Article

Tropical Doctor
0(0) 1–6
! The Author(s) 2020
Clinical spectrum, aetiology and Article reuse guidelines:
sagepub.com/journals-permissions
predictors of acute febrile encephalopathy DOI: 10.1177/0049475520967916
journals.sagepub.com/home/tdo
at a tertiary hospital in south India –
A prospective observational study

Manoj Job1 , Alice J Mathuram2 , Vijay P Turaka3,

Ronald AB Carey3 , Ramya Iyyadurai2 and Sudha J Rajan2

Abstract
Acute febrile encephalopathy is a common syndrome in the tropics with high mortality encountered by emergency phys-
icians. In order to study the aetiology of acute febrile encephalopathy and its mortality and functional outcome over one
year, data on all patients >18 years of age with short duration of fever (<14 days) and altered mental status were collected
and followed up until one month after discharge. Non-infectious aetiology, found in 29%, portends a poor outcome.

Keywords
Acute febrile encephalopathy, acute encephalitis syndrome, heat stroke

After obtaining written informed consent from a leg-

Introduction
ally acceptable representative, a standard proforma was
Acute febrile encephalopathy (AFE) is the clinical syn- used to collect the demography, signs, symptoms,
drome of short duration fever which is either accompa- co-morbidity, and premorbid mRS. Patients were fol-
nied or followed by altered mental status.1 This is a lowed up to discharge and further information such as
common syndrome encountered in the emergency laboratory findings, need for Intensive Care admission,
department.2 The challenge of identifying its aetiology duration of hospital stay and discharge outcome was
has profound implications in terms of management and collected. Follow-up after one month by telephone
outcome.2 interview ascertained vital and functional status by
Although AFE arises from various infectious causes, mRS. Primary aetiology was defined by pre-specified
which vary from region to region and season to season,3 criteria as depicted in Table 1.
in practice non-infectious aetiology such as heat stroke A sample size of 250 was calculated on the propor-
presents none too rarely in tropical countries. tion expected to have mRS > 3 at one month, implying
dependent functional status. From previous studies,1
this was assumed to be 6% with 3% precision for
Methods 95% confidence intervals.
Our prospective cohort was collected from a 2700-bed
university teaching hospital in south India from June
2015 to May 2016. All patients admitted as emergencies 1
Assistant Professor, Department of Medicine, Christian Medical
with AFE as defined by a history of fever for <14 days,
College, Vellore, India
with altered mental status, with or without meningeal 2
Professor, Department of Medicine, Christian Medical College, Vellore,
signs were included in the study. Patients with acute India
3
cerebrovascular disease and head injury were excluded, Associate Professor, Department of Medicine, Christian Medical
as the aim was to study the aetiological spectrum of College, Vellore, India
AFE. Primary outcomes used the modified Rankin
Corresponding author:
score (mRS) at the end of one month and mortality Sudha Jasmine Rajan, Department of Medicine Unit 3, Christian Medical
at the end of hospital stay (Table 1). Secondary out- College, Vellore 632004, Tamil Nadu, India.
comes included predictors for mortality and diagnoses. Email: sudhajasmine@cmcvellore.ac.in
2 Tropical Doctor 0(0)

Table 1. Diagnostic criteria with specific aetiology of acute febrile encephalopathy (AFE).

Clinical diagnosis Diagnostic criteria

Acute viral meningoenceph- AFE with CSF WBC > 5 cells/mm3 AND
alitis of specific aetiology3 Positive CSF viral serology or culture or multiplex PCR AND /OR
EEG/MRI evidence of parenchymal disease AND
Absence of bacterial pathogen on CSF smear or culture AND
Negative tuberculous meningitis criteria
Acute viral meningoenceph- AFE with CSF WBC > 5 cells/mm3 AND
alitis of unknown Negative study for CSF viral pathogen-CSF serology or CSF viral Culture or multiplex PCR
aetiology3 AND /OR EEG/MRI evidence of parenchymal disease AND
Absence of bacterial pathogen on CSF smear or culture AND negative for Tuberculous
meningitis criteria
Tuberculous meningitis4,5 Acute febrile illness with altered mental status or meningitis AND
Ziehl-Neelsen stain of CSF positive for acid fast bacilli OR
CSF culture positive for tuberculous bacilli OR
CSF Xpert-MTB RT-PCR analysis positive for TB OR
Neuroimaging (CT/MRI scan) consistent with TB meningitis (viz. hydrocephalus, basal men-
ingeal enhancement, etc.) OR
Any other source positive for tuberculosis (lymph node biopsy / sputum smear / tissue
biopsy / bone marrow biopsy)
Scrub meningoencephalitis6 Acute febrile illness with clinical features of meningitis or altered sensorium AND
Scrub typhus IgM ELISA positive with or without eschar AND/OR
Response to doxycyline therapy
Bacterial meningitis7 Acute febrile illness with clinical features of meningitis and EITHER:
Gram staining of CSF positive for meningitis causing bacteria OR
CSF culture positive for a known bacterial pathogen OR
Blood culture positive for a known meningitis causing bacteria, i.e. Streptococcus pneumoniae,
etc.
Febrile metabolic Acute febrile illness (differentiated /undifferentiated) with no acute CNS infection AND
encephalopathy2 ANY OF:
Hyponatraemia (sodium < 125 mmol/L) AND
Correction of sodium reversing the altered sensorium OR
Hypernatremia (sodium > 145 mmol/L) OR
Hypoglycaemia (plasma glucose < 2.2 mmol/L) OR Hyperglycaemia (DKA: blood glucose
>13.8 mmol/L, arterial pH < 7.3, bicarbonate <15 mmol/L, and moderate ketonuria or
ketonemia. HHS: blood glucose >33.3 mmol/L, arterial pH > 7.3, bicarbonate >15 mmol/
L, mild ketonuria or ketonemia, and effective serum osmolality >330 mOsm/mL) OR
Hypercalcaemia (serum calcium> 2.5 mmol/L) OR
Hyperthermia (body temperature >40.9 C)
Febrile toxic encephalopathy Acute febrile illness (differentiated /undifferentiated) with no acute CNS infection AND
History of recent consumption of toxins (alcohol or drugs) with altered mental status
CSF: cerebrospinal fluid; WBC: white blood count; PCR: polymerase chain reaction; CSF: cerebrospinal fluid; EEG: electroencephalogram; MRI:
magnetic resonance imaging; MTB RT PCR: mycobacterium tuberculosis real-time PCR; ELISA: enzyme-linked immunosorbent assay; DKA: diabetic
ketoacidosis, HHS: hyperosmolar hyperglycemic state.

Data were entered using Epi-data version 3, and The research proposal for our study was approved
data analysis was done using Statistical Package for by the Institutional Review Board [IRB Min. No. 9450
the Social Sciences (SPSS) software package (version dated 5 June 2015]. There were no ethical issues related
16). Continuous study variables were described using to this study.
mean with standard deviation and discrete variables
were summarised using frequencies with percentages.
The diagnostic predictors for aetiology and outcome
Results
predictors were identified using bivariate analysis and All 73,143 patients admitted to the emergency depart-
logistic regression analysis with odds ratio and 95% ment from June 2015 to May 2016 were screened. On
confidence interval (CI). A P-value <0.05 was con- examination, at presentation, 183/265 patients
sidered significant. (69.05%), who satisfied our inclusion criteria, had
Job et al. 3

Figure 1. (a) Classification of primary etiologies. (b) Consort figure.

documented temperature > 37.2 C. Others had fever Immunodeficiency Virus and 5(3%) tested positive.
during the course of their hospital stay. The mean tem- The mean hospital stay was 8.8 (7.4) days. Some 51
perature at admission was 38 C (SD 2.6). All patients patients (21%) required ICU admission.
were followed up at discharge but 18 were lost to The primary aetiology of all patients is shown in
follow-up at one month. The classification of primary Table 2.
aetiology and the consort figure are shown in Figure 1. The primary outcomes are shown in Table 3.
The sex ratio in the cohort was almost equal (111, Mortality at one month was 20% (n ¼ 51). Nearly a
58% male). Their age ranged from 15 to 86 years. The quarter (59) had a poor functional status (mRS > 3)
mean (SD) age was 46.9 (19.8) years. Major symp- at this time. Mortality at one month was significantly
toms noted were altered sensorium, headache, nuchal higher with non-infectious aetiology (p < 0.05), with
rigidity, vomiting and seizures. Mean Glasgow Coma poorer functional outcome at discharge (p < 0.01) and
Scores (GCS) at admission was 11.8 (3.2). Diabetes in after one month (p < 0.05) when compared to patients
80 (30.2%) and hypertension in 61 patients (23%) were with infectious aetiology.
the major comorbidities. A total of 155 patients Predictors of diagnosis and outcome using univari-
(58.6%) underwent testing for Human ate and multivariate analysis are shown in Table 4.
4 Tropical Doctor 0(0)

Table 2. Primary etiology of acute febrile encephalopathy. low- and middle-income countries where most health-
care is privatised. Our study provides data for ground
Acute febrile encephalopathy (n ¼ 265)
rules in such a setting. Non-infectious causes portend to
Infectious 70.5% (n ¼ 187) n (%) higher mortality, hence timely diagnosis is critical. Heat-
Bacterial 47 (17.7) related illnesses should be considered during the months
Pyogenic 12 (4.5) of April and May in the tropics with core temperature
Scrub typhus 33 (12.4) >40 C, and altered mentation among those > 40 years
Leptospirosis 1 (0.3)
with comorbid conditions and aggressive cooling meas-
ures with evaporative and convection methods should be
Typhoid 1 (0.3)
initiated. Screening for metabolic and toxic causes for
Tuberculosis 45 (17)
encephalopathy through meticulous drug history, alco-
Viral 77 (28.9) hol use along with basic biochemistry for electrolytes,
Specific aetiology 14 (5.2) ammonia and blood glucose facilitates diagnosis of
Herpes Simplex virus (HSV) 3 (1.1) these disorders. Most studies from the region have
Dengue 5 (1.8) excluded the non-infectious causes.1,8–13
Varicella Zoster virus (VZV) 2 (0.7) History of headache and presence of nuchal rigidity
Epstein Barr virus (EBV) 3 (1.1) point to an infectious aetiology over a non-infectious
Adeno virus 1 (0.3) cause. Tuberculosis (24%), scrub typhus (17.6%), viral
Unknown aetiology 63 (23.7) meningoencephalitis (7.5%), pyogenic meningitis
Fungal 2 (0.7) (6.4%) and sepsis associated (7.5%) were the specific
Parasitic 2 (0.7)
infectious aetiologies in our study. One-third of the
patients did not have an aetiological agent identified
Sepsis associated 14 (5.2)
despite extensive workup, due to the rampant use of
Total 187 (70.5)
over the counter antibiotics in the region, unidentified
Non-infectious 29.5% (n ¼ 78) viral aetiology, or aseptic meningitis.
Metabolic
Tuberculous meningitis (TBM) although infre-
Hyperthermia (heat stroke) 48 (18.1)
quently reported in other series1,8,10,13 was a predomin-
Hyponatraemia 7 (2.6)
ant cause of AFE in ours, reflecting the high prevalence
Hyperglycaemia 5 (1.8) of TB in India and better access to diagnostics such as
Hypercalcaemia 1 (0.3) TB polymerase chain reaction (GeneXpert MTB/RIF)
Hepatic encephalopathy 1 (0.3) and liquid culture method (MGIT) that makes early
Toxin diagnosis feasible. Evidence of tuberculosis from
Alcohol related 5 (1.8) lymph node or bone marrow aids in the diagnosis, as
Drugsa 5 (1.8) TBM is due to hematogenous spread of the disease and
Othersb 6 (2.2) as yield from CSF is often low. Scrub typhus is another
Total 78 (29.5) infectious aetiology that presents as AFE and the path-
Grand total 265 (100) ognomonic eschar should be meticulously looked for in
a
the intertriginous regions. The frequency of different
Neuroleptic malignant syndrome.
b aetiologies differed in various studies across north
Autoimmune encephalitis-4, undifferentiated febrile illness with acute
psychosis-1, metastatic brain disease-1. India. Modi et al.1 and Khan et al.8 found pyogenic
meningitis and cerebral malaria as the commonest aeti-
ology whereas Joshi et al.11 and Bhalla et al.2 found
Infections such as scrub typhus, EBV infection, and viral meningoencephalitis. This variation in different
non-infectious aetiology such as heat stroke were found aetiologies among different centres could be due to geo-
to have seasonal variation described in Figure 2. Heat graphical and seasonal variations.
stroke was seen in large numbers in the month of April. One-fifth died at the end of one month and 30%
with AFE had poor functional status at the end of
one month with severe disability requiring assistance
Discussion for activities of daily living. This underscores the high
This study is one of the few from south India on AFE mortality and morbidity associated with acute febrile
among adults. Complete evaluation of AFE requires encephalopathy, consistent with other studies.1,2,9 The
imaging studies (MRI/CT Brain), CSF analysis, cultures prognosis is poor if the patient has a low GCS, requires
(bacterial, mycobacterial, fungal), histopathology and intubation, has diabetes or a poor premorbid func-
multiplex viral PCR and a complete metabolic screen. tional status and non-infectious aetiology.
These facilities are neither available nor affordable in These factors help with prognostication and inform
Job et al. 5

Table 3. Primary outcome of patients with acute febrile encephalopathy.

Outcome n (%) Infectious (n ¼ 187) Non-infectious (n ¼ 78) Total (n ¼ 265)

mRS pre morbid illness, n (%) (n ¼ 187) (n ¼ 78) (n ¼ 265)

0–1 176 (94.1) 68 (87.2) 244 (92.1)
2 11 (5.8) 10 (12.8) 21 (8)
2–3 9 (4.8) 6 (7.7) 15 (5.7)
4–6 2 (1) 4 (5.1) 6 (2.3)
mRS at discharge, n (%) (n ¼ 187) (n ¼ 78) (n ¼ 265)
2 70 (37.6) 43 (55.1)) 113 (42.6)
0–1 117 (62.5) 35 (44.9) 152 (57.3)
2–3 21 (11.3) 14 (17.9) 35 (13.2)
4–6 49 (26.3) 29 (37.2) 78 (29.4)
mRS at one month n (%) (n ¼ 174) (n ¼ 73) (n ¼ 247)
2 49 (28.2) 32 (43.8) 81 (32.7)
0–1 125 (71.2) 41 (56.2) 166 (67.2)
2–3 15 (8.6) 7 (9.6) 22 (8.9)
4–6 34 (19.5) 25 (34.2) 59 (23.8)
Mortality at discharge 16 (8.5) 11 (14.1) 27 (10.2)
Mortality at one month after discharge 30 (16) 21 (26.9) 51 (20.1)

Table 4. Diagnostic and outcome predictors of acute febrile encephalopathy.

Effect of age and comorbid illness on one-month mortality: univariate analysis

Odds ratio 95% confidence interval (CI) P value

Age > 40 years 3.35 1.20–9.33 0.015

Comorbid illness
Diabetes mellitus 2.51 1.33–4.74 0.006
Systemic hypertension 1.94 0.98–3.82 0.063
Previous cerebrovascular disease 2.68 0.73–9.89 0.066
Renal failure 1.28 0.13–12.57 0.538
MRS  2 at discharge 84.70 11.1–646.0 <0.001
Intubation 4.22 1.52–11.69 0.005
GCS < 8 6.74 2.15–21.09 <.001
Effect of age and comorbid illness on diagnostic prediction of non-infectious etiology: Univariate analysis
Age > 40 years 2.18 1.2–3.8 <0.001
Diabetes mellitus 3.78 2.05– 6.96 0.005
Systemic hypertension 2.38 1.31–4.32 0.004
Multivariate analysis
Odds ratio (OR) 95%confidence interval (CI) P value
Diagnosis prediction
Headache 3.03 1.46–6.28 0.003
Nuchal rigidity 4.67 2.08–10.46 <0.001
Outcome prediction
GCS (<8) 4.2 1.8–10.1 0.001
Intubation 2.7 1.07–7.14 0.03
mRS  2 13.9 3.9–49 <0.001
Diabetes 2.1 1.03–4.45 0.04
6 Tropical Doctor 0(0)

45
Funding
40 The author(s) received no financial support for the research,
35 authorship, and/or publication of this article.
30
TBM
25
Undiffernated ORCID iDs
20
15 Scrub Manoj Job https://orcid.org/0000-0002-8670-2577
10 EBV Alice J Mathuram https://orcid.org/0000-0003-2769-1211
5 Hyperthermia Ronald AB Carey https://orcid.org/0000-0002-0899-1510
0 Sudha J Rajan https://orcid.org/0000-0003-4852-9301

References
1. Modi A, Atam V, Jain N, et al. The etiological diagnosis
Figure 2. Seasonal variation of various diseases presenting with and outcome in patients of acute febrile encephalopathy:
acute febrile encephalopathy. a prospective observational study at tertiary care center.
Neurol India 2012; 60: 168–173.
health care decision making and allocation of limited 2. Bhalla A, Suri V, Varma S, et al. Acute febrile encephal-
resources. opathy in adults from Northwest India. J Emerg Trauma
As our study was conducted in a tertiary hospital, Shock 2010; 3: 220–224.
3. Venkatesan A, Tunkel AR, Bloch KC, et al. Case defin-
the findings may not be representative of the commu-
itions, diagnostic algorithms, and priorities in encephal-
nity. Furthermore, PCR could not be done for all itis: consensus statement of the international encephalitis
patients with suspected viral encephalitis on account consortium. Clin Infect Dis 2013; 57: 1114–1128.
of cost and availability. 4. Marais S, Thwaites G, Schoeman JF, et al. Tuberculous
meningitis: a uniform case definition for use in clinical
research. Lancet Infect Dis 2010; 10: 803–812.
Conclusion 5. Thwaites GE, Chau TTH, Stepniewska K, et al.
AFE causes high morbidity burden and mortality. Diagnosis of adult tuberculous meningitis by use of clin-
Tuberculosis and scrub typhus should be considered ical and laboratory features. Lancet 2002; 360:
among differential diagnosis of AFE in India. Non- 1287–1292.
infectious causes portend to a poorer prognosis with 6. Jamil MD, Hussain M, Lyngdoh M, et al. Scrub typhus
meningoencephalitis, a diagnostic challenge for clinicians:
almost half the patients having severe disability at the
a hospital based study from North-East India. J Neurosci
end of one month. Timely diagnosis of heat-related ill- Rural Pract 2015; 6: 488–493.
nesses and metabolic, toxic causes should be made. 7. Bacterial meningitis – Symptoms, diagnosis and treat-
Recognising the geographical location and seasonal ment. BMJ Best Practice. https://bestpractice.bmj.com/
trends is crucial for prompt diagnosis, management, topics/en-us/539 (accessed 10 November 2018).
and prognostication of AFE. 8. Khan R, Quaiser S and Alam S. Clinical profile and prog-
nostic markers of acute febrile encephalopathy (AFE) in
IRB approval adult patients presenting to a North Indian tertiary care
hospital. Int J Nutr Pharmacol Neurol Dis 2015; 5: 95.
IRB Min. No. 9450 dated 05.06.2015.
9. Karmarkar SA, Aneja S, Khare S, et al. A study of acute
febrile encephalopathy with special reference to viral eti-
Authors’ contribution ology. Indian J Pediatr 2008; 75: 801–805.
MJ did the literature review, data collection, and data ana- 10. Singh RR, Chaudhary SK, Bhatta NK, et al. Clinical and
lysis and manuscript preparation. AJM, VPT, RABC, RI etiological profile of acute febrile encephalopathy in east-
were involved in conceptualisation, manuscript editing and ern Nepal. Indian J Pediatr 2009; 76: 1109–1111.
manuscript review. SJR conceptualised, designed the study, 11. Joshi R, Mishra PK, Joshi D, et al. Clinical presentation,
and reviewed data analysis, manuscript editing and review etiology, and survival in adult acute encephalitis syn-
process. drome in rural Central India. Clin Neurol Neurosurg
2013; 115: 1753–1761.
Acknowledgements 12. Beig FK, Malik A, Rizvi M, et al. Etiology and clinico-
The authors thank Dr Jasmine Prasad for technical support epidemiological profile of acute viral encephalitis in chil-
in data analysis. dren of western Uttar Pradesh, India. Int J Infect Dis
2010; 14: e141–e146.
13. Anga G, Barnabas R, Kaminiel O, et al. The aetiology,
Declaration of conflicting interests
clinical presentations and outcome of febrile encephalop-
The author(s) declared no potential conflicts of interest with athy in children in Papua New Guinea. Ann Trop
respect to the research, authorship, and/or publication of this Paediatr 2010; 30: 109–118.
article.