to
60% of blood sugar levels was taken to indicate bacterial
invasion of the subarachnoid space. We found the CSF
a b n o r m a l in 30 of the 31 patients, w i t h 26 of the 30
patients having a lymphocytic response. Ahuja et al 1 had
kept CSF abnormalities as one of the non-essential criteria.
In our modification, it was made mandatory because it
was felt that at primary and secondary health care centers,
one should not be thinking of TBM in patients who have
a normal CSF. This would reduce the sensitivity of the
criteria somewhat but would improve the specificity.
CT scan has been found to be very sensitive for the
diagnosis of TBM. In the present study, CT scan was
abnormal in 70% of patients with definite TBM. Bhargava
et als, in a study of 60 cases of TBM found only three cases
with a normal scan. Hydrocephalus was present in 87% of
children and 12% of adults, basal exudates were seen in
81% while 28% had infarcts. CT scan is n o w freely
available, and though somewhat expensive, has a very
important diagnostic and prognostic role in TBM.
Using the Ahuja criteria, we obtained a sensitivity of
65% and specificity of 75% for the "highly probable"
g r o u p . Using the criteria for " p r o b a b l e TBM, the
sensitivity increased to 96% but at the cost of a reduced
specificity of 38%. A sensitivity of 65% (highly probable
group) implies that 35% of TBM patients will be missed,
while the "probable" criteria gave a 63% false positive rate
suggesting that the trade-off for a higher sensitivity makes
the criteria very unreliable. We therefore modified the
criteria, keeping CSF as one of the mandatory criteria and
i n c l u d i n g M a n t o u x p o s i t i v i t y and a positive family
history in the minor criteria. This modification gave us a
reasonable sensitivity of 83% with a higher specificity of
63%. The false positive rate was also reduced to 38%. Thus
the modified Ahuja criteria w o r k e d better for children
with TBM.
This study has certain lacunae, the sample size is small
and the criteria developed have been validated against a
Indian Journal of Pediatrics, Volume 69--April, 2002
 Diagnostic Criteria for Tuberculous Meningitis

therapeutic response which as discussed is not always the
b e s t w a y to v a l i d a t e criteria. O n e n e e d s to assess the
criteria in a larger p o p u l a t i o n a n d p r o b a b l y test these
criteria against PCR.
REFERENCES
1. Ahuja GK, Mohan KK, Prasad K, Behari M. Diagnostic criteria
for Tuberculous meningitis and their validation. Tubercleand
Lung Dis 1994; 75 : 149-152.
2. Udani PM. Management of Tuberculous meningitis. Indian J
Pediatr 1985; 52 : 171-174.
3. Farer LS, Snider Jr DE. Tuberculosis : current recommen-
dations for cure and control. Postgrad Med 1988; 84 : 58-69.
4. Traub M, Colchester AC, Kingsley DP, Swash M. Tuberculo-
sis of the central Nervous System. Q ] Med 1984; 53 : 81-100.
5. Naidu AK, Gogate A. Early detection of tuberculous
meningitis using one step competitive ELISA. Indian ] Pathol
Microbiol (India) 1997; 40(4) : 531-538.
6. Stegen G, Kenneth J, Kaplan P. Criteria for guidance in the
diagnosis of tuberculosis. Pediatrics 1969; 43 : 260-263.
7. Nair PM, Philip E. A scoring system for the diagnosis of
tuberculosis in children. Indian Pediatr 1981; 18 : 299-303.
8. Bhargava S, Gupta AK, Tandon PN. Tuberculous meningitis.
ACT study. BR ] Radiol 1982; 55 : 189-196.

S T A T E M E N T A B O U T O W N E R S H I P A N D O T H E R PARTICULARS
A B O U T "THE INDIAN J O U R N A L OF PEDIATRICS"
(See Rule 8)

1. Hace of Publication : New Delhi

2. Periodicity of Publication : Monthly

3. Printer's Name Cambridge Printing Works

Nationality Indian
Address B-85, Naraina Industrial Area, Phase-H,
New Delhi-110028

4. Publisher's Name : Dr. Ishwar C. Verma

Nationality : Indian
Address : 125 (2nd Floor), Gautam Nagar
New Delhi-110049

. Editor's Name : Dr. Ishwar C. Verma (Editor-in-Cheif)

Nationality : Indian
Address : 125 (2nd Floor), Gautam Nagar
New Delhi-110049

6. Name & Address of Individuals : The Dr. K.C. Chaudhuri Foundation

who own the newspaper and 125 (2nd Floor) Gautam Nagar,
particulars of shareholders New Delhi-110049
holding more than one per cent
of the total capital

I, I.C. Verma, hereby declare that the particulars given above are true to the best of my
knowledge and belief.
Sd/-
(I.C. Verm,.)

Indian Journal of Pediatrics, Volume 69--April, 2002 303