J Neurosurg Pediatrics 4:176–183,

4:000–000, 2009

Outcome of ventriculoperitoneal shunt placement in Grade

IV tubercular meningitis with hydrocephalus: a retrospective
analysis in 95 patients
Clinical article

Umesh Srikantha, M.B.B.S., M.Ch., Jagadeesh V. Morab, M.B.B.S., Savitr Sastry, M.S.,
Rojin Abraham, M.Ch., Anandh Balasubramaniam, M.Ch., Sampath Somanna, M.Ch.,
Indira Devi, M.Ch., Chandramouli A. Bangalore, M.Ch., and Paritosh Pandey, M.Ch.
Department of Neurosurgery, National Institute of Mental Health and Neuro Sciences (NIMHANS),
Bangalore, Karnataka, India

Object. Hydrocephalus is the most common complication of tubercular meningitis (TBM). Relieving hydro-
cephalus by ventriculoperitoneal (VP) shunt placement has been considered beneficial in patients in Palur Grade II or
III. The role of VP shunt placement in those of Grade IV is controversial and the general tendency is to avoid its use.
Some authors have suggested that patients in Grade IV should receive a shunt only if their condition improves with a
trial placement of an external ventricular drain (EVD). In the present study, the authors assessed the outcome of VP
shunt placement in patients in Grade IV TBM with hydrocephalus to examine the factors predicting outcome and to
determine whether a trial with an EVD is absolutely necessary prior to shunt placement.
Methods. Ninety-five consecutive cases of TBM with hydrocephalus in which the patients underwent VP shunt
placement were retrospectively analyzed, and direct VP shunts were placed whenever possible. An EVD was placed
first only in the presence of deranged blood parameters. Outcomes were assessed both in the short and long term.
Results. The mean patient age was 17.5 years (range 1–55 years). Fifty-two patients underwent direct VP shunt
placement, and the remaining 43 received EVDs first. Overall, 33 and 45% of patients had favorable short- and long-
term outcomes, respectively. Age older than 3 years and duration of altered sensorium ≤ 3 days were predictive of a
favorable short-term outcome. Glasgow Coma Scale score at presentation was predictive of long-term outcome. Of
the patients who did not improve with placement of an EVD prior to VP shunt insertion, 24 and 18% had favorable
short- and long-term outcomes, respectively; this was not significantly different from the outcome in the patients who
underwent direct VP shunt placement.
Conclusions. Direct VP shunt placement is an effective option in patients with Grade IV TBM with hydrocepha-
lus. Age and duration of altered sensorium are predictive of short-term outcome, while Glasgow Coma Scale score at
presentation predicts long-term outcome. Ventriculoperitoneal shunts should be considered even in patients who do
not improve with an EVD. (DOI: 10.3171/2009.3.PEDS08308)

Key Words • tubercular meningitis • hydrocephalus •

ventriculoperitoneal shunt • external ventricular drainage • outcome

T
uberculosis is one of the most common chronic
infections encountered in developing countries. It
is also reported in developed countries because of
AIDS, organ transplantations, and the increasing use of
immunosuppressant medications.6 Central nervous system
tuberculosis accounts for only 5% of all tubercular infec-
tions,8 yet has received a lot of attention because of the
Abbreviations used in this paper: ATT = antitubercular therapy;
EVD = external ventricular drain; GCS = Glasgow Coma Scale;
GOS = Glasgow Outcome Scale; TBM = tubercular meningitis; VP
= ventriculoperitoneal.
increased incidence of complications, and high morbidity
and mortality rate compared to other forms of tuberculo-
sis.6
Hydrocephalus is the most common complication of
TBM. Tandon et al.14 stated that hydrocephalus is nearly
always present in patients who have had TBM for 4–6
weeks, and is more common and develops earlier in chil-
dren than in adults.4,8 Severity of TBM has been graded
by different authors. The grading proposed by Palur and
colleagues,12 used in this study, recognizes 4 grades (Ta-
ble 1).
It is generally agreed that relieving hydrocephalus via

176 J Neurosurg: Pediatrics / Volume 4 / August 2009

Shunt placement in poor-grade TBM with hydrocephalus
TABLE 1: Palur grading system for TBM*
GCS
Grade Clinical Features Score
I no neurological deficit, normal sensorium 15
II neurological deficit present, normal sensorium 15
III altered sensorium, easily arousable; neurological 9–14
deficit present or absent
IV deeply comatose; decerebrate or decorticate postur- 3–8
ing present or absent
* Adapted from Palur et al., 1991.
VP shunt placement is beneficial in patients in Grades II and
III TBM.1,12 The general tendency is to avoid VP shunts in
Grade IV TBM with hydrocephalus because these patients
are considered to have a poor outcome, with reported mor-
tality rates of 40–100%.1,7,10,12 Other authors have suggest-
ed that patients with Grade IV TBM with hydrocephalus
should receive a VP shunt only if their condition improves
to Grade II or III with EVD placement.1,10 Although the
outcome of Grade IV TBM with hydrocephalus is consid-
ered dismal, relieving the hydrocephalus is the only other
treatment we can offer these patients apart from ATT, ste-
roids, and supportive care. Thus, at our institution, we tend
to undertake CSF diversion in most cases.
Although patients in Grade IV TBM have on most oc-
casions received ATT and steroids alone—without a VP
shunt—it is worthwhile to mention that to the best of our
knowledge, there has not been a single study solely dedi-
cated toward establishing the outcome or determining the
prognostic factors in this set of patients. We therefore un-
dertook this study in Grade IV TBM with hydrocephalus
with 3 main objectives: 1) to examine the outcome of VP
shunts in Grade IV TBM with hydrocephalus; 2) to study
the factors predicting outcome; and 3) to assess whether a
trial of EVD is necessary prior to VP shunt placement.
Methods
This study was conducted at the National Institute of
Mental Health and Neurosciences, Bangalore, India, and
is a retrospective analysis of all cases of Grade IV TBM
with hydrocephalus for which VP shunts were placed be-
tween January 2000 and December 2005.
Study Protocol
All patients with a diagnosis of TBM were started on
weight-appropriate doses of 4-drug ATT (isoniazid, rifam-
picin, ethambutol, and pyrazinamide) and steroids (intrave-
nous dexamethasone). In patients in Grade IV (GCS score
≤ 8) with hydrocephalus, the first choice for treatment was
VP shunt implantation. However, in the presence of hema-
tological or biochemical abnormalities such as anemia or
hyponatremia requiring correction prior to general anes-
thesia (hemoglobin < 8 g or Na+ < 130 mEq/L), an EVD
was instituted until the anesthetist considered the patient fit
for shunt implantation surgery. The diagnosis of TBM was
confirmed in all cases by cytochemical analysis of CSF ob-
tained by lumbar puncture.2
J Neurosurg: Pediatrics / Volume 4 / August 2009
Case Selection
Case records obtained in 95 consecutive patients who
met the inclusion criteria were retrospectively analyzed
and their demographic and clinical features, hematological
parameters, type of procedure done (direct VP shunt place-
ment or EVD first) were noted. The available follow-up and
the clinical status at follow-up were also recorded as noted
in each patient’s file, as well as results of telephone inter-
views or mailed questionnaires whenever possible.
Outcome Assessment
Outcome in patients with TBM with hydrocephalus
has been assessed in other studies in terms of GOS score
at follow-up.10,12 We also used GOS score at follow-up as
the criteria for assessing long-term outcome in our patients.
We felt that to truly evaluate the role of the VP shunt, we
would need to consider the short-term outcome, because
long-term outcome could be attributed to a combination of
the shunt, ATT, and steroid therapy. The short-term benefit
was attributed solely to the CSF diversion procedure. Be-
cause of a lack of criteria in the literature for judging what
is considered immediate clinical improvement in these pa-
tients who present with poor GCS scores, we considered
a cut-off of GCS score ≥ 12 at the time of discharge as a
reasonable criterion for clinical improvement. According-
ly, outcome was assessed in 2 forms: short-term and long-
term outcome. Short-term outcome was assessed using the
GCS score at discharge, with scores ≥ 12 considered a fa-
vorable outcome, and scores < 12 considered unfavorable.
Long-term outcome was assessed using the GOS score at
follow-up, with scores of 4 or 5 considered favorable, and
scores of 1–3 considered unfavorable. In children younger
than 2 years of age, the modified GOS score was used.
Statistical Analysis
Statistical analysis was performed with commercially
available software (SPSS 10.0; SPSS, Inc.). Cross-tabs with
chi-square tests were used for assessing outcome. Binary
logistic regression analysis was used to test significance in
a multivariate model. Chi-square tests were used to show
the similarity of the follow-up group to the whole popula-
tion.
Age, duration of symptoms, duration of altered sen-
sorium, GCS score at presentation, presence or absence of
infarcts, and basal exudates were considered as variables
against which to assess outcome.
Results
Ninety-five patients underwent VP shunt placement
at our institution during the study period. The mean pa-
tient age was 17.5 years (range 1–55 years), with 57 pa-
tients (60%) younger than 18 and 33 (35%) younger than
3 years of age. There were 65 male and 30 female patients
(the male-to-female ratio was 1.7:1). The median duration
of symptoms prior to presentation was 21 days and the
median duration of altered sensorium prior to intervention
(either VP shunt or EVD placement) was 3 days. The GCS
score at presentation was 3–6 in 18 patients (19%), and was
either 7 or 8 in the remaining 77 patients (81%). Infarcts
177
 U. Srikantha et al.

and basal exudates were present on preoperative CT scans TABLE 2: Results of the chi-square tests done to compare the
in 31 (32.6%) and 68 (71.6%) patients, respectively. Anemia groups with and without known outcomes*
and hyponatremia requiring correction was observed in 17
(18%) and 54 cases (57%), respectively. No. of Patients (%)
FU No FU
Duration of Follow-Up Variable (40 patients) (55 patients) p Value
A documented outcome was known in 40 cases (42%), age (yrs)
including 11 deaths (9 in-hospital deaths and 2 at follow- group 1 0.26
up). In these 40 cases, the median duration of follow-up
was 12 months (mean 18 months, range 3–65 months).    ≤18 26 (46) 31 (54)
Chi-square tests were performed to compare these 40    >18 14 (37) 24 (63)
cases with the remaining 55 cases, and there were no sta- group 2 0.43
tistically significant differences between the groups with    ≤3 13 (39) 20 (61)
respect to the variables analyzed (Table 2). This finding    >3 27 (43) 35 (57)
implies that the follow-up group is not disparate from the duration of symptoms (days) 0.45
whole group.
≤21 21 (44) 27 (56)
Outcome of Grade IV TBM With Hydrocephalus After VP >21 19 (40) 28 (60)
Shunt Placement duration of altered sensorium
0.56
Outcome was assessed in terms of short-term out- (days)
come (GCS score at discharge in all 95 cases) and long-    ≤3 22 (42) 30 (58)
term outcome (GOS score at follow-up in 40 patients in    >3 18 (42) 25 (58)
whom follow-up was available). GCS score at presentation 0.31
The GCS score at discharge was ≥ 12 in 33% of pa- 3–6 9 (50) 9 (50)
tients, of whom 6% improved to a GCS score of 14 or 15.
The remaining 67% of patients had GCS scores ≤ 11 at 7 or 8 31 (40) 46 (60)
discharge (Fig. 1 upper). basal exudates 0.19
Of the 40 patients with documented follow-up, 18 present 31 (46) 37 (54)
(45%) had a GOS score of 4 or 5 at follow-up, which was absent 9 (33) 18 (67)
considered a favorable outcome (Fig. 1 lower). All 18 of infarcts 0.24
these patients were independent in their activities, and present 11 (35) 20 (65)
most children were attending school and playing. The to-
tal number of deaths in this study was 11 (28%). absent 29 (45) 35 (55)
procedure 0.43
Factors Influencing Outcome direct shunt 21 (40) 31 (60)
We analyzed the factors influencing short- and long- EVD followed by shunt 19 (44) 24 (56)
term outcomes separately. GCS score at discharge 0.24
Short-Term Outcomes. The patient population was ≥12 11 (36) 20 (64)
divided into 2 groups on the basis of age. Taking 18 years <12 29 (45) 35 (55)
as the cut off, we observed that a larger percentage of
* There were no statistically significant differences with respect to any
patients younger than 18 years had GCS scores ≥ 12 at
discharge; this finding showed only a trend toward statis- of the variables. Abbreviation: FU = follow-up.
tical significance (p = 0.08; Table 3). On further analysis,
we found that very young children (younger than 3 years) Long-Term Outcome. In contrast to short-term out-
had a statistically significant, worse short-term outcome come, GCS score at presentation was the sole predictor of
than older patients. Hence, there was a clear indication long-term outcome, assessed in terms of follow-up GOS
that younger patients have a worse outcome compared to score (Table 4). A significantly larger percentage of pa-
their older counterparts. This finding held true even on tients with a GCS scores of 7 or 8 at presentation had
multivariate analysis (binary logistic regression model) a favorable follow-up GOS score (4 or 5). More impor-
where only age (p = 0.03, OR 3.04) emerged as an inde- tantly, GCS score at presentation (p = 0.02, OR 26.47) re-
pendent predictor of outcome (see Table 5). tained its independent prognostic value in the multivari-
Furthermore, patients with a short duration of altered ate model (Table 5). On the other hand, age, duration of
sensorium (≤ 3 days) had a relatively better outcome than symptoms or altered sensorium, and presence or absence
those with a longer duration. This trended toward statisti- of infarcts or basal exudates did not correlate with long-
cal significance (p = 0.059). Presence or absence of in- term outcome.
farcts or basal exudates, duration of symptoms, and, sur-
Direct Shunt Placement Versus EVD First
prisingly, GCS score at presentation did not correlate with
short-term outcome. Imaging studies obtained in patients Of the 95 patients, 52 underwent direct VP shunt
with or without infarcts who had varying outcomes are placement and the remaining 43 received EVDs before
presented in Fig. 2 and discussed in the figure legend. VP shunt placement, for reasons discussed above. Corre-

178 J Neurosurg: Pediatrics / Volume 4 / August 2009

Shunt placement in poor-grade TBM with hydrocephalus
TABLE 3: Summary of the influence of various factors on
short-term outcome in all 95 patients

No. of Patients (%)

Discharge Discharge
No. of GCS Score GCS Score p
Variable Patients ≤11 ≥12 Value
age (yrs)
group 1 0.08
   ≤18 57 42 (74) 15 (26)
   >18 38 22 (58) 16 (42)
group 2 0.02
   ≤3 33 27 (82) 6 (18)
   >3 62 37 (60) 25 (40)
duration of symptoms 0.47
(days)
   ≤21 48 33 (69) 15 (31)
   >21 47 31 (66) 16 (34)
duration of altered sen- 0.059
sorium (days)
   ≤3 52 31 (60) 21 (40)
   >3 43 33 (77) 10 (23)
GCS score at presentation 0.22
3–6 18 14 (78) 4 (22)
Fig. 1. Upper: Bar graph showing short-term outcomes in terms of 7 or 8 77 50 (65) 27 (35)
GCS scores at discharge (GCS @ d/s) in all 95 patients. Lower: Bar
graph showing long-term outcomes in terms of follow-up GOS scores basal exudates 0.20
(GOS @ f/u) in the 40 patients in whom follow-up data were available. present 68 48 (71) 20 (29)
absent 27 16 (59) 11 (41)
lations for outcome were drawn with discharge GCS and
follow-up GOS scores. infarcts 0.11
Of the 52 patients who underwent direct VP shunt place- present 31 24 (77) 7 (23)
ment, 17 (33%) improved to a GCS score ≥ 12 at discharge. absent 64 40 (62) 24 (38)
Among the 43 who received EVDs first, 18 improved to
a GCS score > 8 prior to VP shunt placement. Neverthe-
less, the remaining 25 patients who remained in Grade IV TBM. Although VP shunts have generally been consid-
after EVD insertion also underwent VP shunt placement. ered beneficial in Grade II and III TBM with hydrocepha-
Of these, 12 (48%) had improved to Grade III at discharge. lus,1,10,12 their role in patients in Grade IV is controversial,
Most notably, 6 patients (24%) improved to discharge GCS with several studies recommending VP shunts in these
scores ≥ 12, a percentage not significantly different from cases only if the patient shows improvement after EVD
that in patients who improved to the same level after direct placement.1,10,12 This recommendation is because the un-
VP shunt placement (Fig. 3 upper). This observation stress- derlying cause of altered sensorium is multifactorial and
es the importance of VP shunt insertion even in patients cannot be attributed solely to the presence of hydrocepha-
who do not show improvement with a trial EVD. lus.10 Nevertheless, no study has compared the outcome
Among the 40 patients in whom follow-up was avail- after VP shunt placement in a large homogenous sample
able, 19 underwent EVD placement prior to receiving a VP of patients in Grade IV TBM with hydrocephalus.
shunt, and 21 received a direct VP shunt. Of the 21 patients
who underwent direct VP shunt placement, 13 (62%) had Outcome of VP Shunt Placement in Grade IV TBM With
improved to a GOS score of 4 or 5 at follow-up. Of the 19 Hydrocephalus
patients who received an EVD first, 8 improved to Grade In 1971, Bhagawati3 reported on his experience in a
II or III prior to VP shunt placement, and of these, 3 (38%) limited group of 5 patients with TBM and hydrocepha-
had GOS scores of 4 or 5 at follow-up. Of the 11 patients lus with altered sensorium who received ventriculoatrial
who did not improve after EVD placement, 2 (18%) had shunts. All 5 patients’ conditions improved, and their hemi-
GOS scores of 4 or 5 at follow-up (Fig. 3 lower). plegia, aphasia, and blindness almost completely resolved.
However, recent studies have not been able to duplicate
Discussion these results. In their series of 7 patients, Palur et al.12 re-
ported a mortality rate of 100%. Mathew and colleagues10
Ventriculoperitoneal shunt placement has been the also reported a high mortality rate in their 12 patients with
standard method to relieve hydrocephalus in patients with Grade IV TBM with hydrocephalus. The only patient who

J Neurosurg: Pediatrics / Volume 4 / August 2009 179

 U. Srikantha et al.
TABLE 4: Summary of the influence of various factors on long-
term outcome*

No. of Patients (%)

FU GOS FU GOS
No. of Score Score p
Variable Patients 1–3 4 or 5 Value
age (yrs)
group 1 0.29
   ≤18 26 13 (50) 13 (50)
   >18 14 9 (64) 5 (36)
group 2 0.59
   ≤3 13 7 (54) 6 (46)
   >3 27 15 (56) 12 (44)
duration of symptoms (days) 0.51
≤21 21 12 (57) 9 (43)
>21 19 10 (53) 9 (47)
duration of altered sensorium
0.60
(days)
   ≤3 22 12 (55) 10 (45)
   >3 18 10 (56) 8 (44)
GCS score at presentation 0.02
3–6 9 8 (89) 1 (11)
7 or 8 31 14 (45) 17 (55)
Fig. 2. Preoperative CT scans obtained in 4 different patients who
received VP shunts. All patients had hydrocephalus with periventricu- basal exudates 0.63
lar lucency. A: Scan in a patient with a favorable long-term outcome present 31 17 (55) 14 (45)
(follow-up GOS score of 5). B: Scan showing infarct in the right basal absent 9 5 (56) 4 (44)
ganglia. Despite this, the patient had a favorable outcome. C and D:
Scans obtained in patients who did not improve after VP shunt place- infarcts 0.37
ment. The patient shown in panel C had a poor follow-up GOS score, present 11 7 (64) 4 (36)
and the patient in panel D died of the disease. absent 29 15 (52) 14 (48)
survived in their series showed clinical deterioration in the * Follow-up data were available in 40 patients.
ward from Grade I to IV, and was successfully returned
to normal functional status because of early intervention.
Agrawal et al.1 reported a mortality rate of 40% among ed follow-up. The overall mortality rate among patients
their 6 patients, and all survivors were left with major with available follow-up was 28% (11 of 40), which could
disabilities. In the 28-case series of Karande et al.,7 40% be on the higher side as the follow-up rate was poor.
of patients died and 46% survived with severe disabili-
ties; only 14% made a complete recovery. The authors of Factors Predicting Outcome
another study concluded that although the results of VP Several factors have been identified as prognostic vari-
shunt placement are far from satisfactory, early shunt de- ables in TBM. In their multivariate analysis of prognostic
ployment still remains the best option to prevent long-term variables, Misra et al.,11 identified age, grade of TBM, pres-
neurological sequelae.13 Lamprecht and colleagues9 noted ence of cranial nerve palsies, and severity of hydrocephalus
good outcomes in 8.3% of their 36 patients with Medical as prognostic variables. Kalita et al.,6 in their multivariate
Research Council Grade III TBM, and no good outcomes analysis, have recognized the presence of focal motor defi-
among those who presented with GCS scores of 3–5. The cits as the most important predictor of delayed sequelae.
major limitations of the above series is obviously the small The other factors were GCS score, shunt surgery, and
number of included cases. bacille Calmette-Guérin vaccination. Karande and asso-
In the present study, we evaluated the outcome in 95 ciates7 identified clinical variables predictive of complete
cases of Grade IV TBM with hydrocephalus with known recovery, survival with disability, and death separately. In
outcomes at follow-up in 40 cases. A favorable short-term their study, only presence of deep coma and the absence of
outcome was demonstrated in 33% of our patients. More extrapyramidal movements at presentation were associated
importantly, 18 (45%) of the 40 patients with follow-up with death. Misra et al.11 and Kalita et al.5 identified pres-
information available had a favorable long-term outcome. ence of focal deficits and infarctions as important predic-
This might be an erroneously high percentage of good tors of outcome at 3 and 6 months, respectively. However,
long-term outcomes, however, as it is possible that people all of these factors have been identified in groups that in-
with worse outcomes or who died would not have attend- cluded only a small portion of Grade IV patients. No study

180 J Neurosurg: Pediatrics / Volume 4 / August 2009

Shunt placement in poor-grade TBM with hydrocephalus
TABLE 5: Summary of the influence of various factors on short- and long-term outcome in a multivariate model*

No. of Short-Term Outcome Long-Term Outcome

Variable Patients p Value Exp(B) 95% CI for Exp(B) p Value Exp(B) 95% CI for Exp(B)
age (yrs) 0.03 3.040 1.097–8.430 0.224 0.307 0.046–2.056
≤3 13
>3 27
duration of symptoms (days) 0.871 1.085 0.407–2.891 0.417 1.998 0.375–10.643
≤21 21
>21 19
duration of altered sensorium (days) 0.126 2.141 0.807–5.682 0.194 3.498 0.529–23.13
≤3 22
>3 18
GCS score at presentation 0.438 1.675 0.454–6.181 0.021 26.47 1.633–429.3
3–6 9
7 or 8 31
basal exudates 0.829 0.887 0.297–2.645 0.489 0.499 0.069–3.582
present 31
absent 9
infarcts 0.229 1.931 0.660–5.646 0.388 2.155 0.376–12.34
present 11
absent 29

* Short-term outcome data were available in 95 patients, and long-term outcome in 40.

has so far identified factors predicting outcome in Grade
IV TBM with hydrocephalus.
In our series including only Grade IV TBM with hy-
drocephalus, age and duration of altered sensorium corre-
lated with short-term outcome. A younger age was associ-
ated with a worse short-term outcome, probably because
of the increased production of exudates and the formation
of adhesive leptomeningitis in them.10 The GCS score at
presentation, although not predictive of short-term out-
come, was the sole predictor of outcome in the long-term.
Patients who presented with GCS scores of 7 or 8 fared
better than those who presented with a lower score.
Role of Prior EVD
Palur et al.12 treated all of their Grade IV patients with
direct VP shunt placement and reported a 100% mortal-
ity rate. However, their sample size of 7 was very small.
These authors proposed that patients in Grade IV TBM
with hydrocephalus should undergo VP shunt placement
only if clinical improvement is shown after an EVD trial.
Following this protocol, Mathew et al.10 instituted EVDs
in 12 patients with Grade IV TBM with hydrocephalus,
and only 1 patient showed improvement. This patient’s
condition deteriorated in the ward from Grade I to IV,
and he received immediate medical attention. The patient
subsequently underwent VP shunt placement and was
neurologically normal at follow-up. The other patients
died of TBM. Agrawal and colleagues1 performed VP
shunt placements in 6 patients who showed improvement
after receiving EVDs, and reported a 40% mortality rate;
all surviving patients had severe disabilities.
An interesting point to note here is the observation
made by Mathew et al.10 in Grade III patients. These au-
thors placed EVDs in 13 patients with Grade III TBM
with hydrocephalus, and irrespective of improvement
after EVD placement, they instituted VP shunts in all
patients. Mathew and colleagues observed that of the 6
patients with Grade III TBM with hydrocephalus who did
not improve with an EVD, 4 were functionally normal at
follow-up after VP shunt placement. These authors there-
fore recommended that all patients in Grade III should
be given the benefit of VP shunt placement, irrespective
of their response to a trial of EVD; these results were not
extrapolated to patients in Grade IV, however.
In the present study, all patients in Grade IV received
VP shunts irrespective of their response to a trial of EVD.
Moreover, EVDs were instituted only in the presence of se-
vere biochemical or hematological derangements requiring
correction before the patient could receive general anesthe-
sia. We observed that among the patients who underwent
direct VP shunt placement, 32.6% had favorable short-term
outcomes, and 62% had favorable long-term outcomes.
This observation, in itself, emphasizes the effectiveness
of direct VP shunt implantation in patients in Grade IV
without a trial of EVD. Moreover, even with no response to
EVD, 24% of patients had a favorable short-term outcome,
and 18% had a favorable long-term outcome after shunt
placement. This is similar to the observation of Mathew
et al.10 in Grade III cases, which we have shown holds true
in Grade IV TBM with hydrocephalus as well. We believe
that a trial of EVD for a short duration (48–72 hours) is
inadequate to judge improvement. Cerebrospinal fluid di-
version is the only treatment we can offer these patients
apart from ATT, steroids, and supportive care. Hence, we

J Neurosurg: Pediatrics / Volume 4 / August 2009 181

 U. Srikantha et al.

Fig. 3. Upper: Flowchart showing short-term outcomes in all 95 patients who underwent either direct VP shunt placement
or received an EVD first. Lower: Flowchart showing long-term outcomes in the 40 patients in whom follow-up was available.
Patients are grouped according to whether they received a VP shunt directly or an EVD first. V. = ventricular.

believe that unless there are major contraindications, VP
shunts should be placed in these patients. However, caution
and clinical judgement should be exercised in patients with
GCS scores of 3–6 and extensive infarcts, as the likelihood
of improvement in these patients is very small.
There are obvious limitations to this study, including
its retrospective nature and the relatively poor follow-up in
the included cases. Although the follow-up group has been
shown not to be different from the entire group, whether
the outcome can be projected onto the entire group can
always be questioned. This is the only study to exclusively
examine Grade IV TBM with hydrocephalus with such a
large number of cases. In addition, we have unequivocally
shown the effectiveness of direct VP shunt placement in
terms of both short- and long-term outcome.
even in Grade IV TBM with hydrocephalus managed with
VP shunt placement, ATT, and steroids. Age and duration
of altered sensorium predict short-term outcome, while
GCS score at presentation predicts long-term outcome.
Direct VP shunt placement is an effective option in pa-
tients in Grade IV, and should be considered even in those
who do not show improvement after EVD placement. A
prior trial of EVD is not necessary in all cases, should be
used only as a temporary measure when necessary, and
should be followed up with a VP shunt in most cases.
Disclaimer
The authors report no conflict of interest concerning the mate-
rials or methods used in this study or the findings specified in this
paper.

Conclusions References
In summary, we demonstrated improved outcomes 1. Agrawal D, Gupta A, Mehta VS: Role of shunt surgery in pe-

182 J Neurosurg: Pediatrics / Volume 4 / August 2009

Shunt placement in poor-grade TBM with hydrocephalus
diatric tubercular meningitis with hydrocephalus. Indian Pe-
diatr 42:245–250, 2005
2. Ahuja GK, Mohan KK, Behari PM: Diagnostic criteria for tu-
berculous meningitis and their validation. Tuber Lung Dis
75:149–152, 1994
3. Bhagwati SN: Ventriculoatrial shunt in tuberculous meningi-
tis with hydrocephalus. J Neurosurg 35:309–313, 1971
4. Deshpande DH, Bharucha EP, Mondkar VP: Tuberculous
meningitis in adults: a clinico-pathological study of 18 cases.
Neurol India 17:28–34, 1969
5. Kalita J, Misra UK: Outcome of tuberculous meningitis at 6
and 12 months: a multivariate regression analysis. Int J Tu-
berc Lung Dis 3:261–265, 1999
6. Kalita J, Misra UK, Ranjan P: Predictors of long-term neu-
rological sequelae of tuberculous meningitis: a multivariate
analysis. Eur J Neurol 14:33–37, 2007
7. Karande S, Gupta V, Kulkarni M, Joshi A: Prognostic clinical
variables in childhood tuberculous meningitis: an experience
from Mumbai, India. Neurol India 53:191–195, 2005
8. Kemalogu S, Ozkan U, Bukte Y, Ceviz A, Ozates M: Timing
of shunt surgery in childhood tuberculous meningitis with
hydrocephalus. Pediatr Neurosurg 37:194–198, 2002
9. Lamprecht D, Schoeman J, Donald P, Hartzenberg H: Ventric-
uloperitoneal shunting in childhood tuberculous meningitis.
Br J Neurosurg 15:119–125, 2001
J Neurosurg: Pediatrics / Volume 4 / August 2009
10. Mathew JM, Rajshekhar V, Chandy MJ: Shunt surgery in poor
grade patients with tuberculous meningitis and hydrocepha-
lus: effects of response to external ventricular drainage and
other variables on long term outcome. J Neurol Neurosurg
Psychiatry 65:115–118, 1998
11. Misra UK, Kalita J, Srivastava M, Mandal SK: Prognosis of
tubercular meningitis: a multi-variate analysis. J Neurol Sci
137:57–61, 1996
12. Palur R, Rajshekhar V, Chandy MJ, Joseph T, Abraham J:
Shunt surgery for hydrocephalus in tuberculous meningitis: a
long-term follow-up study. J Neurosurg 74:64–69, 1991
13. Sil K, Chatterji S: Shunting in tuberculous meningitis: a neuro-
surgeons’s nightmare. Childs Nerv Syst 24:1029–1032, 2008
14. Tandon PN, Bhatia R, Bhargava S: Tuberculous meningitis,
in Vinken PJ, Bruyn GW, Klawans HL (eds): Handbook of
Clinical Neurology: Microbial Disease. Amsterdam: El-
sevier, 1988, Vol 52, pp 195–226
Manuscript submitted October 1, 2008.
Accepted March 31, 2009.
Address correspondence to: Paritosh Pandey, M.Ch., Associate
Professor, Department of Neurosurgery, National Institute of Mental
Health and Neurosciences, Bangalore–29, India. email: paritosh2000
@gmail.com.
183