Journal of Infection (2005) 51, 287293

www.elsevierhealth.com/journals/jinf

Dengue infection during pregnancy and

transplacental antibody transfer in Thai mothers
Cecilia Perreta, Pornthep Chanthavanichb, Krisana Pengsaab,
Kriengsak Limkittikulb, Pornsom Hutajaroene, James E.G. Bunnc,
Bernard J. Brabinc,d,*
a

Paediatric Department, Pontificia Universidad Catolica de Chile, Chile

Department of Tropical Pediatrics, Faculty of Tropical Medicine Department, Mahidol University,
Bangkok, Thailand
c
Child and Reproductive Health Group, Liverpool School of Tropical Medicine, University of Liverpool,
Pembroke Place, Liverpool L3 5QA, UK
d
Emma Kinderziekenhuis, Academic Medical Centre, University of Amsterdam, Netherlands
e
Department of Obstetrics and Gynaecology, Rajvithee Hospital, Bangkok, Thailand
b

Accepted 6 October 2004

Available online 23 November 2004

KEYWORDS
Dengue antibodies;
Pregnancy;
Transplacental
transfer;
Congenital infection

Abstract Objectives. The objectives of this study were to estimate dengue

seroprevalence in a population of Thai pregnant women, living in a highly endemic
area and placental transfer of dengue antibodies.
Methods. A cross-sectional seroprevalence study of 245 pregnant women at
delivery.
Results. Dengue HAI antibodies were positive in 94.7%. Maternal age was the only
risk factor associated with dengue infection as older mothers (O20 years) were
significantly more likely to be seropositive than younger women (p!0.0001). Cord
antibody titres varied with maternal age and antibody titre, were significantly higher
in babies born to younger mothers (!20 years) (pZ0.01), and were significantly
correlated with maternal titre. Low birthweight babies had lower transfer ratios for
DEN-2 antibody (1.06) compared to heavier babies (1.36, pZ0.05). No mother or
neonate had dengue IgM detected. Two women were classified as recently, but not
currently infected with dengue virus and we consider it likely these were first
trimester infections. As no infant became infected the fetal infection was 0%.
Conclusions. Younger mothers were more likely to have been recently infected,
resulting in higher antibody titres. Maternal dengue antibody transfer was
proportional to maternal antibody concentration.
Q 2004 The British Infection Society. Published by Elsevier Ltd. All rights reserved.

* Corresponding author.
E-mail address: b.j.brabin@liv.ac.uk (B.J. Brabin).
0163-4453/$30.00 Q 2004 The British Infection Society. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.jinf.2004.10.003

288

Introduction
Many viral infections acquired or reactivated during
pregnancy represent a risk either to mothers or their
babies. Maternal treatment, which could reduce
maternal to child transmission and subsequent
sequelae in neonates is not available for many of
these infections. Dengue infection may cause
disease in neonates born to infected mothers, even
if mothers experience asymptomatic infection in
pregnancy, which commonly occurs. In countries
such as Thailand, with high transmission rates of
dengue infection, most women are seropositive
when they reach childbearing age. For example in
one study 75% were seropositive at 12 years and 97%
at 20 years of age.1 Many cases of dengue hemorrhagic fever (DHF) among adults have been reported
in South East Asia,2 highlighting the concept that
some adults remain susceptible to dengue infection.
Dengue during pregnancy is also particularly important in pregnant travellers from non-endemic
countries to countries where dengue is endemic.
The vertical transmission of dengue has been
infrequently described world-wide, although there
are reports from Cuba, Brazil, Malaysia, and Thailand36 which have occurred during outbreaks.
These report variable neonatal outcomes, from
asymptomatic infection to death.
Four serotypes of dengue virus can infect human
beings. After infection a long lasting serotype
specific immunity is acquired. High antibody titres
of one dengue serotype may be protective to
infection from a different serotype by cross-reaction, although if antibody titres have decreased
they may be non-protective to another serotype,
and could even enhance the immune response
causing DHF.7 The risk of DHF in infants is
associated exclusively with primary infection
related to passively acquired maternal antibodies.
These maternal antibodies are thought to protect
infants from dengue infection while antibody titres
are high, normally during the first 6 months of age.
In older infants and children a higher risk of DHF has
been reported which coincides with decreasing
titres of transplacentally acquired antibodies. A
strong correlation has been observed between
maternal neutralizing antibody titres to DEN-2 and
age of onset of DHF in infants. This is after 3 months
of age with an expected half-life for maternal IgG of
2.6 months. The time period when maternal
antibody is reduced to titres at which immunological enhancement occurs, coincides with the time of
onset of DHF in their infants.8
Even though dengue infection is very common in
endemic countries, attack rates in women of

C. Perret et al.
childbearing age and corresponding maternal to
child transmission rates have not been reported.
The objectives of this study were to estimate the
seroprevalence of dengue in a population of Thai
pregnant women living in a highly endemic area and
placental transfer of dengue antibodies as an
indicator of passively acquired infant immunity. A
secondary aim was to estimate pregnancy infection
risk and the risk of congenital infection if dengue
infection was found in this study population.

Methodology
A cross-sectional seroprevalence study was conducted in Bangkok, Thailand between April and May
of 2000.
A sequential sample of 245 pregnant women
attending Rajavithiee Hospital was requested to
participate at the time of delivery and an informed
consent was obtained. Maternal age, parity, residence, occupation, education, socioeconomic status, ethnic group, weight and gestational age were
recorded. Neonatal gestational age, sex, Apgar
score, birthweight (g) length (mm), head circumference (mm), placental weight (g) and the presence of clinical symptoms were collected.
A delivery blood sample was collected from
mothers and a cord sample from their babies. The
samples were centrifuged and sera were kept at K
20 8C. Maternal and cord samples were tested for
dengue antibodies by an in house ELISA (IgM and
IgG) and haemagglutination inhibition assay (HAI).
These tests were performed at AFRIMS (Army Force
Research Institute of Medical Science) in Bangkok.
ELISA capture IgM and IgG assays were performed as
described previously.9 Plates were sensitised using
goat anti-human IgM or goat anti-human IgG and
patients sera diluted to 1:100 was used. After
overnight incubation antigen solution containing
the four dengue serotypes was added and incubated
for 2 h. Anti-flavivirus IgG HRP conjugate in a
dilution of 1:250 was added and incubated for 1 h
and then, substrate was added. The results were
read by spectrophotometer at a wave length
492 nm, with R40 units considered positive.
HAI was performed as described previously.10
This included acetone extraction of sera to remove
non-specific inhibitors of haemagglutination. The
supernatant was discharged and sera were reconstituted adding one drop of goose red cells.
Supernatant (after centrifugation) was used for
HAI assay. Treated sera were diluted from 1:10 to
1:1280 in four different plates to be tested for the
four dengue antigens. Twenty-five micro litre of

Transplacental transfer of dengue antibodies

each dengue serotype was added to plates and
incubated at 4 8C overnight. Subsequently, 50 ml of
goose red cells was added and incubated at room
temperature for 2 h before reading. The HAI titre
was the last dilution for clear observation of the red
dot. Samples were considered positive for the
presence of dengue antibodies at 1:10 or higher
dilutions.
Dengue primary infection was defined according
to the laboratory (AFRIMS) definition as any serum
containing IgM R40 units with an IgM/IgG ratio
R1.8 and a positive HAI in a dilution !1:1280.
Dengue secondary infection was defined as any
serum containing IgM R40 units and an IgM/IgG
ratio !1.8 and positive HAI in a dilution R1:1280.
Past dengue infection was considered if serum
demonstrated an IgG R40 units, an absence of IgM
and/or (because of few Cve ELISA) a positive HAI !
1:1280. Recent but not current dengue infection
was defined as a negative IgM and positive HAI
R1:1280. Finally, negative infection was defined as
a negative IgM titre and HAI titres !1:10. ELISA IgM
and IgG for Japanese encephalitis (JE) were done
for all samples and cross-reactivity excluded in
those with a positive dengue IgG by calculating the
ratio between JE and dengue IgG antibody.
Statistical analysis used Epi Info version 6.0 and
SPSS 9.0. The non-parametric test KruskalWallis
was used for non-normal distributions and Chisquared values were calculated. For maternal
neonatal pairs, the geometric mean of the maternal
reciprocal HAI dilution, the geometric mean of cord
reciprocal HAI dilution and the geometric mean
cord/maternal HAI titres were calculated. The
latter measures IgG placental transfer. Geometric
means of maternal, cord and cord/maternal ratios
were compared for several risk factors using nonparametric tests. Maternal and cord titres were
compared using a paired sample t-test and the
Pearson correlation coefficient. p values %0.05
were considered statistically significant.
The maternal sample size required to estimate a
prevalence of infection and incidence of infection
during pregnancy was calculated considering the
prevalence of dengue infection expected in this
population. About 95% of pregnant women were
expected to be seropositive.11 The sample size to
estimate the attack rate for dengue infection
during childbearing age, was calculated estimating
a difference of 10% in the prevalence of dengue
antibodies between younger and older pregnant
women. With a level of significance (a) of 5% and
power (1Kb) of 80%, then 120 pregnant women for
each age group (%20 years old and R30 years old)
would be required in order to detect a statistically
significant difference in seroprevalence between

289
groups. This sample size was sufficient to detect a
difference of 20% between maternal and cord
antibody level.
This study received approval from the Ethical
Committee of the Liverpool School of Tropical
Medicine and from Mahidol University.

Results
A total of 245 pregnant women were studied. The
average age was 25.4G5.9 years (range 1440
years), and 80% were less than 30 years of age.
Mean maternal weight was 49.6G0.4 kg. Half of this
sample was classified as lower socioeconomic status
with income less than US$ 200/month. Nearly all
lived in Bangkok (92.2%). Half were primigravida
(48%) (Table 1). Mean birth weight was 2.99G
0.42 kg, 10.2% were low birth weight and 79.5%
were term infants.
Dengue HAI antibodies were positive in 94.7%,
but only 1.6% were positive by ELISA IgG. Seroprevalence was 84.4% in those %20 years, which was
significantly lower than older women O20 years,
whose seroprevalence was 98.3% (p!0.0001)
(Table 2). Four women had been infected by one
serotype alone. Two women (0.8%) were classified
as recently infected, but not currently infected
with dengue virus, because of high HAI dengue
antibody titres, positive IgG, but negative dengue
IgM. These women were both %20 years of age.
Absence of any dengue antibodies was observed in
13 mothers (5.3%) (Table 3). No mother or neonate
had dengue IgM detected.
Maternal age was the only risk factor associated
with dengue infection (pZ0.02). There was no
difference in seroprevalence according to place of
Table 1 Demographic characteristics of the 245 Thai
pregnant women
Variable
Parity
0
1
2
3
4
8
Education
Primary
Secondary
Tertiary
Occupation
Housewives
Workers

% of mothers (nZ245)
48.2
33.5
16.3
1.2
0.4
0.5
56.3
29.8
13.9
53.5
46.5

290

C. Perret et al.

Table 2

Seroprevalence of dengue antibodies

Age, years

Number of
women

Number
positive

Seropositive %

%20
2125
2630
3135
3640

64
65
64
34
18

54
65
62
33
18

84.4a
100
96.9
97.1
100

pZ0.0001 compared to women over 20 years (c2 test).

residence, socioeconomic status, maternal education, ethnic group or parity. Geometric mean
dengue maternal antibody titres were compared
against several variables. HAI DEN antibody titres
varied with maternal age and maximum antibody
titres were in the range of 2030 years of age. Mothers
!30 years had a mean HAI DEN-2 of 44.7G44
compared to 31G28.3 (pZ0.02) in those R30 years.
Cord antibody levels varied with maternal age
and maternal antibody titre. The cord level for HAI
DEN-2 and DEN-3 were significantly higher in babies
born to younger mothers. Cord mean titres for DEN2 were 54.89G63 in babies born to mothers !30
years, compared to 38.3G36.3 for mothers R30
years (pZ0.05) (Fig. 1). The higher the maternal
antibody level the higher the cord antibody levels,
(correlation coefficient R0.76 for the four dengue
serotypes antibodies). Fig. 2 shows the direct
correlation between maternal and cord antibody
levels for DEN-4 antibodies. Similar correlations
were observed for the other dengue serotypes (data
not shown). Cord antibody titre was higher than
maternal antibody for all four dengue serotypes.
(Table 4) The geometric mean cord/maternal
transfer ratio was O1 for all serotypes. The ratio
for HAI DEN-3 antibodies was significantly lower
than the other antibodies (Table 5).
Low birthweight babies (!2500 g) had significantly lower transfer ratios for DEN-2 (1.06)
compared to heavier babies, (1.36 (pZ0.05)).
Table 3
women

Figure 1 Geometric mean of cord HAI DEN-2 titres by

maternal age.

Discussion
Using the HAI DEN method a high seroprevalence of
past dengue infection in pregnant women (94.7%)
was found, but only 1.6% of women were seropositive using an ELISA assay. The sensitivity of the
ELISA used in this study was very low for the
diagnosis of past dengue infection and as a result
this low seroprevalence using ELISA IgG was an
expected result as this method has been standardized for acute infection diagnosis. This in house
ELISA should only be used for diagnosis of acute
infection when sensitivity increases to 97% and
differentiation between primary and secondary
infection is possible but not for the diagnosis of
past infection.9 Seronegative samples occurred in
5.3% of subjects, who would be susceptible to
primary dengue infection, and 1.6% were susceptible to secondary infection as they were seropositive for only one dengue serotype. Altogether, 6.9%

Dengue infection status in Thai pregnant

Dengue diagnosis

Number of pregnant
women

Recent infection
Past infection with
more than 1 serotype
Past infection with 1
serotype
Seronegative
Total

2
226

0.8
92.2

1.6

13
245

5.3
100

Figure 2 Correlation between maternal and cord HAI

DEN-4 antibody titres.

Transplacental transfer of dengue antibodies

Table 4

Geometric mean concentration of maternal and cord dengue antibody

Dengue antibody
HAI DEN-1
HAI DEN-2
HAI DEN-3
HAI DEN-4
a

291

Maternal antibody (NZ245)

Cord antibody (NZ177)

Mean

(SD)

Mean

(SD)

58.10
41.33
80.43
104.04

(53.36)
(41.10)
(78.13)
(145.7)

66.33
50.84
88.69
128.20

(67.92)
(58.39)
(86.24)
(166.8)

Pa

K8.22
K9.51
K8.27
K24.16

0.002
0.0001
0.003
0.0001

Paired sample test.

of pregnant women were susceptible to either

primary or secondary dengue infection. This figure
increased to 15% in the 1420 years age group,
whose seroprevalence was 84.4%. This corresponds
with the fact that two recently infected pregnant
mothers were under 20 years of age. The difference
of seroprevalence between younger mothers and
those above 30 years of age (97.1%) corresponds to
the attack rate of dengue infection during the main
childbearing period (12.7%). The annual incidence
of dengue infection during this 10-year period is,
therefore, estimated at 1.27 and during the 9
months of pregnancy the incidence would be 0.95%.
This figure is comparable to the incidence of
maternal infection (0.8%) found during the present
study. Mothers with recent infection were probably
infected very early during pregnancy as although
IgM negative they still had high titres of HAI
antibody and low titres of IgG.
High titres of these non-neutralizing antibodies
last about 6 months following infection.12 These
mothers are likely to have been infected during, the
first trimester of pregnancy. None of their neonates
became infected defined by the presence of
symptoms or positive IgM, so we can assume that
the fetal transmission rate in this study was 0% for
symptomatic congenital infection. As these
mothers were presumably infected early in pregnancy, it might considered that the risk of congenital infection was very low, when maternal
infection occurs at this time, or that it leads to
miscarriage. The number of infected mothers was
Table 5 Geometric mean cord/maternal dengue
antibody ratios for IgG and HAI antibodies
Dengue antibodies

Cord/maternal transfer ratio (SD)

HAI DEN-1
HAI DEN-2
HAI DEN-3
HAI DEN-4

1.34
1.33
1.17a
1.33

Paired difference

(1.1)
(0.8)
(0.5)
(0.7)

pZ0.007 (for difference with other HAI DEN antibody

ratios).

very small and it cannot be concluded from this

study sample that this is representative of all first
trimester infections. A larger sample, in a country
with a lower seroprevalence during the childbearing age, would be a preferred study group,
particularly if this was an area where outbreaks
occurred. A low rate of fetal transmission is
consistent with other studies which have reported
no cases of congenital dengue infection in neonates
born to mothers infected early in pregnancy.13
Although the latter prospective study of women
with dengue infection early in pregnancy did show
an increase in the rate of early miscarriage (3 of 22
pregnancies), but the fetuses were not investigated
for congenital infection.13 Only one study has
reported neonates who were IgM positive and
these were asymptomatic, and the timing of
maternal infection was not clear.4
Conversely, all reported cases of symptomatic
congenital dengue infection have occurred in
neonates born to mothers infected very late in
pregnancy, mainly when they were symptomatic at
the time of delivery.3,6,14 Accordingly, the gestation of maternal infection is an important factor
related to fetal infection risk, with symptomatic
infection possibly unlikely with early gestational
infection and increased with infection late in
pregnancy. One hypothesis for this might be that
maternal infections occurring close to the time of
delivery would have insufficient protective antibodies to be transferred and consequently direct
viremia into the fetal blood stream may result. This
would explain the poor prognosis in congenital
dengue infections acquired late in pregnancy.
The fetal transmission rate could not be estimated in this study as no infected neonates were
observed, consequently the expected incidence of
congenital infection in this population is unknown.
The congenital dengue infection rate would be
expected to be higher in the 1520 year group as the
proportion of susceptibles in that group is higher
(15%). Maternal age is a risk factor for pregnancy
dengue infection. Younger mothers would be more
likely to become infected, compared to older

292
mothers, who are more likely to be seropositive
and, therefore, less susceptible for dengue
infections.
Neonates with intrauterine infection with dengue virus have IgM antibodies lasting for between 5
and 9 months,4 compared with only 2, 3 months in
their mothers. This might suggest persistence of the
virus due to an immature fetal immune system that
is unable to control the infection. Further research
is warranted to determine the long-term consequences of congenital dengue infection. These
babies would be expected to have a delayed risk
of DHF due to other serotypes, as they would
produce their own antibodies, which are longlasting compared to maternal antibody.
In the present study, dengue maternal antibody
titres correlated with maternal age, as older mothers
had significantly lower antibodies titres. Younger
mothers were more likely to have been recently
infected, resulting in higher titre of antibodies. This
may also relate to the fact that older mothers were
infected several years previously allowing sufficient
time for antibody levels to decrease, especially if
secondary or tertiary infection had not occurred. Nonneutralizing antibodies (HAI antibodies) decrease
rapidly after infection unlike neutralizing antibodies,
which are life-long.
Maternally transferred antibodies have been
related with DHF in infants. They play two roles
during infancy. In the first few months of life passively
transferred dengue antibodies confer protection to
infants until they reach a non-protective concentration, which is !1:10 titre of neutralizing antibody.
Enhancing activity occurs when these lower titre have
been reached, which affects the outcome of infant
dengue infection by increasing the risk of DHF.
Accordingly, both maternal age and maternal dengue
infection could influence the time of onset of DHF in
infants. This explains the strong correlation between
the maternal neutralizing antibody titre and age of
onset of DHF in infants reported by Kliks et al. (1988).8
Maternal dengue antibody transfer was proportional to maternal antibody concentration.
Babies had higher antibody titre compared to their
mothers. This suggests an active transplacental
antibody transfer process.15 The transfer differed
between dengue antibody types. HAI DEN-4 antibodies had the best correlation between maternal
and cord titres. The difference between maternal
and cord titres was greater for HAI DEN-2 and DEN-4
antibodies suggesting better placental transfer and
probably an additional transfer mechanism as this
varies also according to antibody avidity.16
The placental antibody transfer was significantly
decreased in LBW babies. No difference was found
in small-for-gestational age and term babies,

C. Perret et al.
suggesting that low transfer probably occurred
mainly in preterm babies. Low transfer of antibodies has been previously described in preterm
babies and is probably due to an immature placenta
lacking Fcg II receptors. This receptor is expressed
during the third trimester.17 Mothers infected a
long-time ago, and older mothers would be more
likely to not transmit dengue antibodies, as they
would have lower antibody levels. This would
reduce the risk of DHF in their infants.

Acknowledgements
This study was undertaken as part of the Masters
degree in Tropical Paediatrics of the University of
Liverpool at the Liverpool School of Tropical
Medicine.

References
1. Ferguson N, Donnelly C, Anderson R. Transmission dynamics
and epidemiology of dengue: insights from age-stratified
seroprevalence surveys. Philos Trans R Soc Lond B Biol Sci
1999;354:75768.
2. Wali J, Biswas A, Handa R, Aggarwal P, Wig N, Dwivedi S.
Dengue haemorrhagic fever in adults: a prospective study of
110 cases. Trop Doct 1999;29:2730.
3. Chye J, Lim C, Ng K, Lim J, George R, Lam S. Vertical
transmission of dengue. Clin Infect Dis 1997;25:13747.
4. Fernandez R, Rodriguez T, Borbonet F, Vasquez S,
Guzman M, Kouri G. Estudio de la relacion dengue-embarazo
en un grupo de madres cubanas. Rev Cubana Med Trop 1994;
46:768.
5. Figueiredo L, Carlucci R, Duarte G. Estudio prospectivo com
lactantes cuyas maes tiveram dengue durante a gravidez.
Rev Inst Med Trop Sao Paulo 1994;36:41721.
6. Thaithumyanon P, Thisyakorn U, Deerojnawong J, Innis BL.
Dengue infection complicated by severe hemorrhage and
vertical transmission in a parturient woman. Clin Infect Dis
1994;18:2489.
7. Morens D. Antibody-dependent enhancement of infection
and pathogenesis of viral disease. Clin Infect Dis 1994;19:
50012.
8. Kliks S, Nimmannitya S, Nisalak A, Burke D. Evidence that
maternal dengue antibodies are important in the development of dengue hemorrhagic fever in infants. Am J Trop Med
Hyg 1988;38:4119.
9. Innis BL, Nisalak A, Nimmannitya S, et al. An enzyme-linked
immunosorbent assay to characterize dengue infection
where dengue and Japanese encephalitis co-circulate. Am
J Trop Med Hyg 1989;40:41827.
10. Clarke DH, Casals DJ. Techniques for hemagglutination and
hemagglutination inhibition with arthropod-borne viruses.
Am J Trop Med Hyg 1958;7:56173.
11. Chareonsook O, Foy H, Teeranatkul A, Silarug N. Changing
epidemiology of dengue hemorragic fever in Thailand.
Epidemiol Infect 1999;122:1616.

Transplacental transfer of dengue antibodies

12. Vaughn D, Green S, Kalayanarooj S, et al. Dengue in the early
febrile phase: viremia and antibody response. J Infect Dis
1997;176:32230.
13. Carles G, Peiffer H, Talarmin A. Effects of dengue fever during
pregnancy in French Guyana. Clin Infect Dis 1999;28:63740.
14. Poli L, Chungue E, Soulignac O, Gestas P, Kuo P, PapouinRauzy M. Dengue materno-foetale. A propos de 5 cas
observes pendant lepidemie de Tahiti (1989). Bull Soc
Pathol Exot 1991;84:51321.

293
15. Matre R, Tonder O, Endresen C. Fc receptors in the human
placenta. Scand J Immunol 1975;4:7415.
16. Avanzini MA, Pignatti P, Chirico G, Gasparoni A, Jalil F,
Hanson LA. Placental transfer favours high avidity IgG
antibodies. Acta Paediatr 1998;87:1805.
17. Simister NE, Story CM. Human placental Fc receptors and the
transmission of antibodies from mothers to fetus. J Reprod
Immunol 1997;37:123.