35

Journal of Pediatric Neurology 12 (2014) 3543

DOI 10.3233/JPN-140636
IOS Press

Predictors of non-traumatic coma in a

pediatric cohort from a South Indian tertiary
care center: Results of a multivariate analysis

OP

Vykuntaraju K.N. Gowdaa,*, Ningappa B. Bannigidadb, Pragalath Kumarb,

Praveen-Kumar Srikanteswarac, Shivanandab, Govindrajb, Sarala H. Vykuntarajud and
Premalatha Ramaswamyb
aDepartment

of Pediatric Neurology, Indira Gandhi Institute of Child Health, Bangalore, India

of Pediatrics, Indira Gandhi Institute of Child Health, Bangalore, India
cDepartment of Neurology, Bangalore Medical College and Research Institute, Bangalore, India
dDepartment of Anatomy, Kempegowda Institute of Medical Sciences, Bangalore, India

Received 28 September 2013

Revised 31 October 2013
Accepted 19 November 2013

OR
C

bDepartment

AU

TH

Abstract. Objective: To study the etiology and clinical profile of non-traumatic coma in children at tertiary care center and to
determine the predictors of outcome. Methods: One hundred and four consecutive children between 2 mo12 yr were studied.
The clinical signs at admission; vital signs, Glasgow coma scale, respiratory pattern, papillary reflex, extra-ocular movements,
fundus picture and motor deficits were recorded. Etiology of coma was determined by clinical history, examination and relevant
investigations. Their progress was monitored clinically, biochemically and with multi-system monitors. Outcome was recorded
as survived or died. Results: Etiology of coma in 65% cases was intracranial infections; other causes were metabolic (20%).
Sixty-seven percent recovered completely, 16% had residual neurodeficits, 16% died. Survival was better in children with
intracranial infections (13%) as compared to metabolic coma (33%). On multivariate logistic regression, bradycardia, hypotension, abnormal respiratory pattern (especially, ataxic type), duration of coma more than 48 h, Glasgow coma scale < 7 at admission, unequal and non-reactive pupils, papilledema, abnormal extra-ocular movements, motor deficits, signs of meningitis
correlated with mortality. Requirement of ventilatory support and abnormal computerized tomography findings correlated with
mortality. Conclusions: Intracranial infections were the most common cause of non-traumatic coma in children; the most
common cause of death being metabolic coma. Simple clinical signs and relevant investigations served as prognostic indicators
of outcome.
Keywords: Coma, childhood, Glasgow coma scale, intracranial infections, non-traumatic coma

1. Introduction
Coma is a state of unresponsiveness without evidence of awareness of self or environment, a state
___________________________________________
*Correspondence: Dr. Vykuntaraju K.N. Gowda, Bangalore Child
Neurology and Rehabilitation Center, HANS complex, 8/A 1st Main
1st Cross, Manuvana, Near Adhichunchanagiri Choultry, Vijayanagar,
Bangalore, 560040, India. Tel.: +91 80 23301212, +91 9535212556;
Fax: +91 80 26541799; E-mail: [email protected].

from which the patient cannot be aroused by vocal or

sensory stimuli [1]. Coma is a life-threatening emergency and there are variations in its presentation.
Hence, a meticulous collection of clinical information
and investigations are of utmost importance especially,
in pediatric population. The attending pediatrician
must be methodical in his approach, documentation
and initiation of essential screening tests to determine
the etiology of coma. Many a times, the etiology of

1304-2580/14/$27.50 2014 IOS Press and the authors. All rights reserved

V.K.N. Gowda et al. / Predictors of non-traumatic coma

2. Materials and methods

vant investigations. Hematological parameters like

hemoglobin level, total and differential blood cell
counts, erythrocyte sedimentation rate, platelet count
and peripheral smear for malarial parasite. Biochemical tests like blood sugar, blood urea, serum creatinine, serum electrolytes, and urine analysis for ketone
bodies, albumin, sugar, culture and sensitivity. Specific investigation like liver function tests, renal
function tests, cerebrospinal fluid analysis, arterial
blood gas analysis, electroencephalogram, and imaging like X-ray chest, echocardiogram, and computerized tomography (CT) scan of head, magnetic resonance imaging of head whenever necessary. These
children were started on appropriate treatment as per
the standard protocols, and their progress was monitored clinically, biochemically and with multi-system
monitors and appropriate interventions, including
ventilator support was given, whenever required. The
outcome was observed and analyzed at the time of
discharge.
Definitions of study variables were as follows:
Coma: a state of unresponsiveness without evidence
of awareness of self or environment, a state from
which the patient cannot be aroused by vocal or sensory stimuli. Tachycardia: heart rate above the upper
limit for that age. Bradycardia: heart rate less than 60
per minute. Hypertension: blood pressure more than
95th percentile for the age and sex. Hypotension: blood
pressure below 5th percentile for the age and sex.

OR
C

coma cannot be defined, despite the available investigations. Monitoring of a comatose child with modified Glasgow coma scale (GCS) is one of the essential
steps in the management and predicting the outcome.
Neurological assessment has to be done regularly and
if facilities are available, monitoring of intracranial
pressure has to be done. Coma carries a high morbidity and mortality and many studies have indicated that
the longer the duration of coma, poorer is the outcome.
Early institution of appropriate therapy is essential for
better outcome [2]. The incidence of non-traumatic
coma is 30.8 per 100,000 children [3]. Only a few
studies have examined the relationship between clinical variables and outcome of comatose children [4].
The most common cause of non-traumatic coma is
neuroinfections and other causes are toxic-metabolic
conditions, status epilepticus, intracranial bleed and
others [5]. Aims and objectives of the present study
were to study the causes of non-traumatic coma in
children at a tertiary care pediatric center in South
India and to assess the prognostic determinants.

OP

36

AU

TH

This was a prospective observational study conducted at emergency unit of a pediatric tertiary care
teaching and referral hospital. Institutional ethics
committee had approved the study. Informed written
consent was obtained from the parents/guardian of the
subjects.
Inclusion and exclusion criteria; children with nontraumatic coma with GCS of < 12 between the age
group of 2 mo to 12 yr, who were admitted to pediatric
intensive care unit at a tertiary level referral hospital,
were included in the study group. Children with previously known neurological illness and who were
terminally ill and traumatic comas were excluded. The
study period was done between October 2011 and
September 2012.

2.2. Statistical methods

Chi-square and Fisher exact test were used to test
the significance of study parameters. Odds ratio was
used to find the strength of relationship between the
study parameters. Multivariate logistic regression was
used to find the predictors for outcome. P value of <
0.05 was considered significant; P value of 0.05 to
0.01 was taken as moderately significant; P value of
0.01 was taken as strongly significant.

2.1. Study design

3. Results
This prospective study included about 104 children
with non-traumatic coma between the age group of
2 mo to 12 yr. A detailed clinical history was obtained
and a thorough clinical examination including a detailed neurological examination was conducted in all.
The GCS and signs of increased intracranial pressure
were monitored. The patients were subjected to rele-

A total of 104 children with non-traumatic coma

were included in the study. The incidence of nontraumatic coma was 7.3% of pediatric intensive care
unit admission and 2% of general admissions to our
tertiary care teaching hospital. Forty-eight percent
children were in the age group of 15 yr. Sixty percent

V.K.N. Gowda et al. / Predictors of non-traumatic coma

29 (27.88)
50 (48.07)
25 (24.03)

4 (23.52)
12 (70.58)
1 (5.88)

3 (17.64)
7 (41.17)
7 (41.17)

26 (25.00)
5 (4.80)
73 (70.19)

7 (41.17)
0
10 (58.82)

5 (29.41)
4 (23.52)
8 (47.05)

74 (71.15)
7 (6.73)
23 (22.12)

13 (76.47)
2 (11.76)
2 (11.76)

11 (64.70)
1 (5.88)
5 (29.41)

53 (50.96)
29 (27.88)
20 (19.23)
2 (1.92)

11 (64.70)
4 (23.52)
2 (11.76)
0

6 (35.29)
6 (35.29)
3 (17.64)
2 (11.76)

46 (44.23)
27 (25.96)
31 (29.80)

2 (11.76)
6 (35.29)
9 (52.94)

Age
2 mo1 yr
15 yr
> 512 yr
Heart rate
Normal heart rate
Bradycardia
Tachycardia
Blood pressure
Normal
Hypertension
Hypotension
Respiratory pattern
Normal
Hyperventilation
Acidotic breathing
Ataxic respiration
Duration of coma
< 24 h
2448 h
> 48 h

Total numbers
Outcomes
(n = 104)
Neurodeficits Mortality
n (%)
(n = 17)
(n = 17)
n (%)
n (%)

OR
C

Parameters

100% mortality followed by central hyperventilation

which had 20.7% mortality. The neurodeficits and the
mortality were 4.93 times more likely in children
having coma lasting more than 48 h (P < 0.001). The
mortality was 3.4 times more likely in children who
presented with a GCS score < 7 (P = 0 .032). Normal
pupils with a normal response to light were seen in 34
children. Children who had unequal and non-reactive
pupils had 100% mortality. Papilledema was noted in
8.6%. Mortality was 44% in children with papilledema
compared to 14% with a normal fundus examination.
Abnormal extra-ocular movements were noted in two
children. Mortality was 50% in children with abnormal extra-ocular movements compared to 15.5% in
children with normal extra-ocular movements (P =
0.326). Abnormal limb posturing was noted in 8.6%
of children. Mortality was 22% in children with abnormal limb posturing in comparison to 16% with
normal limb posture (P = 0.638). Children with abnormal motor deficit were 5.1 times more likely to
have mortality (P = 0.002). Fifty percent of children
with quadriparesis died as compared to 11% with
hemiparesis. Children with abnormal deep tendon
reflexes (DTRs) were 1.74 times more likely to have a
bad outcome (neurodeficits and mortality, P = 0.275).
Mortality in those with sluggish DTRs was 20.8%
compared to exaggerated (13.6%) and normal DTRs
(6.7%). Tables 3 and 4 respectively depict the relevant
investigations and the etiological distribution in our
patients.
The predominant cause of non-traumatic coma in
the present study was intracranial infections (65%)
followed by metabolic causes (20%). The most
common cause of death was metabolic (33%), followed by neuroinfections (13%). Among the metabolic causes, the predominant cause for death was due
to hepatic encephalopathy five of seven cases (71%).
Among the infectious causes, tuberculous meningitis
(TBM) had high mortality when compared to pyogenic and viral meningoencephaliteis. In pyogenic
meningitis group, 9% died and 13% had neurodeficits.
In 13 cases with TBM 21% mortality and 57% neurodeficits were noted. Out of 19 cases of viral encephalitis, 5% mortality and 16% neurodeficits were
noted. Out of the three cases diagnosed as cerebral
malaria, two recovered completely and one died.
Systemic infections causing toxic encephalopathy
were noted in nine cases, out of which two died.
Metabolic causes contributed to 30 cases. Seven had
hepatic encephalopathy. Six had diabetic ketoacidosis

OP

Table 1
Clinical parameters

5 (29.41)
3 (17.64)
9 (52.94)

AU

TH

were males. The most common symptoms were fever

(83%), convulsions (75%), vomiting (52%), headache
(21%), loose stools (16%), rash (11%), and jaundice
(10%). Sixty-eight percent had coma of less than 24 h
duration and 20% had 2472 h and 12% had more
than 72 h at the time of admission. On examination,
dehydration (12%), bulged anterior fontanels, pallor,
skin rash 7.7% each, icterus in 3.8% and organophosphorus smell in one child were noted. Nutritional
status was normal in 36%; grade III and IV malnutrition were seen in 6% of children. Nutritional status
and outcome were not of statistical significance. Tables 1 and 2 depict the clinical parameters and the
clinical examination findings respectively.
Bradycardia was present in 4.8%. Children with
bradycardia were 26.5 times more likely to have
mortality (P = 0.002). Mortality was more in hypotensive group (22%) compared to hypertensive group
(14%). Abnormal respiration was seen in 49% of
children. Those children with abnormal respiratory
pattern had mortality 2.15 times more than those
children with normal respiration (P = 0.158). Among
the abnormal respiratory patterns, ataxic type had

37

38

V.K.N. Gowda et al. / Predictors of non-traumatic coma

Table 2
Findings on clinical examination and outcomes
Total numbers
(n = 104)
n (%)

Table 3
Investigation parameters and outcomes
Parameters

Total numbers
(n = 104) n (%)

8 (47.05)
6 (35.29)
3 (17.64)

34 (32.69)
64 (61.53)
1 (0.96)
2 (1.92)
3 (2.88)

11 (64.70)
6 (35.29)
0
0
0

6 (35.29)
7 (41.17)
1 (5.88)
2 (11.76)
1 (5.88)

3 (17.64)
7 (41.17)
7 (41.17)

83 (79.80)
9 (8.65)
12 (11.53)

8 (47.05)
6 (35.29)
3 (17.64)

9 (52.94)
2 (11.76)
6 (35.29)

36 (34.61)
44 (42.30)
24 (23.07)

3 (17.64)
12 (70.58)
2 (11.76)

6 (35.29)
6 (35.29)
5 (29.41)

Mortality
(n = 17) n (%)
3 (17.64)
13 (76.47)
1 (5.88)

TH

Glucose (mg/dL)
< 50
11 (10.57)
50180
92 (88.46)
> 180
1 (0.96)
Sodium (mEq/L)
< 120
1 (0.96)
120135
53 (50.96)
135145
45 (43.26)
> 145
5 (4.80)
Cranial computerized tomography
Not done
72 (69.23)
Normal
5 (4.80)
Abnormal
27 (25.96)
Cerebral edema
13 (12.50)
Hydrocephalus
6 (5.76)
Basal exudates
2 (1.92)
Meningeal enhancement
1 (0.96)
Focal lesions
2 (1.92)
Subdural collection
3 (2.88)
Ventilator support
Not required
81 (77.88)
Required
23 (22.11)
< 24 h
9 (8.65)
2448 h
4 (3.84)
> 48 h
10 (9.61)

26 (24.03)
39 (37.50)
39 (37.50)

OR
C

Glasgow coma scale score

36
79
1012
Pupil size
Normal
Dilated and reacting pupils
Dilated and non-reactive pupils
Unequal pupils
Constricted pupils
Motor deficits
No motor deficits
Hemiparesis
Quadriparesis
Deep tendon reflexes
Normal
Increased
Decreased

Outcomes
Neurodeficits
Mortality
(n = 17) n (%) (n = 17) n (%)

OP

Parameters

AU

0
9 (52.94)
7 (41.17)
1 (5.88)

10 (58.82)
0
7 (41.17)
3 (17.64)
2 (11.76)
1 (5.88)
0
1 (5.88)
0
6 (35.29)
11 (64.70)
4 (23.52)
1 (5.88)
6 (35.29)

(one died). One child had Reyes syndrome; one child

with organophosphorus compound poisoning expired;
one child with cobra bite recovered completely with
anti-snake venom treatment. All four cases due to
hypoxia recovered. Six cases had status epilepticus,
all of them recovered without any deficit. All five
cases with dyselectrolytemia recovered completely
with correction. Twenty-seven percent mortality was
noted in the patients who presented with blood sugars
< 50 mg/dL (P = 0.412). Mortality was 20% in normal
serum sodium group compared to 18% with a sodium
level < 135 mEq/L; however it was not statistically
significant (P = 0.14). Twenty sex percent children
expired with abnormal CT scan of head findings
compared to no mortality in normal CT scan findings
(P = 0.04). Ventilator support was received in 23
children. Fifty two percent of ventilated children recovered. Total out of 104 children, 67% children
recovered completely, 16% recovered with some
neurodeficits in the form of monoperesis, hemiplegia,
cranial nerve involvement and 16% expired.

4. Discussion
Pediatric coma has been for long an enigma with
only a few studies and with inconclusive information.

V.K.N. Gowda et al. / Predictors of non-traumatic coma

39

Table 4
Distribution of etiology
Total number
(n =104)
n (%)

23 (22.12)
19 (18.27)
14 (13.46)
3 (2.88)
9 (8.65)

Neurodeficits
(n = 17)
n (%)

Mortality
(n = 17)
n (%)

3 (17.64)
3 (17.64)
8 (47.17)
0
3 (17.64)

2 (11.76)
1 (5.88)
3 (17.64)
1 (5.88)
2 (11.76)

1 (0.96)
6 (5.77)
1 (0.96)
7 (6.73)
5 (4.81)
1 (0.96)

1 (5.88)
1 (5.88)
5 (29.41)

1 (0.96)
1 (0.96)

1 (5.88)

4 (3.88)
6 (5.77)
2 (1.98)
1 (0.96)

OR
C

OP

Neuroinfections
Pyogenic meningitis
Viral meningoencephalitis
Tuberculous meningitis
Cerebral malaria
Systemic infection with toxic encephalopathy
Metabolic
Hypoglycemia
Diabetic ketoacidosis
Reyes syndrome
Hepatic encephalopathy
Dyselectrolytemia
Uremic encephalopathy
Drugs and poisoning
Poisoning
Snake envenomation
Miscellaneous
Cerebral hypoxia
Seizure disorder
Hypertensive encephalopathy
Leukemia

Outcomes

Etiologies

AU

TH

In comatose patients, evidence of widespread damage

of brain stem or cerebral hemispheres at onset usually
predicts death or severe disability. Therefore, the
clinical signs that reflect the extent and severity of
dysfunction of cerebral hemispheres and/or brain stem
have been studied. GCS reflects the integrity of cerebral functions. Eye movements and pupillary responses, express the functions regulated by the brain
stem. The breathing pattern and the signs of abnormal
tone of limbs and posturing indicate the extent and
severity of cerebral hemisphere as well as brain stem
damage. In the present study, the clinical features,
etiology and the laboratory findings of non-traumatic
coma were studied. Our data like the earlier reports on
non-traumatic coma in children indicates that several
clinical variables individually exhibit a significant
dependence to outcome and may therefore be helpful
in predicting outcome. The striking similarity of our
study with Bansal et al. [5] study has been shown in
Table 5.
Earlier studies emphasized the role of multiple
variables in assessing prognosis and cautioned against
the use of isolated one or two variables. Some biosta-

tistical methods can be used to mimic this approach

and to determine the outcome of individual comatose
patients. Jennett et al. [6] used Bayesian analysis to
classify 200 head injury patients into one of two outcome groups, employing information obtained within
24 h of the onset of coma. Stablein et al. [7] used a
logistic regression model, clinical information being
obtained within 1 h of admission. Levy et al. [8] have
diagramed a sequential decision-making process,
based on multiple clinical responses in 500 patients
admitted with non-traumatic coma, to estimate prognosis from the time of admission to 7 d. In this study,
we have used multivariate logistic regression model
for predicting the prognosis in the form of poor recovery (deficits) and mortality (Table 6).
Out of the 104 children, the majority (48%) were
between the age group of 15 yr. Mortality was more
in male children compared to females. Forty nine
percent had abnormal respiration and these children
had mortality of 2.15 times more over those children
with normal respiration. In our study, the mortality
was more in hypotensive group. Five cases had
bradycardia and had high mortality (80%), which was

40

V.K.N. Gowda et al. / Predictors of non-traumatic coma

Table 5
Comparison with the study of Bansal et al. [5]

Present study
Total numbers
Mortality
(n = 104)
(n = 17)
n (%)
n (%)
74 (71.15)
11 (64.70)
7 (6.73)
1 (5.88)
23 (22.12)
5 (29.41)
26 (25.00)
8 (47.06)
78 (75.00)
9 (52.94)
95 (91.34)
13 (76.47)
9 (8.65)
4 (23.52)
102 (98.07)
16 (94.11)
2 (1.92)
1 (5.88)

16 (16.00)
18 (18.00)
19 (19.00)
2 (2.00)
5 (5.00)

4 (11.42)
5 (14.28)
5 (14.28)
0
2 (5.71)

23 (22.12)
19 (18.26)
14 (13.46)
3 (2.88)
9 (8.64)

OR
C

Normal blood pressure

Hypertension
Hypotension
Glasgow coma score < 7
Glasgow coma score > 7
Normal fundus
Abnormal fundus
Normal extraocular movements
Abnormal extraocular movements
Infections
Pyogenic meningitis
Viral meningoencephalitis
Tuberculous meningitis
Cerebral malaria
Others
Metabolic
Diabetic ketoacidosis
Reye syndrome
Hepatic encephalopathy
Others
Recovered
Neurodeficits

Bansal et al. [5]

Total numbers
Mortality
(n = 100)
(n = 35)
n (%)
n (%)
65 (65.00)
19 (54.28)
17 (17.00)
4 (11.42)
18 (18.00)
12 (34.28)
75 (75.00)
26 (74.28)
25 (25.00)
9 (25.71)
60 (60.00)
16 (45.71)
40 (40.00)
19 (54.28)
57 (57.00)
12 (34.28)
43 (43.00)
23 (65.71)

OP

Parameters

1 (1.00)
0
6 (6.00)
5 (14.28)
4 (4.00)
2 (5.71)
8 (8.00)
5 (14.28)
25 (25.00)
40 (40.00)

2 (11.76)
1 (5.88)
3 (17.64)
1 (5.88)
2 (11.76)

6 (5.76)
1 (5.88)
1 (1.44)
1 (5.88)
7 (6.73)
5 (29.41)
7 (6.73)
0
70 (67.30)
17 (16.34)

TH

Table 6
Multivariate logistic regression for prediction of bad prognosis-deficits and mortality
Logit co-efficient
0.462
0.010
0.372
0.159
01.693
0.437
0.061
00.033
0.608
01.194
0.505
00.286
00.976
02.648
0.482
01.318
1.709

AU

Investigations
Malnutrition
Abnormal respiration
Hypotension
Hypertension
Bradycardia
Tachycardia
Febrile
Glasgow coma < 7
Abnormal posture
Abnormal fundus
Abnormal pupil
Decreased deep tendon reflexes
Increased deep tendon reflexes
Signs of meningitis
Abnormal motor deficit
Duration of coma > 24 h
Constant

statistically significant in (P = 0.002) this was due to

raised intracranial pressure. All these observations

Wald
0.563
0.000
0.203
0.017
1.353
0.402
0.009
0.002
0.255
1.219
0.500
0.088
1.981
9.021
0.464
4.165
1.672

P value
0.453
0.986
0.653
0.897
0.245
0.526
0.926
0.962
0.614
0.270
0.479
0.767
0.159
0.003
0.496
0.041
0.196

Adjustment odds ratio

0.63
0.99
0.69
0.85
5.43
0.65
0.94
1.43
0.54
3.30
0.60
1.33
2.65
14.130
0.62
3.74
0.18

were in concordance with those reported by Bansal et

al. [5]. Our study showed that as the duration of coma

V.K.N. Gowda et al. / Predictors of non-traumatic coma

OP

assessments [8].
Some studies have deliberately excluded etiology
from the procedure, so as to determine whether clinical findings at the initial examination could provide
discriminative information about outcome, and because a definitive diagnosis might not always be
available at the onset of coma [8]. However, in our
study we have analyzed the etiologies and found that
there was a significant dependence between the etiology of coma and outcome. Classification errors may
be minimized and serious ones eliminated from prediction procedures by taking etiology into account [4,
9]. The predominant cause of non-traumatic coma of
the present study was intracranial infections accounting for 65%, similar etiological distribution seen
with few other studies in this regard [4,5,9,12,13]. The
most common cause of death was metabolic coma
33%.
Among the infectious causes, TBM had high mortality when compared to pyogenic and viral meningoencephalitis. In pyogenic meningitis, 9% mortality
and 13% had neurodeficits, were comparable to prior
reports [14,15]. In TBM, 21% mortality and 57%
neurodeficits were comparable to Shaha and Gandhi [16]. In viral encephalitis, 5% mortality and 16%
neurodeficits, was noted, which shows a better outcome in our study when compared to other studies.
Others have reported much higher mortality rates of
37.5% [9], 29% [17], 32% [18], 36% [19] and 22.8% [20]
respectively, but these rates were all in cases of Japanese encephalitis. Out of three cases of cerebral malaria, two recovered completely and one child died.
Thapa et al. [21] reported 22% mortality in cerebral
malaria. Systemic infections causing toxic encephalopathy were nine cases, out of which two (22.2%)
died.
Among the metabolic causes the predominant cause
for death was due to hepatic encephalopathy 71%,
which was comparable to other studies [5,13]. Out of
six cases of diabetic ketoacidosis cases one child died.
Kecskes [22] reported 50% of mortality. Dyselectrolytemia contributed to 4.81%; all recovered without
any deficit. Similar results are observed in few studies [4,9,23]. One child with Reyes syndrome died due
to late presentation with increased intracranial pressure. Benakappa et al. [24] reported mortality of 78%
and Kalra et al. [25] observed 90% mortality in children with Reyes syndrome. One child with organophosphorus compound poisoning died. Johnston and
Seshia [9] study observed good recovery [9]. One

AU

TH

OR
C

prolongs mortality also increases (less than 48 h with

22%, more than 48 h coma with 29%). Children who
were in coma for more than 48 h have 4.93 times more
chances of mortality as compared to less than 48 h of
coma. Similar observations were cited in earlier
studies [4,9]. In our study, mortality was 31% with
GCS score < 7 as compared to 11% with GCS score >
7. The mortality was 3.4 times more likely in children
presented with GCS score < 7 (P = 0.032). Similar
findings were observed in prior studies [2,4,5,10].
Levy et al. [8] have excluded the duration of coma,
since by its nature it is a retrospective variable. Johnston and Seshia [9] have opined that duration of coma
correlates significantly with outcome but is no longer
a useful indicator of outcome because it is frequently
impossible to determine the duration of coma in children being treated in intensive care units. Seventeen
percent mortality in children with normal pupil size as
compared to 15.7% in those with abnormal pupil size
was noted, which was not statistically significant.
Bansal et al. [5] have reported 39% and 8% mortality
respectively in those with abnormal pupil size and
with normal pupils. Fundus abnormality in relation to
mortality was statistically similar between the two
studies; mortality was significantly related to the abnormal fundus. Increased mortality was associated
with the abnormal extra-ocular movements (P = 0.326)
which was comparable in both studies. In a study by
Levy et al. [8], abnormal neuro-ophthalmological
signs, at first in combination, then singly after 3 d,
indicated a poor prognosis. On the first day, features
favoring moderate disability or good recovery included verbalization, together with any three reactive
of corneal reflexes, papillary reflexes, oculovestibular
responses or motor responses.
Those who had abnormal DTRs had 1.74 times
more deficit and death. This may be due to involvement of pyramidal tract and mid brain by metabolic
and infective causes. Twelve percent children had
motor deficits who contributed to 61% mortality,
when compared to those without deficits. Hence,
children with motor deficits had 5.1 times more mortality over children without neurodeficits, which was
statistically significant. Similar results were noticed
by prior studies [5,11]. These data underscore the
importance of undertaking a complete neurological
examination at the time of admission. This examination could be done within two or three minutes and is
vital, since high doses of drugs such as barbiturates or
neuromuscular blocking agents interfere with later

41

V.K.N. Gowda et al. / Predictors of non-traumatic coma

and the signs of meningitis. The limitations of present

study are we didnt assess cognition, vision and
hearing and we assessed only immediate outcome at
the time of discharge.
In conclusion, the most common cause of nontraumatic coma in our cohort is neuroinfection while
the most common cause of death was due to a metabolic cause. The bad prognostic indicators are longer
the duration of coma, deeper the coma, abnormal
respiratory pattern, bradycardia, hypotension, abnormal pupillary signs, motor deficits, requirement of
ventilatory support and abnormal CT findings.

References

[1] Bates D. The prognosis of medical coma. J Neurol Neurosurg

Psychiatry 2001; 71 Suppl 1: i20-3.
[2] Margolis LH, Shaywitz BA. The outcome of prolonged coma
in childhood. Pediatrics 198; 65(3): 477-83.
[3] Wong CP, Forsyth RJ, Kelly TP, Eyre JA. Incidence, aetiology, and outcome of non-traumatic coma: a population
based study. Arch Dis Child 2001; 84(3): 193-9.
[4] Seshia SS, Johnston B, Kasian G. Non-traumatic coma in
childhood: clinical variables in prediction of outcome. Dev
Med Child Neurol 1983; 25(4): 493-501.
[5] Bansal A, Singhi SC, Singhi PD, Khandelwal N, Ramesh S.
Non-traumatic coma. Indian J Pediatr 2005; 72(6): 467-73.
[6] Jennett B, Teasdale G, Braakman R, Minderhoud J, Heiden J,
Kurze T. Prognosis of patients with severe head injury.
Neurosurgery 1979; 4(4): 283-9.
[7] Stablein DM, Miller JD, Choi SC, Becker DP. Statistical
methods for determining prognosis in severe head injury.
Neurosurgery 1980; 6(3): 243-8.
[8] Levy DE, Bates D, Caronna JJ, Cartlidge NE, Knill-Jones RP,
Lapinski RH et al. Prognosis in non-traumatic coma. Ann
Intern Med 1981; 94(3): 293-301.
[9] Johnston B, Seshia SS. Prediction of outcome in non-traumatic coma in childhood. Acta Neurol Scand 1984; 69(6):
417-27.
[10] Chaturvedi P, Kishore M. Modified Glasgow coma scale to
predict mortality in febrile unconscious children. Indian J
Pediatr 2001; 68(4): 311-4.
[11] Seshia SS, Seshia MM, Sachdeva RK. Coma in childhood.
Dev Med Child Neurol 1977; 19(5): 614-28.
[12] Ogunmekan AO. Non-traumatic coma in childhood: etiology,
clinical findings, morbidity, prognosis and mortality. J Trop
Pediatr 1983; 29(4): 230-2.
[13] Nayana PP, Serane TV, Nalini P, Mahadevan S. Long-term
outcome in coma. Indian J Pediatr 2005; 72(4): 293-5.
[14] Bhat BV, Verma IC, Puri RK, Srinivasan S, Nalini P. Prognostic indicators in pyogenic meningitis. Indian Pediatr 1987;
24(11): 977-83.
[15] Paul SS, Singh D. Observations on partially treated purulent
meningitis. Indian Pediatr 1979; 16(3): 233-7.
[16] Shah AK, Gandhi VK. Prognosis of tubercular meningitis.
Indian Pediatr 1984; 21(10): 791-5.
[17] Kamala CS, Rao MV, George S, Prasanna NY. Japanese
encephalitis in children in Bellary Karnataka. Indian Pediatr
1989; 26(5): 445-52.
[18] Gowrie-Devi M, Deshpande DH. Japanese encephalitis in

AU

TH

OR
C

child with cobra bite recovered completely with anti-snake venom treatment. Kumar [26] observed 3%
mortality; Lahori et al. [27] reported 5% mortality in
snakebite cases.
All children with hypoxia due to cardio-respiratory
cause recovered. Three cases in our study had respiratory failure, which were managed promptly with
respiratory support. Johnston and Seshia [9] observed
only 14% good outcome with 24% morbidity and 62%
mortality. This was because the study included near
drowning cases, which has got high mortality. Six cases
in our study had status epilepticus; all recovered without any deficit. Johnston and Seshia [9] observed same
100% good recovery. But Aicardi and Chevrie [28]
reported 11% mortality and 57% morbidity and
Meyetal et al. [29] reported 1.9% mortality and 20.5%
morbidity in status epilepticus. Out of the total five
cases of dyselectrolytemia, all made good recovery
and same results were observed by Kansguard et
al. [23].
Johnston and Seshia [9] have observed that children
with some forms of encephalitis may do better than
their initial examination suggested; those with hypoxic-ischemic insult such as cardiorespiratory arrest
do much worse than suggested by findings within 12 h
of the onset of coma; outcome from hepatic encephalopathy is influenced by systemic complications
produced by liver dysfunction. Hence, consideration
of etiology may be a logical step to take after an attempt is made to predict outcome using clinical variables from the initial examination. We should be
cautious with this approach because therapeutic
measures, if instituted early, may improve the outcome in conditions such as coma following hypoxic-ischemic insult or meningitis.
Fifty-two percent ventilated children recovered. It
may be the contributing factor for overall good outcome and of 17% mortality as compared to Bansal et
al. [5] who reported 35% mortality. In our study 70
children recovered completely, 16% children recovered with some neurodeficits in the form of monoparesis, hemiplegia, cranial nerve involvement and
16% mortality were noted. The mortality in our study
was lower when compared to other studies [4,5,9,30]
which may be due to the use of mechanical ventilator
support and prompt therapy. Multivariate logistic
regression showed that the following findings were
the predictors of outcome of coma abnormal respiration, hypotension, bradycardia, GCS < 7, cranial
nerve involvement, abnormal fundus, abnormal DTRs

OP

42

V.K.N. Gowda et al. / Predictors of non-traumatic coma

[25] Kalra V, Sachdev HP, Menon PS. Reyes's syndrome in Delhi.

Indian J Pediatr 1981; 48(394): 585-8.
[26] Kumar V. Accidental poisoning in south west Maharashtra.
Indian Pediatr 1991; 28(7): 731-5.
[27] Lahori UC, Sharma DB, Gupta KB, Gupta AK. Snake bite
poisoning in children. Indian Pediatr 1981; 18(3): 193-7.
[28] Aicardi J, Chevrie JJ. Convulsive status epilepticus in infancy
and childhood. A study of 239 cases. Epilepsia 1970; 11(2):
187-97.
[29] Maytal J, Shinnar S, MoshSL, Alvarez LA. Low morbidity
and mortality of status epilepticus in children. Pediatrics
1989; 83(3): 323-31.
[30] Bondi FS. The prognosis of medical coma in Ibadan: results
of multivariate analysis. Ann Trop Pediatr 1992; 12(1):
87-94.

AU

TH

OR
C

OP

India. Indian J Pediatr 1980; 47(5): 413-7.

[19] Kumar R, Mathur A, Kumar A, Sharma S, Chakraborty S,
Chaturvedi UC. Clinical features & prognostic indicators of
Japanese encephalitis in children in Lucknow (India). Indian
J Med Res 1990; 91: 321-7.
[20] Benakappa DG, Anvekar GA, Viswanath D, George S.
Japanese encephalitis. Indian Pediatr 1984; 21(10): 811-5.
[21] Thapa BR, Marwaha RK, Walia BN, Kumar L, Mehta S.
Falciparum malaria. Indian Pediatr 1987; 24(3): 221-5.
[22] Kecskes SA. Diabetic ketoacidosis. Pediatr Clin North Am
1993; 40(2): 355-63.
[23] Kongsgaard UE, Holst-Larsen H, Lie B, Fauchald P, Nakstad
PH. Severe, symptomatic hyponatremia. Tidsskr Nor Laegeforen 1990; 110(28): 3617-19 (in Norwegian).
[24] Benakappa A. Reye's syndrome in Bangalore. Indian J Pediatr 1991; 58(6): 805-10.

43