You are on page 1of 4

Article

Tropical Doctor
0(0) 1–4
! The Author(s) 2017
Clinical and laboratory factors Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
associated with mortality in dengue DOI: 10.1177/0049475517693993
journals.sagepub.com/home/tdo

Atul Saroch1, Vivek Arya2, Nitin Sinha3, RS Taneja2,


Pooja Sahai4 and RK Mahajan5

Abstract
Dengue is endemic in more than 100 countries, giving rise to an increased number of deaths in the last five years in the
South-East Asian region. We report our findings from a retrospective study of adults admitted with confirmed dengue at
our institution. We studied the clinical and laboratory parameters associated with mortality in these patients. Of the
172 hospitalised patients studied, 156 (90.69 %) recovered while 16 (9.3%) died. Univariate analysis showed altered
sensorium on presentation, lower haemoglobin and haematocrit levels, higher serum creatinine, higher serum trans-
aminase and lower serum albumin levels to be significantly associated with mortality in dengue. Further, using stepwise
multivariate logistic regression, altered sensorium (P ¼ 0.006) and hypoalbuminemia (P ¼ 0.013) were identified as inde-
pendent predictors of mortality in dengue. Identification of these parameters early in the course of disease should
prompt intensification of treatment in dengue cases.

Keywords
Dengue, mortality, fever, New Delhi, predictive factors

Introduction myalgia, arthralgia, rash, a positive tourniquet test


Dengue virus (DENV) is transmitted by vectors Aedes (defined as the presence of  20 petechiae per square
aegypti and Aedes albopictus, and has four serotypes. inch) or leucopenia (defined as a white blood cell
Infection by any one of these may be either asymptom- count < 5.0  109/L); and (2) dengue confirmed by viral
atic or produce severe symptoms even death. According detection (a) PCR or NS1 antigen, or virus culture
to the WHO classification of 2009, dengue can be clas-
sified as with or without warning signs, or severe
(including dengue haemorrhagic fever and dengue 1
Assistant Professor, Department of Medicine, Post Graduate Institute of
shock syndrome).1 Though there are numerous studies Medical Education and Research, Dr Ram Manohar Lohia Hospital,
on factors predicting severity in dengue, there are rela- New Delhi, India
2
tively few studies on factors associated with mortality. Professor, Department of Medicine, Post Graduate Institute of Medical
Education and Research, Dr Ram Manohar Lohia Hospital, New Delhi,
India
3
Methods Associate Professor, Department of Medicine, Post Graduate Institute of
Medical Education and Research, Dr Ram Manohar Lohia Hospital,
Our retrospective study was conducted at the Post New Delhi, India
4
Graduate Institute of Medical Education and Research Senior Resident, Department of Microbiology, Post Graduate Institute of
Medical Education and Research, Dr Ram Manohar Lohia Hospital,
(PGIMER), Ram Manohar Lohia Hospital (RML),
New Delhi, India
New Delhi, India, which caters chiefly as a tertiary refer- 5
Professor, Department of Microbiology, Post Graduate Institute of
ral centre for Delhi and surrounding areas. All patients Medical Education and Research, Dr Ram Manohar Lohia Hospital,
with confirmed dengue were admitted to the designated New Delhi, India
Dengue Ward, and those cases admitted according to
the register between the months of September and Corresponding author:
Atul Saroch, Assistant Professor, Department of Medicine, Post Graduate
October 2015 were included in our study. Inclusion cri- Institute of Medical Education and Research, Dr Ram Manohar Lohia
teria were: (1) fever with at least two of the following Hospital, New Delhi, India.
symptoms: headache, retro-orbital or ocular pain, Email: atulsaroch@gmail.com
2 Tropical Doctor 0(0)

positive; or antibody response (b) IgM seroconversion or Altered sensorium was present in a significantly
IgG titre increase  4 in paired serum samples.2 higher percentage in those who died (37.5%) compared
Patients with mixed infection, multiple medical with survivors (4.5%) (P < 0.001). Of those six patients
co-morbidity or those aged < 14 years (admitted to the who had altered sensorium and died, two had an ence-
paediatric wards) were excluded. Patients with underly- phalitis-like presentation, two had an intracranial
ing medical co-morbidity were excluded to reduce treat- haemorrhage and the remaining two were not evaluated
ment bias, as they were managed with help of different in detail as they died within 24 h of admission. Among
medical specialties with different protocols. the survivors, seven had altered sensorium related to
Cases were divided into two groups: those who sur- hyponatraemia (n ¼ 2), hypernatraemia (n ¼ 2) and an
vived and those who died. Parameters compared were: encephalitis-like presentation (n ¼ 3).
age, sex, place of residence (Delhi or outside Delhi), dur- Other patient characteristics such as gender, area of
ation of fever (i.e. time between onset of fever and admis- residence, age, duration of fever, symptoms of lethargy,
sion), headache, abdominal pain, vomiting, diarrhoea, myalgia, headache, vomiting, diarrhoea, abdominal
myalgia, arthralgia, lethargy, rash, bleeding from any pain, rash and bleeding tendency were not significantly
site, level of consciousness, respiratory rate and mean different (Table 1). Those who died had significantly
blood pressure (MBP). Laboratory parameters compared lower Hb, lower HCT, higher serum creatinine, higher
included: haemoglobin (Hb), haematocrit (Hct), platelet liver transaminases and lower serum albumin compared
counts, total white cell count, serum creatinine, serum to survivors (Table 2).
albumin, aspartate transaminase and alanine transamin- All clinical factors significantly associated with mor-
ase. All patients received care according to WHO guide- tality as detected from the univariate analysis were
lines.2 Data containing all the above information for each included in the multivariate logistic regression analysis.
case at time of admission were recorded on a Proforma. Independently associated factors were altered sensor-
Data were analysed using STATA SE 12 (Stata ium (P ¼ 0.006) and hypoalbuminemia (P ¼ 0.013)
Corp, TX, USA). Variables with normal distribution only.
were summarised as mean and standard deviations
and compared using the Student’s t-test. Variables
with non-normal distributions were summarised as
Discussion
median and interquartile ranges (IQR) and compared Dengue is a significant public health problem, particu-
using the Mann–Whitney U test. Categorical variables larly in India, and the number of related deaths has
were expressed as frequencies and percentages and ana- increased dramatically in recent decades. The mortality
lysed with 2 tests or Fisher’s exact test, as appropriate. rate reported varies from 11%3 to 27.5%,4 somewhat
Univariate analysis to determine all the factors signifi- higher than in our study.
cantly affecting mortality was performed, following Neurologic manifestations, other than headache,
which a stepwise multivariate logistic regression was occur in approximately 4–6% of dengue cases.
conducted using the forward selection method to deter- Neurological complications have been classified into
mine factors independently associated with mortality. dengue virus encephalopathy, dengue virus encephal-
In all tests, P < 0.05 was considered significant. itis, immune-mediated syndromes, neuromuscular com-
plications and dengue-associated stroke. They have
been attributed to different pathophysiological mechan-
Results isms, such as direct central nervous system (CNS) viral
A total of 268 patients were admitted, of whom 96 involvement, metabolic disturbances affecting CNS
patients were excluded because of mixed infection function, haemorrhage and virus-generated auto-
(n ¼ 32), underlying medical co-morbidity (n ¼ 28), immune reactions, leading to CNS inflammation and
incomplete medical records (n ¼ 26) or negative con- demyelination.5,6 Evidence of DENV neurotropism
firmatory results for dengue (n ¼ 10). Out of 28 patients was also found in autopsy studies. DENV antigens
with medical co-morbidity, we found chronic liver dis- have been isolated from brain tissues in fatal dengue
ease (n ¼ 9), coronary artery disease (n ¼ 7), chronic cases.7 Altered sensorium as a factor related to
kidney disease (n ¼ 5), active tuberculosis on treatment increased mortality was documented in retrospective
(n ¼ 4) and diabetes with nephropathy and retinopathy study conducted in Pakistan.8
(n ¼ 3). Of these, causes of death were: coronary artery Hypoalbuminemia may result from increased vascular
disease with congestive cardiac failure (n ¼ 3), chronic permeability, a capillary leak syndrome, negative acute
alcoholic liver disease (n ¼ 1) and active pulmonary phase reactant or hepatic injury. Hypoproteinaemia or
tuberculosis (n ¼ 1). Thus, of the 172 hospitalised hypoalbuminemia were seen in 12.9% in a large study
patients with confirmed dengue included in our study, from Kolkata, India,4 while reports are in the range
156 (90.7%) recovered while 16 (9.3%) died. of 16.5–76% in other studies.9–11 It was likewise
Saroch et al. 3

Table 1. Baseline characteristics and clinical parameters at admission of 172 hospitalised adult
patients with dengue.

Patients who Patients who


survived (n ¼ 156) died (n ¼ 16) P value

Baseline characteristics
Male 98 (62.8%) 9 (56.2%) 0.377
Female 58 (37.2%) 7 (43.8%)
Residential area 0.392
Delhi 135 (86.5%) 13 (81.2%)
Outside Delhi 21 (13.5%) 3 (18.8%)
Clinical parameters
Vital signs
HR median (IQR), beats/minute 90 (82–98) 88 (84–110) 0.330
MAP, median (IQR), mmHg 81.7 (73–86) 71.3 (60–90) 0.160

History and physical examination


Symptoms and signs n (%) n (%)
Myalgia 75 (48.1) 8 (50.0) 0.545
Lethargy 92 (59.0) 9 (56.2) 0.477
Headache 80 (51.3) 9 (56.2) 0.565
Fever > 4 days 84 (53.8) 12 (75.0) 0.366
RR > 18 breaths/min 135 (86.5) 14 (87.5) 1.000
Bleeding 52 (33.3) 4 (25.0) 0.457
Rash 48 (30.8) 3 (18.8) 0.051
Arthralgias 45 (28.8) 4 (25.0) 0.384
Abdominal pain 90 (57.7) 7 (43.8) 0.209
Persistent vomiting 106 (67.9) 8 (50.0) 0.122
Diarrhoea 34 (21.79) 6 (37.5) 0.201
Altered sensorium 7 (4.48) 6 (37.5) <0.01
HR, heart rate; MAP, mean arterial pressure; RR, respiratory rate.

Table 2. Laboratory parameters at admission of 172 hospitalised adult patients with dengue.

Patients who Patients who


survived (n ¼ 156) died (n ¼ 16)

Median (IQR) Median (IQR) P value

Haemoglobin (g/L) 127 (109–143) 113 (102–120) 0.002


Haematocrit (%) 38 (33.7–42.0) 33 (27.7–34.4) <0.001
TLC (cells  109/L) 5.0 (3.4–6.9) 8.2 (4.8–10.9) 0.025
Platelet count (cells per mL) 40,000 30,000 0.738
(median)
Platelet counts
100  109/L 130 (83.33%) 12 (75%) 0.733
>100  109/L 26 (16.66%) 04 (25%) 0.501
Creatinine (mmol/L) 70 (53–88) 176(70–300) 0.003
Albumin (g/L) 33 (29–36) 27 (21–29) 0.003
AST (IU/L) 134.5 (73.7–233) 285 (161–2797) 0.008
ALT (IU/L) 83.5 (47–137) 539 (111–1356) 0.005
IQR, interquartile range; TLC, total leucocyte count.
4 Tropical Doctor 0(0)

significantly associated with mortality in studies from 4. Chowdhury R, Pan K, Sarkar A, et al. Predictors of
Brazil and India.12,13 dengue mortality in a tertiary care hospital at Kolkata:
While other factors such as low Hb, low HCT, low A cross sectional study. Int J Med Sci Public Health 2013;
platelet count and deranged liver function have else- 2: 254–257.
5. Domingues RB, Kuster GW, Onuki-Castro FL, et al.
where been associated with mortality,14 these were
Involvement of the central nervous system in patients
not significant in our findings. This may be explained with dengue virus infection. J Neurol Sci 2008; 267:
by closer attention and treatment of these correctable 36–40.
factors in institutions from where these patients were 6. Puccioni-Sohler M, Soares CN, Papaiz-Alvarenga R,
referred. Early recognition of altered sensorium and et al. Neurologic dengue manifestations associated with
hypoalbuminemia appear, though, to sound the alarm intrathecal specific immune response. Neurology 2009; 73:
for intensification of care in an attempt to reduce 1413–1417.
mortality. 7. Ramos C, Sánchez G, Pando RH, et al. Dengue virus in
the brain of a fatal case of hemorrhagic dengue fever.
Acknowledgements J Neurovirol 1998; 4: 465–468.
8. Almas A, Parkash O and Akhter J. Clinical factors asso-
The authors thank the Medical Record Department at
ciated with mortality in dengue infection in a tertiary care
Dr RML Hospital for the help they extended in retrieving
center. Southeast Asian J Trop Med Public Health 2001;
the patient records.
41: 333–340.
9. Karoli R, Fatima J, Siddiqi Z, et al. Clinical profile of
Declaration of conflicting interests dengue infection at a teaching hospital in North India.
The authors declared no potential conflicts of interest with J Infect Dev Ctries 2012; 6: 551–554.
respect to the research, authorship, and/or publication of this 10. Wong M and Shen E. The utility of liver function tests in
article. dengue. Ann Acad Med Singapore 2008; 37: 82–83.
11. Itha S, Kashyap R, Krishnani N, et al. Profile of liver
Funding involvement in dengue virus infection. Natl Med J India
2005; 18: 127–130.
The author(s) received no financial support for the research,
12. Amancio FF, Heringer TP, de Oliveira Cda C, et al.
authorship, and/or publication of this article.
Clinical profiles and factors associated with death in
adults with dengue admitted to Intensive Care Units,
References Minas Gerais, Brazil. PLoS ONE 2015; 10: e0129046.
1. World Health Organization. Comprehensive guidelines for 13. Schmitz L, Prayag S, Varghese S, et al.
prevention and control of dengue and dengue hemorrhagic Nonhematological organ dysfunction and positive fluid
fever. Geneva: WHO, 2011. balance are important determinants of outcome in adults
2. World Health Organization. Dengue guidelines for diagno- with severe dengue infection: a multicenter study from
sis, treatment, prevention and control - New edition. Geneva: India. J Crit Care 2011; 26: 441–448.
WHO, 2009. 14. Samanta J and Sharma V. Dengue and its effects on liver.
3. Wali JP, Biswas A, Handa R, et al. Dengue haemorrhagic World J Clin Cases WJCC 2015; 3: 125–131.
fever in adults: a prospective study of 110 cases. Trop Doct
1999; 29: 27–30.

You might also like