Dengue in T a ele s: Kine ics o Vi emia and NS1
An igenemia and Thei Associa ions wi h Clinical
Pa ame e s
Elina O. E a
1,2
*
.
, Essi M. Ko honen
3.
, Liina Vou ilainen
3,4.
, Eili Huh amo
3
, Olli Vapalah i
3
, Anu Kan ele
1,2
1Di ision o In ec ious Diseases, Depa men o Medicine, Helsinki Uni e si y Cen al Hospi al, Helsinki, Finland, 2Depa men o Bac e iology and Immunology,
Haa man Ins i u e, Facul y o Medicine, Uni e si y o Helsinki, Helsinki, Finland, 3Depa men o Vi ology, Haa man Ins i u e, Facul y o Medicine, Uni e si y o Helsinki,
Helsinki, Finland, 4Finnish Fo es Resea ch Ins i u e, Van aa Resea ch Uni , Van aa, Finland
Abs ac
Despi e he inc easing numbe s o a el-acqui ed dengue, ew s udies ha e assessed i ologic ma ke s o he disease in
non-endemic popula ions. We examined he kine ics o diagnos ic ma ke s and hei associa ions wi h clinical pa ame e s in
93 pa ien s wi h a el-acqui ed dengue e e . Kine ics analyses sugges ed a longe a e age du a ion o i emia (9 days,
CI95%: 8–10) and non-s uc u al p o ein 1 (NS1) an igenemia (15 days, CI95%: 12–20) han epo ed in endemic popula ions.
While none o he es s su iced alone, he bes diagnos ic co e age was achie ed by combining an ibody de ec ion wi h
RNA o NS1 es ing. S udied by eg ession models, ea ly ela i e le els o i emia and NS1 an igenemia p o ed o be
signi ican ly associa ed wi h se e al clinical pa ame e s: high i emia p edic ed g ea e likelihood and inc eased leng h o
hospi aliza ion, he deg ee o NS1 an igenemia co ela ed posi i ely wi h hema oc i and li e ansaminases, and bo h
i emia and NS1 an igenemia le els nega i ely wi h pla ele coun s in ollow-up. Le els o i emia and NS1 an igenemia
may se e as p edic o s o he clinical mani es a ions in a el-acqui ed dengue.
Ci a ion: E a EO, Ko honen EM, Vou ilainen L, Huh amo E, Vapalah i O, e al. (2013) Dengue in T a ele s: Kine ics o Vi emia and NS1 An igenemia and Thei
Associa ions wi h Clinical Pa ame e s. PLoS ONE 8(6): e65900. doi:10.1371/jou nal.pone.0065900
Edi o : Eng Eong Ooi, Duke-Na ional Uni e si y o Singapo e G adua e Medical School, Singapo e
Recei ed Janua y 2, 2013; Accep ed Ap il 29, 2013; Published June 3, 2013
Copy igh : ß2013 E a e al. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed
use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed.
Funding: This wo k was suppo ed by he Finnish go e nmen al subsidy o heal h science esea ch, he Academy o Finland, and he Finnish Medical
Founda ion. The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip .
Compe ing In e es s: The au ho s ha e decla ed ha no compe ing in e es s exis .
* E-mail: [email p o ec ed]
.These au ho s con ibu ed equally o his wo k.
In oduc ion
Dengue is he mos common mosqui o-bo ne i al disease
wo ldwide, causing 50–100 million in ec ions annually [1]. I is
endemic in mos opical and sub opical pa s o he wo ld,
especially in u ban a eas, many o which a e popula a el
des ina ions. Acco dingly, ongoing dengue epidemics ha e man-
i es ed hemsel es as an inc easing numbe o in ec ions no only in
endemic popula ions, bu also among a ele s e u ning om he
dengue endemic a eas [2,3].
The labo a o y diagnos ics o dengue i us (DENV) in ec ions is
cu en ly based on i us isola ion, and de ec ion o DENV RNA,
non-s uc u al p o ein 1 (NS1), and DENV-speci ic an ibodies
[4,5]. Few labo a o ies can p o ide he ull ange, ye none o
hese assays co e s he en i e disease pe iod [4,5]. Fo a a ional
choice and iming o he es s, he kine ics o he a ious diagnos ic
ma ke s needs o be unde s ood. Fo e icien diagnos ics, i
appea s ha wo o mo e me hods should be combined.
The clinical ou come o DENV in ec ions anges in se e i y
om asymp oma ic o non-speci ic eb ile illness o classical
dengue e e and o se e e dengue cha ac e ized by one o mo e
o he ollowing: plasma leakage, se e e bleeding, and se e e o gan
impai men [5]. Se e e dengue in ec ions should be ea ed in
dedica ed high-dependency uni s, whe e a mo ali y o less han
1% can be achie ed [3]. In a ele s, dengue is a ely a li e-
h ea ening disease; ye , se e e o ms o disease a e also seen [6,7].
Epidemiologic s udies ha e iden i ied se e al isk ac o s o
se e e disease, including p e ious exposu e o he e ologous DENV
se o ype, in ec ion wi h ce ain i al s ains and se o ypes, young
age, emale sex, and ce ain gene a ian s o he hos [8–12]. A
high deg ee o se um i emia and NS1 an igenemia ha e been
associa ed wi h a mo e se e e disease ou come in endemic
popula ions [13–17]. Di e ing esul s ha e also been epo ed
[18,19], hough, and s udies in a el-acqui ed dengue ha e been
lacking.
Recen ly, he impo ance o ea ly diagnosis and isk p edic ion
o clinical ou come ha e been emphasized [3]. In addi ion o
endemic a eas, hese should also be s udied among a ele s who
may di e om endemic popula ions in a numbe o essen ial
espec s, such as lack o p e ious immuni y agains he e o ypic
se o ypes, dissimila i y in p e-exis ing immuni y agains o he
la i i uses (e.g. Japanese encephali is, ick-bo ne encephali is,
yellow e e ), a ie y o in ec ing DENV se o- and geno ypes, as
well as age and gene ic backg ound.
The p esen s udy in es iga ed he clinical and diagnos ic da a
o 93 Finnish a ele s wi h acu e dengue in ec ions, aiming a (1)
desc ibing he kine ics o DENV i emia, NS1 an igenemia, and
DENV-speci ic an ibodies, (2) assessing hei use in diagnos ics as
combina ions, and (3) examining he po en ial co ela ion be ween
PLOS ONE | www.plosone.o g 1 June 2013 | Volume 8 | Issue 6 | e65900
diagnos ic ma ke s ( i emia and NS1 an igenemia) and clinical
pa ame e s in a el-acqui ed dengue.
Me hods
E hics s a emen
As a e ospec i e egis y s udy based solely on pa ien iles and
ou ine, a chi ed blood samples, ins ead o pa ien consen and an
e hics app o al, he s udy equi ed esea ch clea ances om he
ins i u es and he Minis y o Social A ai s and Heal h.
Acco dingly, he s udy was app o ed by he esea ch boa ds o
he Depa men o In e nal Medicine o Helsinki Uni e si y
Cen al Hospi al, Helsinki Uni e si y Hospi al labo a o y (HUS-
LAB), and he Minis y o Social A ai s and Heal h.
S udy design
In Finland, labo a o y diagnos ics o dengue has been based on
se ological es ing pe o med in a single labo a o y o all he
pa ien s h oughou he coun y (HUSLAB, Helsinki Uni e si y
Cen al Hospi al). Fo his s udy, we e ospec i ely eco ded da a
on all he DENV IgM-posi i e pa ien s in Finland 1999–2008 (a
o al o 154 cases). F om hese pa ien s, all samples aken wi hin 21
days since illness onse we e collec ed, including he ea ly IgM-
nega i e se a. Fou pa ien s we e excluded due o suspicion o alse
posi i e IgM. This g oup consis ed o one published case o
Japanese encephali is [20], one labo a o y-con i med ick-bo ne
encephali is (s ong posi i e an i-TBE IgM in se um and
ce eb ospinal luid along wi h a cha ac e is ic clinical pic u e,
and only weakly posi i e se um an i-dengue IgM), and 2 pa ien s
who did no show diagnos ic se ocon e sion in se ial samples.
A e addi ionally excluding hose wi h insu icien clinical
in o ma ion (57 pa ien s), he da a comp ised a o al o 139
samples om 93 pa ien s (1–3 samples pe pa ien ). The i s
samples om a subg oup o pa ien s (72/93) we e e alua ed
p e iously [21]; hese ea ly-phase se a we e included in he p esen
s udy along wi h con alescen samples o analyses o he kine ics
o i ologic ma ke s, and wi h addi ional clinical da a o s udying
he associa ions be ween i ologic ma ke s and clinical pa ame-
e s.
The clinical and labo a o y da a we e analyzed by wo kinds o
s a is ical app oaches: (1) hose explo ing he kine ics o diagnos ic
ma ke s ei he wi h a single- es app oach o as combina ions o
wo es s, and (2) hose examining he associa ions be ween
i ologic ma ke s and clinical pa ame e s. Fo analyses o he
kine ics, we included all 139 se um samples aken wi hin 21 days
since he onse o symp oms (93 pa ien s, 1–3 samples each). Fo
he associa ion analyses, we included a subg oup o 89 pa ien s
wi h da a on he ollowing po en ial con ounding ac o s: age, sex,
day o illness, p esence o co-in ec ions and ch onic diseases, which
we e con olled in he s a is ical analyses.
Collec ion o clinical da a
The ollowing clinical in o ma ion was e ie ed e ospec i ely
om medical eco ds: demog aphic cha ac e is ics (age, gende ,
coun y o bi h), ch onic diseases (long-las ing condi ion ha can
be con olled bu no cu ed, such as diabe es melli us), p e ious
accina ions agains o he la i i uses, a el des ina ion, symp-
oms, esul s o ou ine labo a o y in es iga ions, p esence o co-
in ec ions (addi ional diagnosis o an acu e in ec ious disease o he
han dengue), possible hospi aliza ion and i s leng h. The da e o
he onse o symp oms was designa ed as day 1. Labo a o y
pa ame e s we e classi ied as being no mal, dec eased, o ele a ed
acco ding o cu en e e ence alues. The se e i y o he cases was
assessed acco ding o he WHO dengue classi ica ion c i e ia alid
a he ime o da a collec ion [22].
Vi ological and se ological assays
The an i-DENV IgM es s we e pe o med wi h a comme cial
enzyme immunoassay (EIA) (Focus Technologies) ollowing
manu ac u e ’s ins uc ions. The an i-DENV IgGs we e de e -
mined by an in-house immuno luo escence assay as p e iously
desc ibed [23]. The es s we e ca ied ou as dilu ion se ies o 1:10
o 1:640, employing ace one- ixed DENV-3 in ec ed Ve o E6 cells
as he an igen.
RNA ex ac ion, he de ec ion o RNA by eal- ime e e se
ansc ip ion polyme ase chain eac ion (RT-PCR) ( es ed in 132
samples), and o he NS1 an igen by EIA ( es ed in 135 samples)
we e pe o med as p e iously desc ibed [21]. The se um aliquo s
we e s o ed a 270uC and 220uC o RNA and NS1 de ec ion,
espec i ely. The RNA ex ac ions we e ca ied ou om 100 mlo
se um specimens by he QIAamp Vi al RNA Mini Ki (Qiagen)
acco ding o manu ac u e ’s ins uc ions. Fo he DENV NS1
an igen de ec ion we used he comme cial Pla elia Dengue NS1
Ag EIA assay (Bio-Rad) adhe ing o manu ac u e ’s ins uc ions.
Some o he NS1 Ag es s (69/135) and one-s ep eal- ime RT-
PCRs (69/132) we e ca ied ou sepa a ely p io o his s udy [21].
Fo he 63 se a es ed o RNA o his s udy only, he S a agene
he mocycle was used ins ead o he ABI P ism 7700 Sequence
De ec ion Sys em. The compa abili y and ep oducibili y o he
esul s ob ained wi h he wo pieces o equipmen was asce ained
by means o app op ia e in e nal con ols. The in ec ing DENV
Table 1. Backg ound cha ac e is ics o he 93 a ele s
wi h dengue.
%N
Gende
Male 56 52/93
Female 44 41/93
E hnic o igin
Finnish 88 82/93
O he 12 11/93
Ch onic diseasesa
Gene ally heal hy 79 72/91
Ch onic diseases 21 19/91
His o y o p e ious la i i us accina ionb
Japanese encephali is accine 11 10/92
Yellow e e accine 20 18/92
Tick-bo ne encephali is accine 1 1/92
Any o he abo e 22 20/92
Geog aphic egion isi edb
Sou h-Eas Asia 50 46/92
Sou h Cen al Asia 21 19/92
Cen al Ame ica and Ca ibbean 14 13/92
Sub-Saha an A ica 8 7/92
Sou h Ame ica 7 6/92
Sou hwes Asia 1 1/92
The median age o pa ien s was 37 yea s (in e qua ile ange: 28 o 45 yea s).
a
Da a missing o wo pa ien s.
b
Da a missing o one pa ien .
doi:10.1371/jou nal.pone.0065900. 001
Vi ologic Ma ke s in T a ele s’ Dengue
PLOS ONE | www.plosone.o g 2 June 2013 | Volume 8 | Issue 6 | e65900
se o ype was de e mined by i us isola ion, DENV- yping RT-
PCR [24] and consequen sequencing [21,25], o by di ec
sequencing o RT-PCR p oduc s [24,26] ob ained di ec ly om
se um samples.
The de aul se ing o Finnish a ele s is a p ima y in ec ion,
as he as majo i y a e – apa om a accina ed mino i y –
la i i us naı
¨ e. As high le els o IgG wi h low o e en absen IgM
le els du ing he ea ly acu e phase ha e been associa ed wi h
seconda y in ec ions [5], o he pu pose o his s udy, cases wi h
posi i e RT-PCR esul , IgG i e s $80 (IFA), and a nega i e IgM
(IgM-EIA) es in he acu e phase we e conside ed as possible
seconda y cases.
S a is ical analyses
All s a is ical analyses we e pe o med wi h R so wa e [27].
The kine ics o se um DENV RNA, NS1 an igen, and DENV-
speci ic an ibodies we e s udied using gene alized addi i e mixed
models (GAMM) [28]. In all models, a smoo hing spline o he
numbe o days since illness onse was employed as he
explana o y a iable. Two kinds o uni a ia e models we e
o med: (1) hose wi h Gaussian e o dis ibu ions and an iden i y
link unc ion, whe e con inuous alues o diagnos ic ma ke s
(in e se cycle h eshold [c ] - alue, NS1 a io, IgM index, and
log
10
- ans o med IgG i e) we e used as dependen a iables, and
(2) hose wi h binomial e o dis ibu ions and a logi link unc ion
wi h bina y ou comes (nega i e/posi i e) o he ou diagnos ic
es s as dependen a iables. Simila ly, we s udied he kine ics o
di e en combina ions o wo diagnos ic es s o e he cou se o
illness, using GAMM wi h binomial e o dis ibu ion and a logi
link unc ion, he classes o he dependen a iable being nega i e
in bo h es s/posi i e in ei he o he es s.
Associa ions be ween i ologic ma ke s (se um DENV RNA
and NS1 an igen) and clinical pa ame e s ( ou ine labo a o y es
esul s, symp oms du ing ollow-up, likelihood and leng h o
hospi aliza ion) we e s udied by eg ession models. The ull model
had he clinical pa ame e as he dependen a iable, and
po en ial con ounding ac o s (age, sex, day o illness, p esence
o co-in ec ions and ch onic diseases) as explana o y a iables.
Each o he i ologic ma ke s (se um DENV RNA and NS1
an igen) we e included in he ull model in u n, and a model se
cons i u ing 95% o Akaike weigh s o all models nes ed wi hin he
ull model was hen a e aged [29] using MuMIN package [30] in
R so wa e. Gaussian e o dis ibu ions, an iden i y link unc ion,
and con inuous diagnos ic explana o y a iables ( he in e se c -
alue om RT-PCR [cu -o c - alue – sample c - alue], NS1
a io om an igen EIA) we e applied o models wi h con inuous
alues o ou ine labo a o y indings as dependen a iables. A
Poisson e o dis ibu ion adjus ed o o e dispe sion, a log link
unc ion, and con inuous diagnos ic a iables we e used in models
s udying he du a ion o hospi aliza ion. Fo models examining
bina y dependen a iables (likelihood o abno mal labo a o y
indings/occu ence o di e en symp oms/hospi aliza ion), we
employed binomial e o dis ibu ions, a logi link unc ion, and
bina y diagnos ic ou comes (se um DENV RNA/NS1 posi i i y).
Table 2. Clinical cha ac e is ics o he 93 a ele s wi h
dengue.
%N
Clinical symp oms
Fe e 93 86/92
Rash 71 65/92
Headache 64 59/92
Myalgia 55 51/92
Fa igue 51 47/92
O he gas oin es inal symp oms 46 42/92
Nausea 45 41/92
A h algia 34 31/92
Respi a o y symp oms 27 25/92
Hemo hagic mani es a ions 24 22/92
Vomi ing 24 22/92
Re o-o bi al pain 17 16/92
P u i us 9 8/90
Shock 0 0/92
Rou ine labo a o y indings in ollow-up
a
Anaemia (,134/117 g/L) 9 5/56
Ele a ed Hb (.167/155 g/L) 13 11/87
Low Hc (,39%/35%) 9 5/56
Ele a ed Hc (.50%/46%) 8 7/87
Leukopenia (,3.4610
9
/L) 67 59/88
Th ombocy openia (,150610
9
/L) 78 68/87
Ele a ed AST (.45/35 U/L) 78 49/63
Ele a ed ALT (.70/45 U/L) 61 51/83
Ele a ed c ea inine (.100/90 mmol/L) 25 10/40
Pa ien s posi i e o he a ious diagnos ic es s
b
PCR 74% 67/90
NS1 79% 72/91
IgM
c
100% 93/93
IgG 98% 91/93
Se o ype iden i ied
DENV-1 26 24/93
DENV-2 8 7/93
DENV-3 24 22/93
DENV-4 3 3/93
No known 40 37/93
Co-in ec ions
d
A leas one co-in ec ion 19 17
e
/91
Gas oin es inal in ec ion 9 8/91
Respi a o y ac in ec ion 4 4/91
U ina y ac in ec ion 2 2/91
O he bac e ial disease 2 2/91
Mala ia 2 2/91
O he pa asi ic disease 1 1/91
Hospi aliza ion 79 73
/92
a
Re e ence alues o males/ emales in pa en heses.
b
A leas once du ing he 21 days since illness onse . The iming o se um
sampling was no s anda dized.
c
Selec ion c i e ion o en e ing he s udy.
d
Diagnosed acco ding o cu en p ac ice.
e
O hese pa ien s, 9/17 (53%) we e posi i e o PCR o NS1 a some poin . The
pa ien s nega i e o PCR and NS1 p o ided samples on illness days 6–21.
Du a ion o hospi aliza ion: median 4 days, in e qua ile ange 3 o 6 days.
Abb e ia ions: ALT, alanine ansaminase; AST, aspa a e ansaminase; DENV,
dengue i us; Hb, hemoglobin; Hc , hema oc i .
doi:10.1371/jou nal.pone.0065900. 002
Table 2. con .
Vi ologic Ma ke s in T a ele s’ Dengue
PLOS ONE | www.plosone.o g 3 June 2013 | Volume 8 | Issue 6 | e65900
Explana o y a iables we e conside ed s a is ically signi ican (on
0.05 le el), i he 95% con idence in e als o hei coe icien s
excluded ze o.
Resul s
Pa ien cha ac e is ics
Table 1 and Table 2 summa ize he backg ound cha ac e is ics
and clinical da a o he s udy popula ion. The majo i y o he
pa ien s we e Finnish adul s wi h no unde lying ch onic diseases.
Mos o he dengue cases we e acqui ed in Asia (72%).
All pa ien s we e classi ied as ha ing an uncomplica ed dengue
in ec ion. Mos o he cases we e ca ego ized as p ima y in ec ions.
Only h ee pa ien s we e conside ed o ha e a p obable seconda y
in ec ion: in he i s sample hey had a conside able an i-DENV
IgG le el concomi an ly wi h a nega i e an i-DENV IgM, ye hei
samples we e posi i e o DENV-RNA. No o he cases wi h a
conside able ea ly IgG and low o non-exis en IgM we e
de ec ed. None o he suspec ed seconda y cases had ecei ed
la i i al accines in he pas .
The majo i y (79%) we e admi ed o hospi al, he mean leng h
o hospi aliza ion being 4.5 (SD 2.5) days. Thei ea men was
symp oma ic, and some pa ien s we e also gi en an imic obials,
ei he because hey we e ini ially suspec ed o ha e a bac e ial
in ec ion o because a simul aneous co-in ec ion (19% [17/91] o
he cases) was diagnosed. The mos common co-in ec ion was
bac e ial gas oen e i is.
All ou DENV se o ypes we e de ec ed in his pa ien
popula ion (Table 2). DENV-1 (24/56) and DENV-3 (22/56)
we e he p edominan se o ypes, accoun ing o 82% (46/56) o
cases wi h iden i ied se o ype.
Kine ics o diagnos ic ma ke s
Figu e 1 shows he kine ics o plasma i emia, NS1 an igen-
emia, and DENV-speci ic IgM and IgG an ibodies in he 93
pa ien s s udied. DENV RNA and NS1 an igen we e de ec ed
al eady on he 2
nd
day o illness ( he ea lies se um samples), wi h a
p obabili y o 91% (95% CI: 82–96% o PCR posi i i y and 80–
96% o NS1 posi i i y) (Figu e 1E, 1F). On he 7 h day o illness,
65% (95% CI: 54–74%) o he se a we e s ill posi i e o DENV
RNA, and 80% (95% CI: 71–87%) we e posi i e o NS1. Pa ien s
we e likely o emain PCR posi i e up o illness day 9 (95% CI:
day 8–10), whe eas NS1 posi i i y would las longe , un il illness
day 15 (95% CI: day 12–20). On day 21, none o he se a p o ed
posi i e o DENV RNA, while 20% we e s ill posi i e o NS1
(Figu e 1E, 1F).
Du ing he i s 12 days o illness, all se um samples om
pa ien s wi h DENV-1 in ec ion (n = 35), bu only 69% (n = 32) o
hose om pa ien s wi h DENV-3 in ec ion we e NS1 posi i e
(p = 0.0003, Fishe ’s exac es ). In he 13 samples aken a e day
12, his di e ence was no longe seen. The kine ics o i emia
showed no se o ype-dependen di e ences (da a no shown).
Figu e 2 shows he sensi i i ies o he combina ions o wo
diagnos ic es s o e he i s h ee weeks o illness. When RNA
de ec ion was combined wi h IgM de ec ion, only 1% (1/120) o
he se um samples emained nega i e ( he nega i e sample aken
on illness day 14). The combina ion o NS1 and IgM de ec ion
emained nega i e in 3% (4/120) o he samples. Al hough no
p ima ily in ended o iden i ying acu e in ec ion, he combina-
ions wi h IgG de ec ion we e also s udied. When ei he RNA o
NS1 es s we e used oge he wi h IgG de ec ion, 2% (3/129) o
3% (4/130) o he samples we e nega i e, espec i ely. When
combining IgM and IgG de ec ion, he esul s p o ed nega i e o
10% (12/124) o he samples; when using RNA and NS1 de ec ion
oge he , 25% (32/127) p o ed nega i e.
Associa ions be ween i ologic ma ke s and clinical
pa ame e s
Associa ions be ween i ologic ma ke s (se um i emia/NS1
an igenemia) and clinical pa ame e s we e s udied wi h eg ession
models in a subg oup o 89 pa ien s who p o ided da a on he
ollowing po en ial con ounding ac o s con olled in he analyses
(age, sex, day o illness, p esence o co-in ec ions and ch onic
diseases). All associa ions epo ed below we e s a is ically signi -
ican on he 0.05 le el when con ounding ac o s we e con olled
o . Table S1, and Figu es S1 and S2 show he associa ions
be ween i ologic ma ke s and clinical pa ame e s. Tables S2 and
S3 p o ide he a e aged coe icien s o all explana o y a iables
and hei 95% CIs o all models s udied.
Associa ions be ween i emia and clinical
pa ame e s. While con olling o po en ial con ounding ac-
o s, e.g. day o illness, ela i e le els o DENV RNA in he i s
se um sample co ela ed nega i ely wi h leukocy e and pla ele
nadi s in ollow-up (Figu e S1, Table S3). In addi ion, o e all PCR
posi i i y a he ime o p esen a ion p edic ed a highe p obabili y
o leukopenia, h ombocy openia, and ele a ed alanine and
aspa a e ansaminases (ALT, AST) du ing he illness (Tables
S1 and S2).
No associa ions we e ound be ween ini ial DENV RNA
posi i i y and he a ious symp oms du ing he illness (Table S2).
Pa ien s ound DENV RNA posi i e a he ime o admission
p o ed mo e likely o be hospi alized han RNA nega i e pa ien s
(Tables S1 and S2). Fu he mo e, a high ela i e amoun o se um
DENV RNA a he ime o p esen ing o hospi al inc eased bo h
he p obabili y and du a ion o hospi aliza ion (Figu e S2, Table
S3).
Associa ions be ween NS1 an igenemia and clinical
pa ame e s. Rela i e le els o NS1 an igen a he ime o
admission co ela ed nega i ely wi h leukocy e and pla ele nadi s,
and posi i ely wi h maximum le els o hemoglobin, hema oc i
and li e ansaminases (AST, ALT) in ollow-up (Figu e S1, Table
S3). O e all NS1 posi i i y a he ime o hospi al admission was
associa ed wi h a highe p obabili y o leukopenia, h ombocy o-
penia, ele a ed ALT, and AST (Tables S1 and S2). The odds
a ios o de eloping leukopenia, h ombocy openia, o ele a ed
ALT, we e a leas wice as high o RNA posi i e as o NS1
posi i e pa ien s (Table S1).
NS1 posi i i y a he ime o admission was ound o be
associa ed wi h a igue and abdominal pain/dia hea o e he
cou se o illness (Table S2).
No associa ion was ound be ween ea ly NS1 posi i i y o
ela i e amoun o se um NS1 and p obabili y/du a ion o
hospi aliza ion (Tables S1, S2, S3, Figu e S2).
Discussion
The need o be e diagnos ics o dengue and p edic ion o
disease ou come is becoming inc easingly u gen [3]. On accoun
o di e ences in pa ien cha ac e is ics, sepa a e da a on endemic
popula ions and a ele s om non-endemic egions a e essen ial.
Ou p e ious s udy sugges ed ha he kine ics o DENV i emia
and NS1 an igenemia may di e in a ele popula ions. This
s udy u he explo es he kine ics o diagnos ic ma ke s, as well as
hei use as combina ions, and co ela ions wi h clinical pa am-
e e s in a el-acqui ed dengue.
Vi ologic Ma ke s in T a ele s’ Dengue
PLOS ONE | www.plosone.o g 4 June 2013 | Volume 8 | Issue 6 | e65900
Kine ics o i emia
The cu en da a sugges a longe span o se um i emia in
a el-acqui ed dengue han ha epo ed in endemic popula ions
[4,5,16,17]. Vi al RNA was de ec able on a e age un il illness day
9; in endemic se ings an a e age du a ion o 5 o 7 days has been
epo ed [4,5,16,17]. This inding migh be ela ed o he high
p opo ion o p ima y in ec ions among a ele s, he clea ance o
i emia being slowe in p ima y han seconda y in ec ions [16,17].
Kine ics o NS1 an igenemia
NS1 an igen was al eady de ec able in he ea lies se um
samples, consis en wi h p e ious s udies in endemic a eas [31–
33]. Howe e , as in i emia, he NS1 posi i i y las ed longe han
ha epo ed in endemic popula ions. This migh be due o slowe
clea ance o he an igen in p ima y in ec ions, o a highe
sensi i i y o he assay in he absence o p e-exis ing an i-NS1
an ibodies. Compa ed o an a e age du a ion o 5 o 7 days
Figu e 1. Kine ics o diagnos ic ma ke s. Rela i e amoun s (A–D) and he p obabili y o a posi i e esul (E–H) o DENV-RNA, NS1 p o ein, and
DENV speci ic IgM and IgG an ibodies in he se um samples o 93 pa ien s wi h acu e dengue e e . Solid lines indica e p edic ed means (A–D), and
p obabili ies (E–H) om gene alized addi i e mixed models (GAMM), and dashed lines hei 95% con idence in e als. In A–D, he ci cles se e o
illus a e indi idual obse a ions; in E–H, he ci cles show he posi i e/nega i e es esul s a each gi en ime poin , he size o he ci cle being
p opo ional o he numbe o obse a ions.
doi:10.1371/jou nal.pone.0065900.g001
Vi ologic Ma ke s in T a ele s’ Dengue
PLOS ONE | www.plosone.o g 5 June 2013 | Volume 8 | Issue 6 | e65900

eco ded in endemic se ings [5,19,31,32], NS1 posi i i y las ed on
a e age un il illness day 15, he span o NS1 an igen de ec ion
being also signi ican ly longe han ha o se um i emia.
O e he i s 12 days o illness, NS1 posi i i y was ound mo e
o en in pa ien s in ec ed wi h DENV-1 han DENV-3, consis en
wi h some p e ious s udies [19,34]. The NS1 EIA es is supposed
o de ec NS1 an igens o all se o ypes equally [31]; a be e
sensi i i y o se o ype 1 has been sugges ed, howe e [17].
Al oge he , he long de ec ion span o NS1, a ailabili y o he es ,
and high speci ici y in compa ison o an ibody es ing [31,35]
make he assessmen o NS1 a use ul ool o diagnosing a ele ’s
dengue.
Diagnos ic co e age
When sc u inizing he i s 3 weeks a e onse , none o he
diagnos ic me hods could alone se e as a su icien diagnos ic ool
o he whole ime pe iod, as expec ed. Combining RNA o NS1
de ec ion and an ibody es ing esul ed in a high o e all sensi i i y,
which acco ds wi h esul s om p e ious s udies [19,21,33,36,37].
In se ings whe e RNA de ec ion is no eadily a ailable, he
combina ion o IgM and NS1 appea s o be a easible app oach.
The ad an ages o NS1 o e RNA de ec ion a e low cos and ease
o es ing. Mo eo e , he long de ec ion span o NS1 an igen
enhances diagnos ic speci ici y in he la e phase o he disease.
Vi ologic ma ke s and clinical pic u e
Dengue in a ele s is conside ed mos ly a mild disease [6], ye
se e e o ms o he disease a e occasionally seen [7]. None o ou
cases we e classi ied as dengue hemo hagic e e o dengue shock
synd ome acco ding o he WHO c i e ia alid a he ime o da a
collec ion. A signi ican p opo ion o he s udy pa ien s we e
hospi alized, howe e ; he decision was aken on he basis o he
clinicians’ e alua ion o he pa ien s’ gene al condi ion. In Finland
i is cus oma y o hospi alize eb ile a ele s whene e he e is any
eason o belie e ha hei clinical s a us may ge wo se, ega dless
o whe he a diagnosis has been es ablished o no . In mos cases
dengue was no diagnosed un il a e he pa ien was admi ed o
hospi al.
Despi e he ac ha he dengue cases we e p edominan ly mild,
eg ession analyses e ealed se e al associa ions be ween he
ela i e le els o i ologic ma ke s (DENV RNA and NS1) and he
a ious clinical pa ame e s s udied. The ini ial le els o i emia
p edic ed bo h likelihood and du a ion o hospi aliza ion. No ably,
he clinicians we e unawa e o hei pa ien s’ i emia (o NS1
an igenemia) s a us a he ime o illness, and he numbe o days
0.0
0.2
0.4
0.6
0.8
1.0
PCR
NS1
ei he
ei he , 95% CI
PCR
IgM
ei he
ei he , 95% CI
0.0
0.2
0.4
0.6
0.8
1.0
PCR
IgG
ei he
ei he , 95% CI
NS1
IgM
ei he
ei he , 95% CI
5 101520
0.0
0.2
0.4
0.6
0.8
1.0
NS1
IgG
ei he
ei he , 95% CI
5 101520
IgM
IgG
ei he
ei he , 95% CI
Days a e onse o illness Days a e onse o illness
P obabili y o posi i e esul
Figu e 2. Kine ics o diagnos ic combina ions. The solid lines and shaded a eas p o ide he p edic ed p obabili ies and hei 95% con idence
in e als o a leas one o he wo diagnos ic es s showing a posi i e esul on a gi en day o illness. Non-solid lines indica e he p edic ed
p obabili ies o posi i e esul o each es alone. The ci cles illus a e ime poin s o samples nega i e in bo h es s, he size o he ci cle being
p opo ional o he numbe o obse a ions.
doi:10.1371/jou nal.pone.0065900.g002
Vi ologic Ma ke s in T a ele s’ Dengue
PLOS ONE | www.plosone.o g 6 June 2013 | Volume 8 | Issue 6 | e65900
om he onse o illness was con olled in he s a is ical analyses; i
no , bo h o hese could ha e had a con ounding e ec on he
associa ions obse ed.
The ini ial le els o NS1 an igenemia we e no associa ed wi h
hospi aliza ion, bu co ela ed posi i ely wi h some cen al
labo a o y pa ame e s, such as hema oc i and li e ansaminases.
No ably, bo h se um i emia and NS1 an igenemia le els
co ela ed nega i ely wi h pla ele coun s o e he cou se o illness.
S udies conduc ed in endemic se ings ha e ela ed high se um
i emia and NS1 an igenemia wi h se e e dengue [13–17]; his
ela ion has no been con i med in some o he s udies, hough
[18,19]. To ou knowledge, co ela ions be ween i ologic and
clinical ma ke s ha e no p e iously been assessed in a ele s. To
s udy hese associa ions, a p ospec i e s udy design wi h
s anda dized da a collec ion would ha e been supe io o he
e ospec i e app oach. Due o he ela i ely low incidence o
dengue among Finnish a ele s, a p ospec i e s udy would ha e
been di icul o conduc in he cu en se ing, hough. One o he
limi a ions o he e ospec i e esea ch app oach was he lack o
comp ehensi e da a on p e-exis ing la i i al immuni y. Howe e ,
o he known po en ial con ounding ac o s (day o illness, and age,
sex, and p esence o co-in ec ions and ch onic diseases) we e
con olled in he analyses explo ing he co ela ions be ween
i ologic and clinical ma ke s. Despi e in ec ions being p edom-
inan ly mild in his s udy, he obse ed co ela ions wi h clinical
pa ame e s (e.g. hema oc i and h ombocy openia) sugges ha
i emia and NS1 an igenemia may se e as p edic o s o he
clinical mani es a ions in a el-acqui ed dengue.
Conclusions
This s udy sugges s a longe a e age du a ion o se um i emia
and NS1 an igenemia in a el-acqui ed dengue han ha
obse ed in endemic a eas, emphasizing he impo ance o
ca ying ou indi idual s udies among di e en pa ien popula-
ions. Ou esul s highligh he impo ance o combining
diagnos ic es s, such as NS1 and IgM de ec ion, o p o ide be e
diagnos ic co e age. The ea ly appea ance and long de ec ion
span o NS1, easy pe o mance and apidi y o he es , as well as
high speci ici y o dengue i uses as compa ed o an ibody es ing
make he assessmen o NS1 a easible ool o diagnosing
a ele ’s dengue. Bo h i emia and NS1 an igenemia showed
co ela ion wi h se e al clinical pa ame e s, sugges ing a po en ial
ole in p edic ing disease ou come in a ele pa ien s.
Suppo ing In o ma ion
Figu e S1 Co ela ions be ween i ologic ma ke s and
labo a o y pa ame e s. Associa ions be ween ela i e
amoun s o DENV RNA and NS1 an igen in he i s se um
sample, and he minimum/maximum alues o hemoglobin,
leukocy es, pla ele s, ALT, and AST du ing ollow-up. Lines
indica e p edic ions om a e aged linea models ( hick line =
s a is ically signi ican , hin line = non-signi ican associa ion),
and open symbols he obse ed da a. In hemoglobin plo s, solid
line and open do s indica e he p edic ions and da a o women,
dashed lines and open squa es hose o men.
(EPS)
Figu e S2 Co ela ions be ween i ologic ma ke s and
hospi aliza ion. Associa ions be ween ela i e amoun s o se um
DENV RNA/NS1 an igen a he ime o p esen ing o hospi al
and p obabili y and leng h o hospi aliza ion. The ci cles indica e
he deg ee o se um i emia/NS1 an igenemia o indi idual
pa ien s a he ime o p esen ing o hospi al and he choice o
ea men place (ou -pa ien /hospi alized)/leng h o hospi aliza-
ion. The size o he ci cles is p esen ed in p opo ion o he
numbe o obse a ions.
(EPS)
Table S1 Odds a ios (OR) and p edic ed occu ence o
abno mal labo a o y alues du ing ollow-up depending
on ini ial RNA o NS1 posi i i y.
(DOCX)
Table S2 Pa ame e es ima es and hei 95% con i-
dence in e als om a e aged model se s explo ing he
connec ion be ween he ini ial DENV RNA/NS1 an igen
posi i i y and he p obabili ies o abno mal clinical
pa ame e s, hospi aliza ion, and di e en symp oms
du ing ollow-up.
(DOCX)
Table S3 Pa ame e es ima es and hei 95% con i-
dence in e als om a e aged model se s explo ing he
connec ion be ween ini ial amoun s o se um DENV
RNA/NS1 an igen and clinical pa ame e s in ollow-up.
(DOCX)
Acknowledgmen s
The au ho s wish o hank Minna Ulmanen and Ki s i Ra¨iha¨ o echnical
assis ance.
Au ho Con ibu ions
Concei ed and designed he expe imen s: EOE EMK LV EH OV AK.
Pe o med he expe imen s: EMK EH. Analyzed he da a: LV.
Con ibu ed eagen s/ma e ials/analysis ools: OV. W o e he pape :
EOE EMK LV EH OV AK. Collec ed clinical da a: EOE AK.
Re e ences
1. WHO. (2012) Dengue and se e e dengue. Fac shee N
o
117, e ised Janua y
2012. Gene a, Wo ld Heal h O ganiza ion.
2. Wilde -Smi h A, Schwa z E. (2005) Dengue in a ele s. N Engl J Med 353:
924–932.
3. Simmons CP, Fa a JJ, Nguyen , Wills B. (2012) Dengue. N Engl J Med 366:
1423–1432.
4. Peeling RW, A sob H, Peleg ino JL, Buchy P, Ca dosa MJ, e al. (2010)
E alua ion o diagnos ic es s: Dengue. Na Re Mic obiol 8: S30–8.
5. WHO. (2009) Dengue: Guidelines o diagnosis, ea men , p e en ion and
con ol – New edi ion. Wo ld Heal h O ganiza ion.
6. Baa en GG, Sonde GJ, Zaaije HL, an Gool T, Kin JA, e al. (2011) T a el-
ela ed dengue i us in ec ion, The Ne he lands, 2006–2007. Eme g In ec Dis
17: 821–828.
7. Wichmann O, Gascon J, Schunk M, Puen e S, Siikamaki H, e al. (2007) Se e e
dengue i us in ec ion in a ele s: Risk ac o s and labo a o y indica o s. J In ec
Dis 195: 1089–1096.
8. Hals ead SB. (1970) Obse a ions ela ed o pa hogenesis o dengue hemo -
hagic e e . VI. Hypo heses and discussion. Yale J Biol Med 42: 350–362.
9. Kho CC, Chau TN, Pang J, Da ila S, Long HT, e al. (2011) Genome-wide
associa ion s udy iden i ies suscep ibili y loci o dengue shock synd ome a
MICB and PLCE1. Na Gene 43: 1139–1141.
10. Ande s KL, Nguye NM, Chau NV, Hung NT, Thuy TT, e al. (2011)
Epidemiological ac o s associa ed wi h dengue shock synd ome and mo ali y in
hospi alized dengue pa ien s in Ho Chi Minh Ci y, Vie nam. Am J T op Med
Hyg 84: 127–134.
11. Rico-Hesse R, Ha ison LM, Salas RA, To a D, Nisalak A, e al. (1997) O igins
o dengue ype 2 i uses associa ed wi h inc eased pa hogenici y in he Ame icas.
Vi ology 230: 244–251.
12. Guzman MG, Hals ead SB, A sob H, Buchy P, Fa a J, e al. (2010) Dengue: A
con inuing global h ea . Na Re Mic obiol 8: S7–16.
13. Vaughn DW, G een S, Kalayana ooj S, Innis BL, Nimmanni ya S, e al. (2000)
Dengue i emia i e , an ibody esponse pa e n, and i us se o ype co ela e
wi h disease se e i y. J In ec Dis 181: 2–9.
14. Lib a y DH, Endy TP, Houng HS, G een S, Kalayana ooj S, e al. (2002)
Di e ing in luences o i us bu den and immune ac i a ion on disease se e i y in
seconda y dengue-3 i us in ec ions. J In ec Dis 185: 1213–1221.
Vi ologic Ma ke s in T a ele s’ Dengue
PLOS ONE | www.plosone.o g 7 June 2013 | Volume 8 | Issue 6 | e65900
15. Lib a y DH, Young PR, Picke ing D, Endy TP, Kalayana ooj S, e al. (2002)
High ci cula ing le els o he dengue i us nons uc u al p o ein NS1 ea ly in
dengue illness co ela e wi h he de elopmen o dengue hemo hagic e e .
J In ec Dis 186: 1165–1168.
16. T icou V, Minh NN, Fa a J, T an HT, Simmons CP. (2011) Kine ics o
i emia and NS1 an igenemia a e shaped by immune s a us and i us se o ype
in adul s wi h dengue. PLoS Negl T op Dis 5: e1309.
17. Duyen HT, Ngoc TV, Ha do T, Hang VT, Kieu NT, e al. (2011) Kine ics o
plasma i emia and soluble nons uc u al p o ein 1 concen a ions in dengue:
Di e en ial e ec s acco ding o se o ype and immune s a us. J In ec Dis 203:
1292–1300.
18. Chau TN, Ande s KL, Lien le B, Hung NT, Hieu LT, e al. (2010) Clinical and
i ological ea u es o dengue in Vie namese in an s. PLoS Negl T op Dis 4:
e657.
19. Duong V, Ly S, Lo n T y P, Tuiskunen A, Ong S, e al. (2011) Clinical and
i ological ac o s in luencing he pe o mance o a NS1 an igen-cap u e assay
and po en ial use as a ma ke o dengue disease se e i y. PLoS Negl T op Dis 5:
e1244.
20. Leh inen VA, Huh amo E, Siikama¨ki H, Vapalah i O. (2008) Japanese
encephali is in a Finnish a ele on a wo-week holiday in Thailand. J Clin
Vi ol 43: 93–95.
21. Huh amo E, Hasu E, Uzca egui NY, E a E, Nikka i S, e al. (2010) Ea ly
diagnosis o dengue in a ele s: Compa ison o a no el eal- ime RT-PCR, NS1
an igen de ec ion and se ology. J Clin Vi ol 47: 49–53.
22. WHO. (1997) Dengue haemo hagic e e : diagnosis, ea men , p e en ion and
con ol, 2
nd
ed. Gene a, Wo ld Heal h O ganiza ion.
23. Vene S, Mangia ico J, Niklasson B. (1995) Indi ec immuno luo escence o
se ological diagnosis o dengue i us in ec ions in Swedish pa ien s. Clin Diagn
Vi ol 4: 43–50.
24. Lancio i RS, Calishe CH, Guble DJ, Chang GJ, Vo ndam AV. (1992) Rapid
de ec ion and yping o dengue i uses om clinical samples by using e e se
ansc ip ase-polyme ase chain eac ion. J Clin Mic obiol 30: 545–551.
25. Huh amo E, Uzca egui NY, Siikamaki H, Saa inen A, Piipa inen H, e al. (2008)
Molecula epidemiology o dengue i us s ains om Finnish a ele s. Eme g
In ec Dis 14: 80–83.
26. Sca amozzino N, C ance JM, Jouan A, DeB iel DA, S oll F, e al. (2001)
Compa ison o la i i us uni e sal p ime pai s and de elopmen o a apid,
highly sensi i e hemines ed e e se ansc ip ion-PCR assay o de ec ion o
la i i uses a ge ed o a conse ed egion o he NS5 gene sequences. J Clin
Mic obiol 39: 1922–1927.
27. R De elopmen Co e Team. (2011) R: A language and en i onmen o
s a is ical compu ing. R ounda ion o s a is ical compu ing. 2.13.0.
28. Wood S. (2011) gamm4: Gene alized addi i e mixed models using mgc and
lme4. R package e sion 0.1–2.
29. Bu nham KP, Ande son DR. (2000) Model selec ion and in e ence : A p ac ical
in o ma ion- heo e ic app oach. New Yo k (NY): Sp inge . 353.
30. Ba on K. (2011) MuMIn: Mul i-model in e ence. R package e sion 1.0.0.
31. Dussa P, Labeau B, Laga hu G, Louis P, Nunes MR, e al. (2006) E alua ion o
an enzyme immunoassay o de ec ion o dengue i us NS1 an igen in human
se um. Clin Vaccine Immunol 13: 1185–1189.
32. Alcon S, Tala min A, Deb uyne M, Falcona A, Deubel V, e al. (2002)
Enzyme-linked immunoso ben assay speci ic o dengue i us ype 1 nons uc-
u al p o ein NS1 e eals ci cula ion o he an igen in he blood du ing he acu e
phase o disease in pa ien s expe iencing p ima y o seconda y in ec ions. J Clin
Mic obiol 40: 376–381.
33. Hu D, Di B, Ding X, Wang Y, Chen Y, e al. (2011) Kine ics o non-s uc u al
p o ein 1, IgM and IgG an ibodies in dengue ype 1 p ima y in ec ion. Vi ol J 8:
47.
34. Fox A, Le NM, Simmons CP, Wolbe s M, We heim HF, e al. (2011)
Immunological and i al de e minan s o dengue se e i y in hospi alized adul s
in Ha Noi, Vie Nam. PLoS Negl T op Dis 5: e967.
35. Hunspe ge EA, Yoksan S, Buchy P, Nguyen VC, Seka an SD, e al. (2009)
E alua ion o comme cially a ailable an i-dengue i us immunoglobulin M es s.
Eme g In ec Dis15: 436–440.
36. F y SR, Meye M, Semple MG, Simmons CP, Seka an SD, e al. (2011) The
diagnos ic sensi i i y o dengue apid es assays is signi ican ly enhanced by
using a combined an igen and an ibody es ing app oach. PLoS Negl T op Dis
5: e1199.
37. Lindeg en G, Vene S, Lundk is A, Falk KI. (2005) Op imized diagnosis o acu e
dengue e e in Swedish a ele s by a combina ion o e e se ansc ip ion-PCR
and immunoglobulin M de ec ion. J Clin Mic obiol 43: 2850–2855.
Vi ologic Ma ke s in T a ele s’ Dengue
PLOS ONE | www.plosone.o g 8 June 2013 | Volume 8 | Issue 6 | e65900