Epidemiol. Infect., Page 1 of 5.

© Cambridge University Press 2014

doi:10.1017/S0950268814002647

Risk factors associated with an outbreak of dengue fever/dengue

haemorrhagic fever in Hanoi, Vietnam

D. T. T. TOAN 1 *, L. N. HOAT 1 , W. HU 2 , P. WRIGHT 3 AND P. MARTENS 4

1
Hanoi Medical University, Institute of Training for Preventive Medicine and Public Health, Vietnam
2
School of Public Health and Social Work, Institute of Health and Biomedical Innovation, Queensland University
of Technology, Australia
3
The Medical Committee Netherlands-Vietnam
4
International Centre for Integrated Assessment and Sustainable Development, Maastricht University,
Maastricht, The Netherlands

Received 17 June 2014; Accepted 15 September 2014

SUMMARY
Dengue fever/dengue haemorrhagic fever (DF/DHF) appears to be emerging in Hanoi in recent
years. A case-control study was performed to investigate risk factors for the development of DF/
DHF in Hanoi. A total of 73 patients with DF/DHF and 73 control patients were included in
the study. The risk factor analysis indicated that living in rented housing, living near uncovered
sewers, and living in a house discharging sewage directly into to ponds were all significantly
associated with DF/DHF. People living in rented houses were 2·2 times more at risk of DF/DHF
than those living in their own homes [adjusted odds ratio (aOR) 2·2, 95% confidence interval
(CI) 1·1–4·6]. People living in an unhygienic house, or in a house discharging sewage directly to
the ponds were 3·4 times and 4·3 times, respectively, more likely to be associated with DF/DHF
(aOR 3·4, 95% CI 1–11·7; aOR 4·3, 95% CI 1·1–16·9). These results contribute to the
understanding of the dynamics of dengue transmission in Hanoi, which is needed to implement
dengue prevention and control programmes effectively and efficiently.

Key words: Dengue fever, dengue haemorrhagic fever, risk factors, Hanoi.

I N T RO D U C T I O N reported DF/DHF infections has been stated to be

Dengue fever/dengue haemorrhagic fever (DF/DHF) around 96 million per year. However, many cases go
unreported and it is estimated that the real number
is a fast-spreading vector-borne disease associated
with a significant public health impact. Dengue is re- could be as high as 390 million (95% confidence inter-
ceiving attention all over the world for its epidemic ex- val 284–528) cases [2].
pansion and high mortality rate [1]. Before 1970, DF/ Vietnam is located in the heart of the endemic area
DHF was detected in only nine countries but has now for DF/DHF. It is recognized as a major cause of
spread to over 100 countries. The global number of mortality and morbidity in Vietnam and ranks
among the top ten communicable disease in terms of
overall health burden [3]. In 1958, the first DHF
* Author for correspondence: Dr D. T. T. Toan, Biostatistics and case was described in Vietnam and the first reported
Medical Informatics Department, Institute of Training for outbreak occurred in southern Vietnam in 1963,
Preventive Medicine and Public Health, Hanoi Medical University,
Vietnam.
resulting in 116 deaths [4]. The estimated morbidity
(Email: [email protected]) and mortality rates between 1979 and 2005 were
2 D. T. Toan and others

33–462/100 000 and 0·1–2·7/100 000, respectively, per backward stepwise logistic regression model was ap-
year. Although dengue transmission occurs in both plied to identify the risk factors for DF/DHF. The in-
rural and urban areas in Vietnam, 73% of the popu- dependent variables consisted of: age, education,
lation live in rural areas and therefore the majority occupation, having been in an epidemic area within
of DF/DHF cases and deaths are from these areas [5]. past 6 months; place of living: in urban or rural area,
Hanoi in Northern Vietnam is a large city where in a rented house or near an open sewer; and living con-
DF/DHF appears to be emerging as a major public ditions such as unhygienic environment, other family
health concern. In 2009, Hanoi experienced its largest member had DF/DHF during past 6 months, presence
ever recorded outbreak. In this study we set out to in- of mosquitoes, presence of larvae in water containers,
vestigate which risk factors are associated with DF/ directly discharging sewage to ponds, and being
DHF in patients admitted to hospitals in Hanoi. flooded during the rainy season. The significance level
Identifying the risk factors can help to design and was set at P < 0·2. Associations are expressed in
apply effective preventive and control strategies. terms of odds ratios (ORs) with 95% confidence inter-
vals (CIs). All analyses were performed using Stata
statistical software, v. 12.1 (StataCorp., USA).
M AT E R IA L S AN D M E T H O D S
This study was performed in Hanoi at the National
R E S ULTS
Hospital of Tropical Diseases, the Hospital of
Hanoi Medical University and the Infectious The characteristics of the 73 DF/DHF patients and 73
Diseases Department of Bach Mai Hospital, from control patients are shown in Table 1. The mean age
August 2009 to March 2010. Approval from the hos- of the case group and the control group were compar-
pital administration was obtained before approaching able. The distribution of education level and occu-
the patients. We performed a prospective matched pation were quite similar between the case and
case-control study on patients who were diagnosed control groups; most individuals had college edu-
with DF/DHF; control patients were from the same cation and many were students.
hospital but negative for DF/DHF. The criteria for Table 2 displays variables related to the residence of
notification of DF were based on the guidelines of the study population. In both groups, three quarters
the Ministry of Health, 1999, on surveillance, diag- resided in the inner districts of Hanoi. More than
nosis and treatment of dengue. The guidelines stipu- half of the participants in the case group were living
late that individuals are suspected to have dengue in rented houses, significantly higher than in the con-
when they have acute febrile illness (538 °C) of 2–7 trol group, with only one third. People in the DF/
days’ duration, with 52 of the following non-specific DHF group were living in smaller houses with a larger
manifestations of DF: headache, retro-orbital pain, household size compared with those in the control
myalgia, arthralgia, rash, haemorrhagic manifesta- group. In particular, a significantly higher proportion
tions, and leucopenia [6]. The control group was of people in the case group were living in homes near
matched to the study group for sex and age (±5 suitable breeding areas for mosquitoes, such as open
years). The characteristics of the 73 DF/DHF patients sewers.
and the 73 control patients are shown in Table 1. Table 3 presents the adjusted and unadjusted ORs
All cases and control patients were interviewed dur- and CIs of variables affecting DF/DHF included in
ing their period of hospitalization by two specially the final logistic regression model. The analysis
trained interviewers who administered a standard revealed that living in a rented house, living near
questionnaire. The questionnaire included infor- open sewers and untreated water discharging directly
mation on demographic characteristics of the patients into nearby ponds/lakes were all significantly asso-
such as age, education level and occupation as well as ciated with DF/DHF. Living in a rented house
variables related to their residence: location of house, increased risk by 2·2 times (aOR 2·2, 95% CI 1·1–
type of house, source of water supply, and type of 4·6). Living in an unhygienic house or one directly dis-
water containers in the house. The questionnaire had charging sewage into ponds increased risk by 3·4 times
been pre-tested before being used in the study. and 4·3 times, respectively (aOR 3·4, 95% CI 1–11·7;
χ2 and Mann–Whitney tests were used to test for dif- aOR 4·3, 95% CI 1·1–16·9). Detecting mosquitoes in
ferences in demographic characteristics between DF/ the house or living near an open sewer constituted a
DHF patients and control patients. A multivariable very high risk (aOR 6·3, 95% CI 0·7–59; aOR 6·9,
 Risk factors for DF/DHF in Vietnam 3

Table 1. Demographic characteristics of dengue fever/dengue haemorrhagic fever (DF/DHF) and control patients

Characteristic DF/DHF (n = 73) Controls (n = 73) P

Sex Male 36 (49·32%) 36 (49·32%)

Female 37 (50·68%) 37 (50·68%) >0·05*
Age (years) Mean ± S.D. 28·1 ± 9·4 27·4 ± 9·9 >0·05†
Range 17–75 18–78
Education Elementary school 9 (12·3%) 10 (13·7%)
Middle school 10 (13·7%) 4 (5·5%) >0·05*
High school 11 (15·1%) 6 (8·2%)
University/college 43 (58·9%) 53 (72·6%)
Occupation Office worker 17 (23·3%) 24 (32·9%) >0·05*
Business person 7 (9·6%) 12 (16·4%)
Home worker 4 (5·5%) 7 (9·6%)
Pupil/student 36 (49·3%) 22 (30·1%)
Other 9 (12·3%) 8 (11·0%)

* By χ2 test.
† By Mann–Whitney test.

Table 2. Housing variables of dengue fever/dengue haemorrhagic fever (DF/DHF) and control patients

Variables DF/DHF (n = 73) Controls (n = 73) P

Living area Inner district 54 (74·0%) 55 (75·3%) >0·05*

Outer district 19 (26·0%) 18 (24·7%)
Accommodation status Rented 41 (56·2%) 25 (34·3%) <0·05*
Owned 32 (43·8%) 48 (65·7%)
Type of housing Brick construction 43 (58·9%) 35 (47·9%)
Temporary house 7 (9·6%) 7 (9·6%)
Old condominium 11 (15·1%) 15 (20·6%) >0·05*
New condominium 0 (0·0%) 3 (4·1%)
Other 12 (16·4%) 13 (17·8%)
Area of household (m2) Mean ± S.D. 54·1 ± 47·1 68·2 ± 69·9 >0·05†
Range 8–200 9–350
Number of people in household Mean ± S.D. 4·2 ± 2·3 4·9 ± 4·3 >0·05†
Range 1–10 1–19
House with water storage containers Yes (water tank without 22 (10·7%) 21 (1·3%) >0·05*
cover)
No 53 (70·8%) 54 (72·0%)
House with water storage containers Yes 11 (15·1%) 13 (17·8%) >0·05*
infested with larvae and/or pupae No 52 (71·2%) 56 (76·7%)
Don’t know 10 (13·7%) 4 (5·5%)
House environment Pond/lake/river 16 (19·3%) 21 (24·1%) >0·05*
Open sewer 33 (39·7%) 13 (14·9%) <0·01*
Garbage collection point 7 (36·8%) 12 (63·2%) >0·05*

* By χ2 test.
† By Mann–Whitney test.

95% CI 0·9–71·9, respectively), but the difference be- much less frequently in suburban districts, which is
tween the case and control groups was not significant. consistent with the statistical reports on the DF epi-
demic in 2009. The highest morbidity rate was found
in the 15–30 years age group; at higher ages the rate
DI S C US S IO N became very low. Dung & Cam reported in 2003
The results of this study revealed that DF/DHF in that more than 90% of dengue morbidity was in indi-
Hanoi appeared mainly in downtown districts and viduals aged between 15 and 25 years [7]. This could
4 D. T. Toan and others

Table 3. Crude and adjusted logistic regression odds higher rates of morbidity. This is consistent with
ratios on selected variables other national and international studies [9–17]. In
Brazil [10] the epidemic was associated with proximity
Logistic to uncontrolled waterways and stagnant water in
Independent Crude OR regression
variables Grouping (95% CI) aOR (95% CI)
tanks, gutters, and cans. In Pakistan, poor condition
of the house, such as uncovered toilet water tank or
Living in Yes 2·2 (1·1–4·8) 2·2 (1·1–4·6) leaking water pipes, was a highly significant risk factor
rented house
for the presence of Aedes foci [10]. In Vietnam, a study
Other family Yes 2 (0·9–4·6) 1·96 (0·9–4·3)
member had in Binh Thanh, Dong Thap [11] found that while most
DF/DHF families stored their drinking water, only three quar-
during past 6 ters of them covered their water storage containers.
months Nowadays, although Hanoi residents seldom use
Unhygienic Yes 3·3 (0·9–12) 3·4 (1–11·7) water storage containers like jars or pots, the large
house
Mosquitoes in Yes 6·8 (0·7–64·2) 6·3 (0·7–59·5)
area of natural water surfaces available makes an
house annual epidemic almost unavoidable [12]. Because
Near open Yes 6·9 (0·7–67·1) 7·9 (0·9–71·9) dengue-spreading mosquitoes spend three stages of
sewer their life-cycle in water, it is recommended to reduce
Discharging Yes 5·9 (1·3–26·4) 4·3 (1·1–16·9) mosquito living sites, e.g. long-term stagnant water
sewage
storage inside and outside the house, open sewers,
directly into
ponds and natural water surfaces.
According to around half of the respondents, their
aOR, Adjusted odds ratio; CI, confidence interval; DF, den- homes and the surrounding environment are not really
gue fever; DHF, dengue haemorrhagic fever. clean. Although the difference between case and con-
trol groups was not statistically significant, it did
reflect how the unsanitary, polluted living environ-
be related to differences in lifestyle, time spent out- ment in the modern environment of the capital is
doors near vectors, sleeping without mosquito nets favourable for the vector and the disease.
or other aspects of inadequate disease prevention Environmental factors have a powerful influence on
among the young. Another explanation could be the appearance of DF/DHF. Simple interventions
that the disease can create lifelong immunity for the could help; in our study, people living in houses
individual, so that older persons who have been ex- with uncovered water tanks were 7·9 times more likely
posed more often may have more resistance, decreas- to get DF. Those in houses discharging sewage
ing the morbidity rate. directly into ponds had a 5·9 times higher risk than
Considering how many cases were young people, it those in houses with sanitary sewage systems. For
is perhaps not surprising that around half of the posi- civil authorities it is important to note that people
tive cases were students. The second dominant group who stay in rented accommodation had double the
was office workers. This result differs from the result risk of those owning their homes. These influences
of a study in Kamphaeng Phet, Thailand [8] where of the environmental factors have also been found in
65% of the patients were farmers. The discrepancy is other studies. A 2001 study in Brazil revealed that
probably due to a different job distribution in the people living in a slum area had a nearly 10 times
two regions studied. Students may also be prominent higher risk of DF [10]. In An Giang, Vietnam, people
in the cases in our study because many students live living in homes with gardens or water tanks had a
in rented accommodation, which was another risk fac- threefold higher risk [12].
tor for dengue. The combination of cramped rented This study was conducted soon after Hanoi had ex-
living accommodation, an unsanitary environment, perienced heavy flooding in August 2009 which
infrequent cleaning, little knowledge of DF, and little reduced the selection and recall bias often found in a
interest in prevention activities may explain why the case-control study. However, our data came from
morbidity rate in the rental group is rising. the three major hospitals for infectious diseases in
People living near stagnant water like ponds, lakes Hanoi, and may have missed factors related to cases
and rivers or open sewers, or favourable mosquito that did not present to these hospitals. Our logistic
breeding places like garbage collection points, had regression model was set up to identify factors that
 Risk factors for DF/DHF in Vietnam 5

may influence dengue transmission dynamics. In a 4. Ha DQ, Ninh TU. Virological surveillance of dengue
complex epidemiological setting, transmission dynam- haemorrhagic fever in Vietnam, 1987–1999. Dengue
Bulletin 2000; 24: 18–23.
ics involve interactions among people, dengue viruses,
5. Tien TKN, et al. Epidemiological analysis of deaths
vectors and ecosystems, where biotic and abiotic associated with dengue haemorrhagic fever in southern
determinants have both direct and indirect influences Vietnam in 1999–2000. Dengue Bulletin 2001; 25: 28–32.
on transmission. Our model cannot account for all 6. Vietnam Ministry of Health. Guidelines on the surveil-
possible variables. More information is needed to lance, diagnosis, and treatment of dengue hemorrhagic
further unravel the predictors of dengue transmission fever. Hanoi, Vietnam, 1999.
7. Dung NT, Cam NN. Dengue fever/dengue hemorrhagic
in Hanoi, which will help to inform planning of pre-
fever in Hanoi – 2003. Journal of Vietnam Preventive
vention and control measures. Medicine 2005; 1: 73–77.
8. Koenraadt CJ, et al. Dengue knowledge and practices
and their impact on Aedes aegypti populations in
CO NC LU SI O N S Khamphaeng Phet, Thailand. American Journal of
This integrated analysis of the eco-social determinants Tropical Medicine and Hygiene 2006; 74: 692–700.
9. Nguyen ND. Knowledge, attitude and practice on den-
of DF transmission risk contributes to improved
gue/dengue hemorrhagic fever prevention among
understanding of the dynamics of dengue transmission mothers in Ho Chi Minh City, Vietnam. Journal of
in Hanoi. Our study shows various well-known risk Ho Chi Minh City’s Medical Practice 1999; 2: 119–124.
factors of DF which are components of human behav- 10. Heukelbach J, et al. Risk factors associated with an out-
iour, can be addressed through modifications of mu- break of Dengue fever in a favela in Fortaleza, north-
nicipal services. The social and demographical east Brazil, Tropical Medicine & International Health
2001; 6: 635–642.
contexts and lifestyles of people in Hanoi will signifi-
11. Huong LTT, et al. Knowledge, attitude and practice on
cantly influence transmission and all information on dengue prevention among residences in Binh Thanh
this population will support dengue prevention and commune, Thanh Binh District, Dong Thap province,
control programmes. Vietnam. Journal of Preventive Medicine 2009; 11:
36–42.
12. Be PV, et al. Evaluation on knowledge, attitude and
ACK N OWL E D G E ME N T S practice of people about dengue fever in two communes
of An Giang province, Vietnam in 2004. Journal of
We thank Dr Nguyen Nhat Cam, Vice-Director of Preventive Medicine 2005; 3: 56–64.
Hanoi Preventive Medicine Center and his staff for 13. Nga NT. Epidemiology of dengue fever/dengue hemor-
the assistance on collecting the epidemiological data. rhagic fever in Hai Ba Trung District, Hanoi City.
This study was funded by the Netherlands Higher Master thesis of Hanoi Medical University, Vietnam,
Education (NPT) project on ‘Strengthening teaching 8/1982.
14. Toan DTT, et al. Hot spot detection and spatio-
and research capacity in preventive medicine in
temporal dispersion of dengue fever in Hanoi,
Vietnam’. Vietnam. Global Health Action 2013; 6: 18632.
15. Itrat A, et al. Knowledge, awareness and pratices re-
garding Dengue fever among the adult population of
R E FE RE NC E S Dengue hit cosmopolitan. PLoS ONE. Published
1. Gubler DJ. Dengue and dengue hemorrhagic fever. online: 9 July 2008. doi:10.1371/journal.pone.0002620.
Clinical Microbiology Reviews 1998; 11: 480–496. 16. Nalongsack S, et al. Knowledge, attitude and practice
2. Bhatt S, et al. The global distribution and burden of regarding dengue among people in Pakse, Laos.
dengue. Nature 2013; 496: 504–507. Nagoya Journal of Medical Science 2009; 71: 29–37.
3. Tien TKN, et al. Predictive indicators for forecasting 17. Syed M, et al. Knowledge, attitudes and practice
epidemic of dengue/dengue haemorrhagic fever through regarding dengue fever among adults of high and low
epidemiological, virological and entomological surveil- socioeconomic groups, Journal of Pakistan Medical
lance. Dengue Bulletin 1999; 23: 34–39. Association 2010; 60: 243–247.