RESEARCH ARTICLEMolecular Diagnosis of Chagas Disease in.docx

RESEARCH ARTICLE
Molecular Diagnosis of Chagas Disease in
Colombia: Parasitic Loads and Discrete
Typing Units in Patients from Acute and
Chronic Phases
CarolinaHernández1,2,ZulmaCucunubá1,2,CarolinaFlórez1,2,M
arioOlivera1,2,
CarlosValencia1,2,PilarZambrano2,CieloLeón1,2,JuanDavidRa
mírez3*
1 RedChagasColombia,InstitutoNacionalde
Salud,Bogotá,Colombia,2 Grupode Parasitología,
InstitutoNacionalde Salud,Bogotá,Colombia,3 Grupode
InvestigacionesMicrobiológicas-UR(GIMUR),
Programa de Biología,Facultad de CienciasNaturalesy
Matemáticas, Universidad el Rosario,Bogotá,
Colombia
* [email protected]
Abstract
Background
The diagnosis of Chagas disease is complex due to the dynamics
of parasitemia in the clini-
cal phases of the disease. The molecular tests have been
consideredpromissorybecause
they detect the parasite in all clinicalphases.Trypanosoma
cruzipresentssignificant
geneticvariabilityand is classified into six DiscreteTyping Units

TcI-TcVI(DTUs)with the
emergenceof foreseen genotypes withinTcI as TcIDomand TcI
Sylvatic. The objective of
this study was to determinethe operatingcharacteristicsof
molecular tests (conventional
and Real Time PCR) for the detectionofT.cruziDNA, parasitic
loads and DTUs in a large
cohortof Colombianpatients fromacute and chronicphases.
Methodology/PrincipalFindings
Sampleswere obtainedfrom708 patients in all clinical phases.
Standarddiagnosis (direct
and serological tests)and molecular tests (conventional PCR and
quantitativePCR) target-
ing the nuclear satelliteDNA region.The genotypingwas
performedby PCR using the
intergenicregionof the mini-exon gene, the 24Sa, 18S and A10
regions.The operating
capabilitiesshowed thatperformanceof qPCR was
highercomparedto cPCR. Likewise,
the performance of qPCR was significantlyhigher in acute phase
comparedwith chronic
phase. The median parasitic loads detectedwere 4.69 and 1.33
parasite equivalents/mLfor
acute and chronicphases. The main DTU identifiedwas TcI
(74.2%).TcIDomgenotype

was significantlymorefrequent in chronicphase comparedto
acute phase (82.1%vs
16.6%).The medianparasitic load for TcIDomwas
significantlyhighercomparedwith TcI
Sylvatic in chronic phase (2.58vs.0.75 parasite equivalents/ml).
PLOS NeglectedTropical Diseases |
DOI:10.1371/journal.pntd.0004997 September 20,2016 1 / 20
a11111
OPEN ACCESS
Citation: Hernández C, Cucunubá Z, Flórez C,
Olivera M, Valencia C, Zambrano P, et al. (2016)
Molecular Diagnosis of Chagas Disease in Colombia:
Parasitic Loads and Discrete Typing Units in Patients
from Acute and Chronic Phases. PLoS Negl Trop Dis
10(9): e0004997.doi:10.1371/journal.pntd.0004997
Editor: Alain Debrabant, US Food and Drug
Administration, UNITED STATES
Received:April 15, 2016
Accepted: August 22, 2016
Published: September 20, 2016
Copyright: © 2016 Hernández et al. This is an open
access article distributedunder the terms of the
Creative Commons AttributionLicense, which permits
unrestricteduse, distribution,and reproduction in any

medium, provided the original author and source are
credited.
Data Availability Statement: All relevant data are
within the paper and its SupportingInformationfiles.
Funding: This work was supportedby Departamento
Administrativo Nacional de Ciencia y Tecnologíade
Colombia ‘‘Francisco José de Caldas –
COLCIENCIAS’’ and ‘‘Unión Temporal Programa
Nacional de Investigación para la prevención,control
y tratamiento integral de la enfermedadde Chagas
en Colombia’’, Grant Number 380- 2011, code 5014-
537-30398.The funders had no role in study design,
data collection and analysis, decision to publish, or
preparationof the manuscript.
http://crossmark.crossref.org/dialog/?doi=10.1371/journal.pntd.
0004997&domain=pdf
http://creativecommons.org/licenses/by/4.0/
Conclusions/Significance
The molecular tests are a precise tool to complement the
standarddiagnosis of Chagas dis-
ease, specifically in acute phase showing high
discriminativepower. However, it is neces-
saryto improve the sensitivityof molecular tests in chronic
phase. The frequencyand
parasitemiaof TcIDomgenotype in chronic patientshighlight its
possible relationshipto the
chronicityof the disease.

Author Summary
Chagas disease is a neglected tropical disease caused by the
parasite Trypanosoma cruzi
that shows tremendous genetic diversity evinced in at least six
Discrete Typing Units and
massive genetic diversity within TcI. Two clinical phases exist
where acute phase shows
high parasitemia and chronic phase shows low and intermittent
parasite dynamics. One
particularity of the disease is the diagnosis, because the
parasitemia is highly variable dur-
ing the phases of the disease. Molecular tests allow detecting
DNA of the parasite in all
clinical phases. Herein, we determined the operating
characteristics of two molecular tests
(cPCR and qPCR) to evaluate the performance of these tests for
diagnosis of Chagas dis-
ease in 708 Colombian patients. We determined the parasitic
loads and DTUs to assess
how is the behaviour of these characteristics in relation to the
clinical phases. We found
that the performance of qPCR was higher compared to cPCR
and the molecular tests are a
precise tool for diagnostic of Chagas disease, mainly in the
acute phase. The parasitemia
was higher in the acute phase compared to chronic phase and
the DTU predominant in
Colombian patients was TcI. The behaviour of TcIDom
genotype in the chronic phase
patients evidenced possible relationship with the chronicity of
the disease.
Introduction
Chagas disease is a zoonotic parasitic disease caused by the

protozoan Trypanosoma cruzi. It is
considered a public health problem in Latin-America, where
approximately 6 million people
are currently infected [1]. The acute phase of the disease is
characterised by usually mild fever
that in a small proportion of cases can be accompanied by
myocarditis and other lethal compli-
cations. Most of the patients continue through the chronic phase
that is initially characterised
by an asymptomatic clinical course during two or three decades,
and about 30% of the infected
patients will develop heart or digestive complications
afterwards [2].
T. cruzi parasite shows significant genetic variability and
classified into at least six Discrete
Typing Units TcI-TcVI (DTUs), that present associations with
the geographical distribution,
epidemiological transmission cycles, insect vectors and clinical
manifestations of Chagas dis-
ease [3–5]. Recent studies suggest the occurrence of an
emerging clade within TcI named TcI-
Dom which is distributed in the Americas and associated with
domestic cycles of transmission
and human infection [6–10]. Recently, a genotype detected in
anthropogenic bats and named
as TcBat has been described in Panama, Ecuador, Colombia and
Brazil including a case of
human infection in Colombia [11–14].
The diagnosis of Chagas disease is complex due to the dynamics
of parasitemia in the phases
of the disease. During the acute phase the parasitemia is high,
therefore the diagnosis is per-
formed by direct parasitological tests [15,16]. Nevertheless,
direct parasitological tests are not

Diagnosis and Genotyping ofT.cruzi in Acuteand ChronicPhases
Competing Interests: The authors have declared
that no competinginterests exist.
useful in the chronic phase due to the low and intermittent
parasitemias. Therefore, the diag-
nosis of Chagas disease in the chronic phase is determined by
serological tests such as ELISA:
enzyme-linked immunosorbent assay, IFA: indirect
immunofluorescence assay or HAI: Hem-
agglutination Inhibition Test [17–19]. Recently, molecular
techniques such as cPCR (conven-
tional PCR) and qPCR (quantitative real-time PCR) have been
considered as supportive
diagnostic tests due to their ability to determine parasitic loads
of T. cruzi in all clinical phases
of the disease [20–22]. The operating characteristics of
molecular tests for diagnosis of T. cruzi
infection have varied according to clinical phase and technical
specifications. Sensitivity for
identifying chronic infection with cPCR has ranged between 22
and 75% [23,24] and in both
cases with a specificity of 100%. Contrastingly, for qPCR,
sensitivity has ranged between 60
and 80% [22,25,26] in chronic phase and between 88% and
100% for acute phase [25,26],
whereas specificity is between 70–100% [26–28]. Sampling
methods have not been always
clearly stated and the role of these techniques for diagnosis of

Chagas disease in the different
clinical phases still remains poorly understood.
The objective of this work was to determine the operating
capabilities of qPCR and cPCR
targeting the satellite nuclear DNA region, compared with
standard diagnosis methods for
acute and chronic Chagas disease. Additionally, we evaluated
the plausible associations
between parasitic load and DTUs in Colombian patients from
the acute and chronic phases to
untangle the natural course of T. cruzi infection in terms of
parasite dynamics.
MaterialsandMethods
Participants
All patients who attended the Colombian National Health
Institute (Overall 985 individuals)
seeking diagnostic tests for Chagas disease in acute (113
patients) or chronic phase (872
patients) between 2004 and 2015 were considered as potential
participants. Inclusion criteria
were: i. Clinical or epidemiological suspect of Chagas disease
in acute or chronic phase ii. Not
having received aetiological treatment for Chagas disease iii.
Positive serological tests for Cha-
gas disease (IFA, ELISA and/or HAI) iv. Adequate blood and
serum samples available for per-
forming diagnostic tests according to the clinical phase. v.
Acceptance to participate and sign
the informed consent.
Ethicalstatement
The Technical Research Committee and Ethics Research Board
at the National Health Institute

in Bogotá, Colombia approved the study protocol CTIN-014-11.
Participation was voluntary
and patients were asked for informed written consent
authorising to take blood and serum
samples and access information on their clinical records.
Samplesize calculationand samplingmethods
The total sample size (N) was calculated for test binary
outcomes and separately for each clini-
cal phase: acute and chronic. Considering, n = Z2 S (1−S) d2,
where for a confidence level of
95% (1- α, with α = 0.05) Z is inserted by 1.96, and a maximum
marginal error of estimate, d, is
a desired value for precision based on researchers judgment,
and S is a pre-determined value of
sensitivity [29]. Based in previous studies, for the acute phase S
was pre-established at 92% and
with d at 8% [25,26], whereas for chronic phase S was pre-
established at 60% with d at 5% [22–
26]. Then, N = n /P, being P the estimated prevalence in this
specific population under study.
Given this is a selected population, composed of patients with
some suspicion of the infection
and remitted to a reference centre, P was specified at 60% in
suspected cases for both acute and
chronic phases. This value was obtained as an approximation
based on the laboratory records
at the NHI (Bogota, Colombia). The minimum total sample sizes

were then calculated as
N = 74 and N = 615 for suspected cases in acute and chronic
phases respectively. The tests
were performed to all subjects without knowing their previous
clinical status. Clinical evalua-
tion was conducted simultaneously to all individuals as part of
the study to determine health
status and then to the confirmed cases to evaluate heart
complications. The inclusion of partici-
pants was conducted retrospectively for the period 2004 to
2012, and prospectively between
2013 and 2015. At the end, a total of 86 suspected acute
patients and 622 suspected chronic
patients were included in the study (Table 1).
Clinicalclassification
Acute phase. a suspected case was defined as an individual with
> 7 days of fever accom-
panied or not by hepatomegaly or splenomegaly. The patient
was considered with acute Chagas
disease if additionally to symptoms tested positive by
parasitological tests (Strout, micro-strout,
blood thick smear, or hemoculture) [15] or presented positive
results to two serological tests
over the course of the following weeks [30,31]. The patients
were classified as negative to Cha-
gas disease otherwise noted.
Chronic phase. individuals without criteria for acute phase but
with clinical or epidemio-
logical suspicion of Chagas disease. The patients were
confirmed as positive T. cruzi infection
when tested positive to two serological tests (IFA, ELISA
and/or HAI). It was then classified as
chronic undetermined (when no evidence of signs or symptoms

of heart complications were
evinced) or chronic determined otherwise.
The risk factors classification was conducted through a survey
applied to each of the patients
included in the study. A series of questions were asked such as
the place of birth, knowledge of
vector insects, blood donations and/or organ transplantation,
housing type and presence of
cardiac symptoms based on previous evaluated questionnaires
(Survey 1) [32]. Patients whose
serological tests were negative were classified into two groups
according to the presence or
absence of risk factors. The patients, who had one or more risk
factors, were categorized as
"negative with risk factors" and those patients that did not have
any risk factors were catego-
rized as "negative without risk factors".
Laboratorytests
Parasitological methods. The direct parasitological methods
were performed (Strout,
micro-strout, blood thick smear, or hemoculture) according to
the methodology described by
Table 1. Generalcharacteristicsofpatientsincludedinthestudy.
General characteristics Acutephaseb N = 86 Chronicphasec N =
622
Positive Negative Positive Negatived
Patientsnumber(N) 708 71 15 481 141
Age,median(Q1-Q3)a 48 (47–49) 31 (26–35) 27 (23–30) 51 (50–

53) 37 (39–41)
Sex, n (%)
Female 428 (60.4) 26 (36.6) 8 (53.3) 313 (65.1) 60 (42.5)
Male 280 (39.6) 45 (63.4) 7 (46.7) 168 (34.9) 81 (52.4)
a Age in years
b Positive patientswere those who had positive
directparasitological tests, symptomatology and/orserological
tests. Negative patientscomprise a group of
febrilepatientswith negative serologyfor Chagas disease and
diagnosedwith dengue.
c Positive patientswere those who had two positive serological
tests and negative patientswere those with two negative
serological tests.
d Twenty-nine were negative without risk factor and 112
negative with risk factor
doi:10.1371/journal.pntd.0004997.t001
Freilij et al., 1983 [33]. The results were considered positive
when morphology compatible with
the T.cruzi was observed. All samples were analysed without
knowledge of the clinical status or
other test.

Serological tests. Enzyme-linked immunosorbent assay
(ELISA), immunofluorescence
antibody assay (IFA) or hemaglutination-inhibition assay (HAI)
were originally standardized
at the National Health Institute [34] with T. cruzi strains
belonging to TcI. All serological tests
were conducted in duplicate and positive and negative controls
were used for each assay.
ELISA test was considered as positive when absorbance was
�0.300, IFA when titres were �1/
32 and HAI when titres were �1/32.) (S1 Appendix). All
samples were analysed without
knowledge of the clinical status or other tests. The
indeterminate results in the serology tests
(ELISA and IFI) were resolved by use of HAI test.
Molecular diagnostic tests. 10 mL of blood samples were
collected and stored with equal
volume of Guanidine Hydrochloride 6M, EDTA 0.2 M buffer,
pH 8.00 (GEB) and subsequently
stored at 8°C. 5mL of serum was frozen at -70°C as described
elsewhere [25,35]. 300 μL aliquots
of GEB were employed and 5μL of IAC plasmid (40pg/μL) were
added as internal control. The
samples were submitted to DNA extraction using the High Pure
PCR Template Roche kit
according to Duffy et al., 2013. Conventional PCR (cPCR) and
multiplex quantitative PCR
(qPCR) for detection of satellite DNA of T. cruzi and IAC
plasmid DNA were performed as
reported elsewhere [23,26]. The qPCR test was considered
positive when the amplification
exceeded the threshold of fluorescence 0.01 and cPCR when
was observed a DNA fragment of
166 bp in the electrophoresis. The positive samples for satellite
nuclear PCR (qPCR and cPCR),

were confirmed by kPCR. Parasitic loads by qPCR were
measured as parasite equivalents per mL
according to Moreira et al., 2013, using a TcI strain as standard
curve (MHOM/CO/01/DA] [22].
All samples were analysed without knowledge of the clinical
status or other tests (S1 Appendix).
DTUs discrimination. PCR was performed using five different
molecular markers aimed
at detecting the six DTUs and the two subdivisions of TcI
previously described by other authors
(TcIDom and TcI sylvatic) as recommended elsewhere [36–41]
(S1 Appendix and S1 Fig).
Statisticalanalysis
Operating characteristics of the molecular tests were estimated
by comparing against standard
diagnosis (described above). Sensitivity, specificity, positive
(+LR) and negative likelihood ratio
(LR-), predictive values (PV), diagnostic precision (DP), Area
under the curve (AUC), and
Kappa index (K) were estimated for each phase of the disease
(acute and chronic), the clinical
stage of chronic patients (determined and undetermined) and
according to DTUs and TcI
genotypes identified (TcI sylvatic/TcIDom) (S2 Appendix).
Results are presented as percent-
ages, with corresponding 95% confidence intervals (95% CI).
Additionally, operational capabil-
ities in chronic patients were calculated in two ways: the first
including negative patients
without risk factors since they are the true negative and the
second including all the negative
patients (with and without risk factors). Due to over dispersion
of parasitic loads, medians and
quartiles are presented. Comparisons are based on Mann-

Whitney test between clinical phases,
chronic clinical stages and the different T. cruzi DTUs and
genotype groups identified. A p
value at <0.05 was considered as statistically significant. All
analysis was performed in Stata:
Data Analysis and Statistical Software version 12.
Results
Generalcharacteristicsof the patients included in the study
Overall, 985 patients were included, 872 suspected of chronic
and 113 of acute infection. Gen-
eral demographic characteristics are shown in Table 1. Out of
the initial potential participants,
http://www.stata.com/
http://www.stata.com/
27 and 129 were excluded for incomplete samples to perform all
analysis from the acute and
chronic groups, respectively and 121 from the chronic group
due to absence of clinical informa-
tion (Fig 1). The inclusion of patients was prospective, whereas
the sample collection was both ret-
rospective (for the period 2004–2011) and prospective (for the
period 2012–2015). This means
that for the retrospective component the samples were part of
the repository. The repository con-
sists of 144 samples, collected between 2004 and 2011, and
corresponds to serum samples stored

at (-80°C). In these samples, serological tests were repeated and
it was found that the results were
the same that they had been reported at the time of collection of
samples and molecular tests were
performed. The prospective component consists of 564 samples,
collected in the period between
2012 and 2015, and maintained in guanidine hydrochloride
solution until processing.
In patients from the acute phase, the qPCR test was positive in
95.7% of the patients and
cPCR in 84.5%. In patients from the undetermined chronic
phase, qPCR was positive in 68.0%
of the cases and in 55.4% by cPCR. In the cardiac chronic
phase, qPCR positivity was 59.1%
and 58.6% by cPCR. The positive samples for satellite nuclear
PCR (qPCR and cPCR), were
confirmed by kPCR. In patients that were negative by serology
but with risk factors cPCR
(2.6%) and qPCR (3.6%) were positive. In febrile and negative
patients without risk factors
both tests were negative in all samples. In all samples analyzed
we detected the internal amplifi-
cation control for both cPCR and qPCR, the average Ct value in
all samples tested was 21.
Fig1. Algorithmfor
selectionandclassificationofpatients.Therewere selected 708
patients,71 in acutephase, 15 febrilenegatives, 481 in chronic
phaseand 141 negatives. *RF:Risk Factor.
doi:10.1371/journal.pntd.0004997.g001

Operatingcharacteristicsof molecularmethodsvs
standarddiagnostic
tests
The operating characteristics including all negatives patients of
chronic phase (Negatives with
and without risk factors) are presented in Tables 2 and 3.
Performance of qPCR was higher compared to cPCR in both
acute (AUC 0.98 vs 0.92) and
chronic phase including only negatives with risk factors (0.82
vs 0.78) (Fig 2). Likewise, the per-
formance was significantly higher in acute compared with
chronic phase and in overall a speci-
ficity higher than sensitivity particularly in chronic phase
(Tables 2, 4 and 5).
T.cruziparasitic loads and clinicalphase
Parasitic loads were determined in samples that tested positive
by qPCR. Significantly different
median values were detected in acute (4.69 parasite
equivalents/mL) versus chronic phase (1.33
parasite equivalents/mL). A statistically median difference was
also found between determined
and undetermined chronic phase (Fig 3).
Table 3.
Operatingcharacteristicsofmoleculartestsinchronicphases(undete
rminedanddetermined)ofChagasdisease includingallnegatives
patients(withandwithoutriskfactors).
Operatingcharacteristics ChronicundeterminedphaseN = 278/419
Chronicdetermined phaseN = 203/344

qPCR(95%CI) cPCR (95%CI) qPCR(95%CI) cPCR(95%CI)
Sensitivity 67.9 (62.3–73.1) 55.4 (49.5–61.1) 59.1 (52.2–65.6)
58.6 (51.7–65.1)
Specificity 97.2 (92.9–98.8) 97.9 (93.9–99.2) 97.2 (92.9–98.8)
97.9 (93.9–99.2)
PPV 97.9 (94.8–99.2) 98.0 (94.5–99.3) 96.7 (92.0–98.7) 97.5
(93.0–99.1)
NPV 60.6 (54.1–66.7) 52.7 (46.6–58.6) 62.2 (55.7–68.4) 62.1
(55.6–68.2)
DP 77.8 (73.6,81.5) 69.6 (65.1–73.9) 74.7 (69.86–79.0) 74.7
(69.9–79.0)
LR+ 24.0 (14.6–39.3) 26.0 (13.4–50.5) 20.8 (12.6–34.4) 27.5
(14.2–53.5)
LR- 0.33 (0.32–0.33) 0.45 (0.44–0.46) 0.42 (0.41–0.43) 0.42
(0.41–0.43)
K 0.57 (0.48–0.65) 0.44 (0.36–0.52) 0.52 (0.42–0.61) 0.51
(0.42–0.61)
AUC 0.83 (0.79–0.86) 0.77 (0.72–0.80) 0.77 7(0.73–0.82) 0.77
(0.73–0.82)
PPV: Positive predictive value; NPV: Negative predictive value;
DP: diagnostic precision; LR+: positive likelihood ratio;LR-:
negative likelihood ratio. N =
(Positive gold standard/ total assayed).

Table 2.
Operatingcharacteristicsofmoleculartestsinacuteandchronicphase
sincludingallnegativepatients(withandwithoutriskfactors).
Operatingcharacteristics AcutephaseN = (71/86) ChronicphaseN
= (481/622)
qPCR(95%CI) cPCR(95%CI) qPCR(95%CI) cPCR(95%CI)
Sensitivity 95.7 (88.3–98.5) 84.5 (74.3–91.2) 64.2 (59.8–68.4)
56.8 (52.3–61.1)
Specificity 100.0(79.6–100.0) 100.0(79.6–100.0) 97.1 (92.9–
98.8) 97.9 (93.9–99.2)
PPV 100.0(94.6–100.0) 100.0(93.9–100.0) 98.7 (96.7–99.5)
98.9(96.8–99.6)
NPV 83.3 (60.8–94.2) 57.69(38.9–74.5) 44.3 (38.9–49.9) 39.8
(34.9–45.1)
DP 96.5 (90.2–98.8) 87.2 (78.5–92.8) 71.7 (68.0–75.1) 66.0
(62.3–69.6)
LR+ Undefined 22.6 (13.82–37.09) 26.68(13.8–51.5)
LR- 0.04 (0.02–0.1) 0.15 (0.13–0.18) 0.37 (0.36–0.37) 0.44
(0.43–0.45)
Kappaindex 0.89 (0.7–1.1) 0.62 (0.5–0.9) 0.43 (0.36–0.49) 0.36
(0.29–0.42)
AUC 0.98 (0.91–0.99) 0.92 (0.88–0.96) 0.81 (0.77–0.84) 0.77
(0.74–0.81)

negative likelihood ratio. N =
(Positive gold standard/ total assayed). When the specificity is
100% the positive likelihood ratio is undefined.
Fig2. ROC
curvesofmoleculartestsinclinicalphasesofChagasdisease.A.
AcutephaseB. ChronicphaseC. Chronicundetermined phase
D. Chronicdeterminedphase.
Molecular testsperformanceaccordingtoT.cruziDTUs and
clinical
phase
In samples that tested positive (n = 407) by cPCR, the DTUs
TcI-TcVI and TcI (TcI Dom, TcI

Sylvatic) were evaluated. The distribution of DTUs was 74.2%
for TcI, 17.2% for TcII, 1.48%
for TcIII, 0.5% for TcV and 6.7% for mixed infections. For the
latter seven different combina-
tions were identified: TcIDom/TcII/TcV, TcIDom/TcII,
TcIDom/TcISylv , TcIDom/TcISylv/
TcII, TcIDom/TcISylv/TcIII, TcIDom/TcIV, TcISylv/TcII. With
respect to TcI, the genotyping
was feasible in 290/302 samples. Out of them, 28.7% were
classified as TcI Sylvatic and 71.4%
as TcIDom. The median load parasitic value for TcII (4.68
parasite equivalents/mL) was signif-
icantly different to the one for TcI (2.87 parasite
equivalents/mL) and TcIII (1.72 parasite
equivalents/mL) (Fig 4).
The genotype distribution according to clinical phase evidenced
that TcIDom was sig-
nificantly more frequent in chronic phase compared with acute
phase (Table 6). The oper-
ating characteristics of molecular tests for the different
genotypes were calculated,
observing that the sensitivity for identifying TcII was slightly
higher than for TcI, mainly
for qPCR (S1 Table). The median parasitic load for TcIDom
was significantly higher (2.58
parasite equivalents/ml) compared with TcI Sylvatic (0.76
parasite equivalents/ml) in
chronic phase (Fig 5).
Discussion
Operatingcharacteristicsof molecularmethodsagainststandard
diagnostic tests
The main limitation involved in this study is the fact that there
is not a gold standard test for

all clinical phases of Chagas disease. Particularly for chronic
phase, the best comparators are
serological tests but these techniques measure the immune
response and not the relative pres-
ence of the parasite. This particular situation impacts the
evaluation of new diagnostic tests.
This is reflected mainly in the kappa index (Tables 2 and 4) that
presented very low values in
the undetermined and determined chronic phases. Unfortunately,
it has not a simple solution
and more understanding of the course of the infection is still
needed.
Table 4. Operating characteristicsofmolecular
testsinchronicphaseofChagasdiseaseincluding
onlynegativeswithoutriskfactors.
Operatingcharacteristics ChronicphaseN = (481/510)
qPCR(95%CI) cPCR(95%CI)
Sensitivity 64.2 (59.9–58.4) 56.8 (52.3–61.1)
Specificity 100.0(88.3–100.0) 100 (88.3–100.0)
PPV 100.0(98.8–100.0) 100.0(98.6–100.0)
NPV 14.4 (10.2–19.9) 12.2 (8.7–17.0)
DP 66.3(62.1–70.2) 59.2 (54.9–63.4)
LR+ Undefined
LR- 0.36 (0.35–0.36) 0.43 (0.42–0.44)
Kappaindex 0.17(0.1213–0.218) 0.13 (0.09–0.17)

DP: diagnosticprecision; LR+: positive
likelihood ratio;LR-: negative likelihood ratio. When the
specificity is 100% the positive likelihood ratio is
undefined. N = (Positive gold standard/ totalassayed)
Table 5.
Operatingcharacteristicsofmoleculartestsinchronicphases(undete
rminedanddetermined)ofChagasdisease includingonlynega-
tiveswithoutriskfactors.
Operatingcharacteristics ChronicundeterminedphaseN = 278/307
Chronicdetermined phaseN = 203/232
qPCR(95%CI) cPCR(95%CI) qPCR(95%CI) cPCR(95%CI)
Sensitivity 68.0 (62.3–73.2) 55.4 (49.5–61.1) 59.1 (52.2–65.7)
58.6 (51.8–65.2)
Specificity 100 (88.3–100) 100.0(88.3–100.0) 100 (88.3–100)
100.0(88.3–100.0)
PPV 100 (98.0–100) 100.0(97.6–100.0) 100 (96.9–100)
100.0(96.8–100.0)

NPV 24.6 (17.7–37.1) 18.9 (13.5–25.9) 25.9 (18.7–34.7) 25.6
(18.5–34.4)
DP 71.0 (65.7–75.8) 59.6 (54.0–64.9) 64.2 (57.9–70.1) 63.6
(57.2–69.6)
LR+ Undefined
LR- 0.32 (0.31–0.33) 0.44 (0.43–0.45) 0.41 (0.39–0.41) 0.41
(0.40–0.43)
K 0.29 (0.21–0.36) 0.19(0.12–0.25) 0.26(0.17–0.35) 0.26 (0.17–
0.34)
negative likelihood ratio. When
the specificity is 100%the positive likelihood ratio is undefined.
Fig3. Comparativeanalysisofparasitic
loadsinpatientswithChagasdisease.Distributionof parasitic load
and medianson the
basis of theclinicalphases. Theoutlierswere removed fromthe
graph for convenience. * p < 0.05 ** p<0.01 *** p < 0.001.

loadsforDTUs.Distributionof parasitic load and medianson
thebasis
of theT.cruziDTUs. For convenience the outlierswere removed
for the graph. * p < 0.05 ** p<0.01 *** p < 0.001.
Table 6.
FrequencyofDTUsandTcIgenotypesfromclinicalphasesofChagasd
iseasepatients.
DTUs Clinicalphases P value
Acute Chronic
N = 68 N = 332
N (%) (95%CI) n (%) (95%CI)
TcI 54 79.4 69.3 88.9 241 72.6 67.8 77.4 0.25
TcII 6 9.0 2.0 15.9 64 19.3 15.0 23.5 0.04
TcIII 4 6.0 0.2 11.7 2 0.6 0.2 1.4 <0.001
TcV - - - - 2 0.6 0.2 1.4 0.53
Mixed 4 6.0 0.2 11.7 23 6.9 4.2 9.7 0.15
TcI Genotypes N = 48 N = 235
TcI Sylvatic 40 85.1 74.8 95.4 42 17.9 12.9 22.8 <0.001
TcI Dom 8 16.6 4.6 25.2 193 82.1 77.2 87.0 <0.001

DTU: Discrete Unit Typing; bold text: p value at <0.05
The results obtained for the molecular diagnosis in acute phase
were optimal in terms of
sensitivity for both qPCR (95.7%; 95%CI: 88.3–98.5) and cPCR
sensitivity (84.5%; 95%CI:
74.3–91.2), and same specificity. Although the results are
showing a potential superior perfor-
mance of the sensitivity of qPCR compared with cPCR, this
difference needs a cautious inter-
pretation. This might be explained due to the fact that detection
by qPCR increases the
sensitivity and specificity because of the hybridization of the
Taqman probe in the amplicon,
whereas in the case of the cPCR it requires a considerable
amount of amplicon so that it can be
observed in agarose gels [25,26] In addition, the confidence
intervals were slightly overlapped,
meaning that there is some indication of this difference but it is
not statistically significant, so
not definitive. The performance of the molecular tests in the
acute phase is explained because
there are large numbers of parasites, for example in cases of
reactivation in immunosuppressed
patients and in oral outbreaks. The values obtained for LR
evinced the high probability that
positive results correspond to diseased patients (LR+) and the

low probability that the diseased
patients present negative results (LR-). In addition, the DP was
very optimal specifically for
qPCR test confirming that this molecular test is very useful for
the diagnosis in the acute phase,
considering that the direct diagnosis is complex when the
parasitemia is low (As is the case of
the acute patients detected more than a month after the infection
where the parasitemia nor-
mally begins to decrease due to the control of the immune
response) and are required many
tests for the confirmation of the acute cases (direct tests,
serology tests and clinical informa-
tion). Regarding the predictive power of molecular tests in the
acute phase, these tests are very
good predictors of the disease presence when positive results
are obtained (PPV) but their per-
formance as predictors of absence of the disease are less (NPV).
However, it is worth noting
that the predictive values depend on disease prevalence in the
evaluated population.
The analysis of operational capabilities in the chronic phase was
conducted in the first
instance including only negative patients without risk factors or
true negatives. For the chronic
loadsforTcIGenotypesintheclinicalphases.Distribution of
parasitic load and medianson the basis of theTcIgenotypesin
theacuteand chronicphases. For convenience the
outlierswere removed for thegraph. * p < 0.05 ** p<0.01 *** p
< 0.001.

phase, qPCR sensitivity was 64.2% and 56.8% for cPCR and in
concordance with previous
reports obtained by qPCR that have shown sensitivity ranging
from 60–80% and 20–70% for
cPCR [22–24,26,28,42]. These sensitivity results may be due to
low and intermittent parasitic
loads during chronic phase.
The performance of qPCR was better than cPCR in the chronic
undetermined phase, while
that was very similar between the two tests in the determined
chronic phase (Tables 3 and 5).
The discriminative power of the two molecular tests was
acceptable in the chronic phase. For
qPCR, the AUC and DP values obtained (Tables 3 and 5) were
better for the undetermined
phase than for determined phase. The differences between
undetermined and determined
phases for qPCR of the chronic phase can be explained by the
natural course of the disease, in
which the parasitic load decreases while increases the infection
time. This is supported by sev-
eral studies showing that there is no relationship between the
evolution of the cardiac form of
the disease and parasitemia but it declines with time as
observed in this study [43,44]. Also,
some studies show that cardiac form is mainly related to
different types of strains, increased
parasitemia, reinfection or immune system disorders in chronic

patients [45,46]. In the cPCR
AUC values were the same for both phases, while the value of
DP was best for the determined
phase. Possibly, this is because the detection limit of the cPCR
is lower than qPCR, for this rea-
son the cPCR behaves similarly in the two phases. In the two
stages of the chronic phase, there
is a high probability that patients with negative results in the
molecular tests have the disease
(LR-) and these tests are not good predictors of the absence of
the disease (NPV) (Table 5).
Therefore, the use of molecular methods as diagnostic tests is
not appropriate due to the better
performance displayed by serology. The probability that the
results are positive is high in dis-
eased individuals with respect to healthy individuals (LR +) and
the molecular tests are excel-
lent predictors of the presence of disease (PPV). Thus, these
tests could be used in situations in
which the diagnosis is doubtful, allowing the confirmation of
the parasite in diseased patients,
which is of great importance for example when monitoring
etiological treatment. However, it
is necessary to improve the sensitivity, which can be performed
by analysing serial samples for
each patient as seen in some studies in which such sensitivity
improved from 69.2% to 85.2%
with the addition of a second sample or conducting DNA
extraction from a larger volume of
the sample [47,48].
In addition, the operating capabilities of patients in chronic
phase were calculated including
all negatives by serology with and without risk factors (Table 1,
N = 141). It was observed in
the group of negative patients with risk factors a positivity of

2.6% (3 patients) by cPCR and
3.6% (4 patients) by qPCR, possibly due to an
immunosuppression issue in these patients pre-
venting the detection of antibodies or infection. Three patients
are from the department of
Casanare, which is an endemic area, and five patients had less
than 24 years of age suggesting a
recent infection. Also, all patients reported to know the vectors
and have lived during his/her
childhood in homes with features such as thatched or
‘barheque’, floor or wood and/or tread
walls of earth, wood or ‘barheque’. Two of the seven patients
that were negative by serology
and had risk factors, whose ages were 36 and 51 showed the
presence of symptoms at cardiac
level. In this group of 7 patients, 4 presented the ELISA
absorbance values greater than 0.200
and 4 detectable titles in the IFA (1/8 and 1/16).
As the operating capabilities calculated including all negative
patients, a small percentage of
decreased specificity in the two platforms was observed (S3
Appendix). The positivity of these
serologically negative patients that generated the decrease can
probably be explained because
cases of recent infection or patients with some form of
immunosuppression that has generated
the absence of detectable antibodies. In fact, in the group of
acute patients, 4 patients whose
serology was negative showed positive PCR, in these patients
the detection was achieved by
direct parasitological methods. Regarding the molecular
techniques, given that in all PCR runs

were included negative controls including reagents controls, a
plausible contamination with
parasite DNA is discarded. Significantly, the DP and AUC
values showed no obvious changes
unlike the values obtained for the NPV and the Kappa index, in
which there was a marked
increase. However, the changes obtained do not change the
interpretation of the usefulness of
the test in the clinical setting, but can show that there are few
cases where serological tests may
have false negatives as noted previously using cPCR by
Ramirez et al., 2009 [23]. Even though
serological tests are considered the best current option for the
diagnosis of Chagas disease, in a
meta-analysis of high quality tests their sensitivity has been
estimated at 90% [49]. Given this,
we believe that an improvement of diagnostic tests for Chagas
disease is needed for both serol-
ogy and PCR techniques. An appropriate use of the comparator
as gold standard and the inclu-
sion of different phases of the disease are crucial to understand
the utility of different
diagnostic tests.
To our knowledge, this is the first study to include statistical
calculation of the sample,
which allowed the analysis of operating characteristics of the
molecular tests in all clinical
phases of Chagas disease. In addition, this study is the first in
analysing the two PCR platforms
(qPCR and PCR) for the same target (stDNA) in patients from

all clinical phases of Chagas dis-
ease. The conventional technique was included, due to the vast
use of this technique in the
diagnosis and its ease implementation in laboratories with
restricted equipment (a Real Time
PCR machine is not available) [23,24,28]. Lastly, acute patients
had a less median age than
chronic phase patients and in turn the largest number of acute
cases are male. This possibly is
because economic activity in endemic areas is developed by
males that assist to the field and
this facilitates direct patient contact with the vector and
therefore with the parasite. On the
other hand, females ratio and the median age were higher in
chronic phase patients that are
usually detected by screening blood banks or present cardiac
abnormalities in chronic phase,
then the detection occurs at a greater age. Additionally, in
Colombia most blood donors are
women facilitating their diagnosis.
T.cruziparasitemia,DTUs and clinicalphases
Regarding the parasitemia, it is observed that the median
parasitemia was higher in acute
patients compared to chronic phase, which is expected given the
dynamics of parasitemia in
the disease [25,26]. As for the group of chronic patients, the
herein reported median of parasi-
temia is similar to those previously reported for Colombia
[22,26]. In addition, the difference
in medians between cardiac chronic and undetermined chronic
stages was statistically signifi-
cant, being higher in the undetermined chronic phase unlike the
findings described by Ramirez
et al, 2015 [26], in which statistically significant difference was
not detected. However, our

results are in accordance with the natural history of the disease
where parasitic loads decrease
with the chronicity of the infection and this is probably
associated with the type of strain and/
or the immune response [2].
The DTU with highest frequency was TcI, both in acute and
chronic patients, consistent
with findings previously reported in Colombia [8,39,50,51].
Followed by TcII most often
detected in chronic than acute patients. These findings are
congruent due to the predominance
of TcII in domestic cycles of transmission for the case of
Colombia [50]. Regarding the parasitic
loads of the DTUs detected, we observed that TcII had higher
median parasitemia than other
DTUs, consistent with the number of copies that has been
reported in the DNA nuclear satel-
lite region being higher for TcII than for TcI [52–54]. These
findings highlight the importance
of using the most representative DTU to generate the standard
curves for quantification
[22,25,26]. In addition, in murine models TcII shows higher
parasitemias than TcI when per-
forming individual and mixed infections [55].
In this study, acute cases are likely caused by vector
transmission and possible oral route. In
most of the cases TcI (TcI sylvatic), TcII and TcIII infection

was observed. These findings are
consistent with previously documented reports for acute patients
where DTUs associated with
the sylvatic cycle of transmission were depicted [4,5,40,51,56–
61]. An interesting finding was
the detection of TcV in the patients surveyed. This DTU has
been already reported in dogs and
Rhodnius prolixus from eastern Colombia but this would be the
first report of TcV human
infection in the country [50]. It is necessary to conduct further
studies to understand the host-
parasite associations of this foreseen DTU in patients from
northern areas of the continent. It
is well known that TcV infection is endemic in Bolivia, Brazil
and Argentina but in Colombia
is a novel case that requires further investigation; in fact high-
resolution markers have been
applied to the few isolates of Colombian TcV showing a tailored
hybrid profile suggesting a
Pan-American import from south America [62]. The DTU TcVI,
is mainly detected in the
South Cone of Latin America. Normally associated with
megavisceral syndromes and some
cases of congenital heart disease [4]. In Colombia, TcVI has
been very rare and almost infre-
quent. In fact it is limited to a report in which was detected in
humans and R. prolixus isolates
(4% and 1.4% respectively). In addition, in different studies
with a considerable number of
patients conducted in Colombia it was not detected, confirming
the low prevalence of the DTU
in the country [39,51,63].
Recently, it has been highlighted the emergence of a genotype
named as TcIDom and associ-
ated to human infection and domestic transmission cycles via

different molecular markers
[5,6,8,64–66]. Other studies have shown the presence of TcI
Sylvatic genotype in tissue and TcI-
Dom in bloodstream of patients with Chagas cardiomyopathy
[41]. In murine models was
observed that TcIDom produced high parasitemia and low tissue
invasion, a process that allows
an adaptation to the host prolonging its permanence and likely
generation of chronicity, opposite
process to what happened with the TcI sylvatic strains [67]. In
accordance with these previous
findings, our results show that in chronic patients the frequency
and parasitemia of TcIDom geno-
type were significantly higher in chronic patients than in acute
patients, supporting the hypothesis
that this genotype may be related to chronicity in patients with
Chagas cardiomyopathy.
In conclusion, the molecular diagnostic tests are becoming a
precise tool to complement the
standard diagnostic methods for Chagas disease. This study
shows that in general qPCR has a
better performance than cPCR. Also, the results confirm that
PCR is highly specific for both
acute and chronic clinical phases, whereas sensitivity is
acceptable for acute phase but still very
low for chronic patients. This situation could be partially
explained by the higher parasitic
loads detected in acute phase and the intermittent nature of the
parasite release to the blood-
stream in chronic phase. We explored for the first time in a
large cohort of Chagas disease
patients the DTU parasitemia and the natural course of
infection. This type of studies is
required in Latin-America for a better understanding of disease
progression and molecular epi-

demiology of Chagas disease. This makes PCR a potential tool
for its use in acute phase diagno-
sis in a routine basis, and potentially for determining
aetiological treatment failure when tests
positive but not substantially useful when tests negative and
these results must be interpreted
cautiously as in the clinical trials previously published [21,68].
Further research is needed to
improve the sensitivity of this test and the mandatory
deployment of new diagnostic tests.
SupportingInformation
S1 Fig. Algorithm for genotyping of T. cruzi DTUs. Molecular
characterization of T.cruzi by
five molecular markers and genotyping of TcI DTU in two
genotypes TcI Dom and �TcI Sylv:
TcI Sylvatic.
(JPG)
http://journals.plos.org/plosntds/article/asset?unique&id=info:d
oi/10.1371/journal.pntd.0004997.s001
S1 Table. Operating characteristics of molecular test for DTUs
and genotypes TcI.
(DOC)
S1 Appendix. Methodology of reference tests employed in the
study.
(DOCX)

S2 Appendix. Detailed methodologyof the molecular tools
employed in the study.
(DOC)
S3 Appendix. Sensitivity and specificity calculations.
(DOC)
S4 Appendix. STARD. Standards for Reporting of Diagnostic
Accuracy Checklist.
(DOCX)
S5 Appendix. Research protocol. Project: Characterization of a
cohort of patients with Chagas
disease, its etiological treatment, adverse events and therapeutic
response.
(PDF)
Acknowledgments
Parasitology group and Red Chagas Colombia, National
Institute of Health, Colombia.
AuthorContributions
Conceived and designed the experiments: CH ZC JDR.
Performed the experiments: CH CF CL.
Analyzed the data: CH ZC MO CV PZ JDR.
Contributed reagents/materials/analysis tools: ZC JDR.
Wrote the paper: CH ZC JDR.
References
1. WHO. Weekly epidemiological record—Chagas diseases in
LatinAmerica:an epidemiological update

basedon 2010estimates.2015p. 5–13.Available:
http://www.who.int/wer/2015/wer9006.pdf
2. BernC. Chagas’Disease. N Engl J Med.2015;373(5):456–66.
Available: doi:http://www.nejm.org/doi/
abs/10.1056/NEJMra1410150PMID:26222561
3. ZingalesB, Andrade SG, Briones MRS, Campbell DA,
ChiariE, Fernandes O, et al. A new consensus
for Trypanosoma cruzi intraspecific nomenclature:
Secondrevision meetingrecommendsTcI to TcVI.
MemInst Oswaldo Cruz.2009;104(7):1051–4.PMID:20027478
4. ZingalesB, MilesMA,Campbell DA,
TibayrencM,MacedoAM,Teixeira MMG,et al.The revised Try-
panosomacruzisubspecific nomenclature:Rationale,
epidemiological relevance and researchapplica-
tions. Infect GenetEvol. 2012;12(2):240–53. Available:
doi:http://dx.doi.org/10.1016/j.meegid.2011.
12.009PMID:22226704
5. MessengerLA, MilesMA,BernC. Betweena bug and a
hardplace:Trypanosomacruzigeneticdiver-
sity and theclinicaloutcomesof Chagasdisease. ExpertRev Anti
Infect Ther. 2015;13(8):995–1029.
Available: doi:
http://informahealthcare.com/doi/abs/10.1586/14787210.2015.1
056158 PMID:
26162928
6. Llewellyn MS, MilesMA,CarrascoHJ, Lewis MD, Yeo M,
Vargas J, et al.Genome-Scale Multilocus
Microsatellite Typing of Trypanosoma cruziDiscreteTyping
Unit I Reveals Phylogeographic Structure
and SpecificGenotypes Linked to HumanInfection.PLoS Pathog.

2009;5(5):e1000410. Available: doi:
http://dx.plos.org/10.1371/journal.ppat.1000410
PMID:19412340
7. Zumaya-EstradaF, MessengerL, Lopez-Ordonez T, Lewis
MD, Flores-Lopez C, Martínez-IbarraAJ,
et al. NorthAmerican import?Chartingtheoriginsof an
enigmaticTrypanosoma cruzidomesticgeno-
type. Parasit Vectors. 2012;5(1):226.
oi/10.1371/journal.pntd.0004997.s002
oi/10.1371/journal.pntd.0004997.s003
oi/10.1371/journal.pntd.0004997.s004
oi/10.1371/journal.pntd.0004997.s005
oi/10.1371/journal.pntd.0004997.s006
oi/10.1371/journal.pntd.0004997.s007
http://www.who.int/wer/2015/wer9006.pdf
http://www.nejm.org/doi/abs/10.1056/NEJMra1410150
http://www.nejm.org/doi/abs/10.1056/NEJMra1410150
http://www.ncbi.nlm.nih.gov/pubmed/26222561
http://dx.doi.org/10.1016/j.meegid.2011.12.009
http://informahealthcare.com/doi/abs/10.1586/14787210.2015.1

056158
http://dx.plos.org/10.1371/journal.ppat.1000410
8. RamírezJD, Tapia-calle G, GuhlF. Geneticstructureof
Trypanosomacruzi in Colombia revealed by a
High-throughputNuclearMultilocus Sequence Typing
(nMLST)approach. BMC Genet.2013;14:2–10.
9. Segovia M,CarrascoHJ, MartínezCE, MessengerLA,Nessi A,
Londoño JC, et al.Epidemiologic
SourceTracking of OrallyTransmittedChagasDisease,Venezuela.
Epidemiol SourceTrack Chagas
Dis. 2013;19(7):1098–101.
10. LeónCM,Hernández C, MontillaM,RamírezJD. Retrospective
distribution of Trypanosoma cruzi I
genotypesin Colombia.MemInstOswaldo Cruz.2015;110(3):387–
93. Available: http://www.scielo.br/
scielo.php?script=sci_arttext&pid=S0074-
02762015000300387&lng=en&nrm=iso&tlng=en doi:10.
1590/0074-02760140402 PMID:25946157
11. RamírezJD, Tapia-Calle G, Muñoz-CruzG,Poveda C,
RendónLM,HincapiéE, et al.Trypanosome
species in neo-tropical bats:Biological,evolutionaryand
epidemiological implications. Infect Genet
Evol. 2014;22:250–6. doi:
10.1016/j.meegid.2013.06.022PMID:23831017
12. Marcilia, LimaL, Cavazzana M,Junqueira a C V, Veludo
HH,MaiaDa Silva F, et al. A new genotypeof
Trypanosoma cruziassociated withbats evidenced by
phylogeneticanalysesusingSSU rDNA,cyto-

chromeb and HistoneH2B genesand genotypingbasedon
ITS1rDNA.Parasitology. 2009;136
(6):641–55.doi:10.1017/S0031182009005861PMID:19368741
13. LimaL, Espinosa-Álvarez O, OrtizPA, Trejo-Varón JA,
CarranzaJC, PintoCM,et al. Geneticdiversity
of Trypanosoma cruzi in bats, and multilocus phylogeneticand
phylogeographical analysessupporting
Tcbatas an independent DTU(discretetypingunit).Acta Trop.
2015;151(1):166–77.Available: http://
dx.doi.org/10.1016/j.actatropica.2015.07.015
14. RamírezJ, HernándezC, MontillaM,ZambranoP, FlórezAC,
Parra E, et al.FirstReportof HumanTry-
panosomacruzi InfectionAttributedto TcBatGenotype. Zoonoses
Public Health.2013;1–3.
15. Feilij H, MullerL,
GonzalezCappaSMM,CappaSMG.DirectMicromethod for
Diagnosis of Acuteand
Congenital Chagas ‘ Disease. J ClinMicrobiol.1983Aug;
18(2):327–30. PMID:6413530
16. FloresMA,Trejos A, Paredes AR. Strout’s
concentrationmethodin the diagnosisof acuteChagas dis-
ease. Bol ChilParasitol. 1966;21:38–9.PMID:4958150
17. Da Silveira JF, Umezawa ES, LuquettiAO. Chagasdisease:
recombinantTrypanosoma cruziantigens
for serologicaldiagnosis.Trends Parasitol.2001;17(6):286–91.
PMID:11378036
18. Umezawa ES, BastosSF, CouraJR, Levin MJ, GonzalezA,
Rangel-Aldao R, et al. An improved sero-
diagnostic test for Chagas’diseaseemployinga mixtureof
Trypanosomacruzi recombinant antigens.

Transfusion.2003;43(1):91–
7.Available:http://www.ncbi.nlm.nih.gov/pubmed/12519436PMI
D:
12519436
19. BrittoC, CardosoMA,Vanni C, Hasslocher-MorenoA, Xavier
SS, Oelemann W, et al.Polymerase
chainreactiondetectionof Trypanosoma cruzi in humanblood
samplesas a tool for diagnosisand
treatmentevaluation. Parasitology.1995;110(3):241–7.
20. ISGlobal-CRESIB.VIII tallersobre la enfermedad de Chagas
importada:Avances en el tratamiento
antiparasitario. 2012;30–35.
Available:https://issuu.com/isglobal/docs/viii_taller_de_chagas
_2012/4
21. MorilloCA, Marin-NetoJA, Avezum A, Sosa-Estani S,
RassiA, RosasF, et al. Randomized Trial of
Benznidazole for ChronicChagas’ Cardiomyopathy. N Engl J
Med.2015;373(14):1–12.Available:
http://www.nejm.org/doi/full/10.1056/NEJMoa1507574#t=articl
e
22. MoreiraOC, RamírezJD, Velázquez E, MeloMF, Lima-
FerreiraC, GuhlF, et al. Towards theestablish-
mentof a consensus real-time qPCR to monitorTrypanosoma
cruziparasitemiain patientswithchronic
Chagasdiseasecardiomyopathy: A
substudyfromtheBENEFITtrial.Acta Trop. 2013;125(1):23–31.
doi:10.1016/j.actatropica.2012.08.020PMID:22982466
23. RamírezJD, GuhlF, Umezawa ES, MorilloCA, Rosas F,
Marin-Neto JA., et al. Evaluation of adult
chronicChagas’heartdiseasediagnosisby molecular and
serological methods.J ClinMicrobiol.2009

Dec;47(12):3945–51. doi:10.1128/JCM.01601-
09PMID:19846646
24. DuarteL, FlorezO, RinconG, GonzalezC. Comparison of
seven diagnostic tests to detectTrypano-
somacruzi infection in patients in chronicphaseof Chagas
disease.ColombMed.2014;45(2):61–6.
PMID:25100890
25. DuffyT, Cura CI,RamirezJC, AbateT, Cayo NM,Parrado R,
et al. AnalyticalPerformance of a Multi-
plex Real-TimePCR Assay Using TaqMan Probesfor
Quantificationof Trypanosoma cruziSatellite
DNA in BloodSamples.PLoS NeglTrop Dis.
2013;7(1):e2000.Available: doi:http://journals.plos.org/
plosntds/article?id=10.1371%2Fjournal.pntd.0002000PMID:233
50002
26. RamírezJC, Cura CI,MoreiraC, Lages-silva E, Juiz N,
Velázquez E, et al.AnalyticalValidation of
Quantitative Real-TimePCR Methodsfor Quanti fi cationof
Trypanosoma cruziDNA in BloodSamples
fromChagasDiseasePatients. J Mol Diagnostics. 2015;17(5).
http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0
074-
02762015000300387&lng=en&nrm=iso&tlng=en
http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0
074-
02762015000300387&lng=en&nrm=iso&tlng=en
http://dx.doi.org/10.1590/0074-02760140402

http://dx.doi.org/10.1590/0074-02760140402
http://dx.doi.org/10.1017/S0031182009005861
http://dx.doi.org/10.1016/j.actatropica.2015.07.015
https://issuu.com/isglobal/docs/viii_taller_de_chagas_2012/4
http://www.nejm.org/doi/full/10.1056/NEJMoa1507574#t=articl
e
http://dx.doi.org/10.1128/JCM.01601-09
http://journals.plos.org/plosntds/article?id=10.1371%2Fjournal.
pntd.0002000
http://journals.plos.org/plosntds/article?id=10.1371%2Fjournal.
pntd.0002000
27. PironM,Fisa R, CasamitjanaN, López-ChejadeP, Puig L,
Vergés M,et al. Development of a real-time
PCR assay for Trypanosoma cruzidetectionin blood
samples.Acta Trop. 2007;103(3):195–200.
PMID:17662227
28. SchijmanAG, BisioM,OrellanaL, Sued M,DuffyT,
MejiaJaramilloAM,et al. International study to

evaluate PCR methodsfor detectionof Trypanosoma cruziDNA
in blood samples fromChagasdis-
ease patients. PLoS NeglTrop Dis. 2011;5(1):e931.Available:
doi:http://journals.plos.org/plosntds/
article?id=10.1371/journal.pntd.0000931 PMID:21264349
29. Hajian-Tilaki K. Samplesize estimationin diagnostic
teststudiesof biomedical informatics. J Biomed
Inform.2014Apr;48:193–204.
doi:10.1016/j.jbi.2014.02.013PMID:24582925
30. INS. Protocolode vigilancia en saludpublica Chagas.2014p.
28.Available: http://www.ins.gov.co/
temas-de-interes/Chagas/01%20Protocolo%20Chagas.pdf
31. Shikanai-Yasuda MA,CarvalhoNB. Oral transmission of
chagasdisease.Clin Infect Dis. 2012;54
(6):845–52.doi:10.1093/cid/cir956PMID:22238161
32. Castellanos-Dominguez YZ, Cucunubá ZM,OrozcoLC,
Valencia-HernandezCA, LeónCM,Florez
AC. Risk factors associatedwithChagasdisease in
pregnantwomenin Santander, a highlyendemic
Colombianarea.2016;21(1):140–8.
doi:10.1111/tmi.12634PMID:26578246
33. Feilij H, MullerL, GonzalezCappaSM.Directmicromethodfor
diagnosisof acuteand congenital Cha-
gas’ disease.J Clin Microbiol. 1983;18(2):327–30.
PMID:6413530
34. LópezMC, DuqueBeltránS, OrozcoVargas LC, CamargoD,
GualdrónLE, CáceresE, et al. Inmuno-
diagnóstico de la infección chagásicaporELISA.Biomédica.
1999;19(2):159–63.

35. MeloMF, MoreiraOC, Tenório P, LorenaV, Lorena-Rezende
I, Júnior WO, et al. Usefulness of real
timePCR to quantifyparasite load in
serumsamplesfromchronicChagasdiseasepatients.Parasit
Vectors. 2015; 8:154.Available: doi:
http://parasitesandvectors.biomedcentral.com/articles/10.1186/
s13071-015-0770-0PMID:25890282
36. SoutoRP, FernandesO, MacedoAM,Campbell D
a.,ZingalesB. DNAmarkersdefinetwo majorphylo-
genetic lineagesof Trypanosoma cruzi.MolBiochemParasitol.
1996;83(2):141–52.PMID:9027747
37. LiarteDB, MurtaSMF, SteindelM, Romanha AJ.
Trypanosoma cruzi:MultiplexPCR to detectand clas-
sify strainsaccordingto groupsI and II.Exp
Parasitol.2009;123(4):283–91.Available: doi:http://dx.doi.
org/10.1016/j.exppara.2008.12.005PMID:19133262
38. BrisseS, Verhoef J, TibayrencM.Characterisationof largeand
smallsubunit rRNAand mini-exon
genesfurthersupportsthe distinction of six Trypanosomacruzi
lineages. IntJ Parasitol. 2001;31
(11):1218–26. PMID:11513891
39. RamírezJD, GuhlF, Rendón LM,RosasF, Marin-Neto JA,
MorilloC a. ChagasCardiomyopathy Mani-
festations and TrypanosomacruziGenotypes Circulating in
ChronicChagasic Patients. PLoS Negl
Trop Dis. 2010;4(11):e899.Available:
doi:http://dx.plos.org/10.1371/journal.pntd.0000899 PMID:
21152056
40. VillaLM,GuhlF, ZabalaD, Ramírez JD, UrreaDA,
HernándezDC, etal.The identification of two Trypa-
nosomacruzi I genotypes fromdomesticand sylvatic transmission

cycles in Colombiabasedon a sin-
gle polymerasechainreactionamplificationof the spliced-leader
intergenicregion.MemInstOswaldo
Cruz.2013;108(7):932–5. doi:10.1590/0074-0276130201
PMID:24037107
41. BurgosJM,DiezM,ViglianoC, BisioM,RissoM,DuffyT,
etal.Molecular IdentificationofTrypanosoma
cruziDiscreteTyping Units in End�Stage
ChronicChagasHeartDisease and Reactivation afterHeart
Transplantation.Clin Infect Dis. 2010;51(5):485–95.Available:
doi:http://cid.oxfordjournals.org/
lookup/doi/10.1086/655680 PMID:20645859
42. Castroa M,Luquettia O, Rassia, Rassi GG,ChiariE,
GalvãoLMC. Bloodcultureand polymerase
chainreactionfor thediagnosisof thechronicphaseof
humaninfection withTrypanosomacruzi.Para-
sitolRes. 2002;88:894–900. PMID:12209329
43. Pereira J, Wilcox H, CouraJR. The evolution of
chronicchagasiccardiopathy. I. The influence of para-
sitemia.Rev Soc Bras MedTrop. 1992;25(2):101–
8.PMID:1308935
44. CastroC, PrataA, MacedoV. The influenceof the
parasitemiaon theevolution of the chronicChagas’
disease.Port Rev Soc Bras Med.2005;36:1–6.
45. MateusJ, Lasso P, Pavia P, RosasF, RoaN, Valencia-
Hern??ndez CA, et al.Low Frequency of Circu-
latingCD8+ T StemCellMemoryCells in
ChronicChagasicPatients withSevere Formsof theDisease.
PLoS NeglTrop Dis.
2015;9(1):e3432.Available:doi:http://journals.plos.org/plosntds/
article?id=10.

1371/journal.pntd.0003432 PMID:25569149
46. BustamanteJM,Novarese M,Rivarola HW, Lo PrestiMS,
FernándezAR, EndersJE, et al.Reinfec-
tionsand Trypanosoma cruzistrainscan determinetheprognosisof
thechronicchagasiccardiopathy
in mice.Parasitol Res. 2007;100(6):1407–10.PMID:17372764
http://journals.plos.org/plosntds/article?id=10.1371/journal.pntd
.0000931
.0000931
http://dx.doi.org/10.1016/j.jbi.2014.02.013
http://www.ins.gov.co/temas-de-
interes/Chagas/01%20Protocolo%20Chagas.pdf
http://www.ins.gov.co/temas-de-
interes/Chagas/01%20Protocolo%20Chagas.pdf
http://dx.doi.org/10.1093/cid/cir956
http://dx.doi.org/10.1111/tmi.12634
http://parasitesandvectors.biomedcentral.com/articles/10.1186/s
13071-015-0770-0
http://parasitesandvectors.biomedcentral.com/articles/10.1186/s
13071-015-0770-0

http://dx.doi.org/10.1016/j.exppara.2008.12.005
http://dx.plos.org/10.1371/journal.pntd.0000899
http://dx.doi.org/10.1590/0074-0276130201
http://cid.oxfordjournals.org/lookup/doi/10.1086/655680
http://cid.oxfordjournals.org/lookup/doi/10.1086/655680
.0003432
.0003432
47. LanaM De, LopesL a, MartinsHR,BahiaMT, Machado-de-
AssisGF, Wendling AP, et al.Clinicaland
laboratorystatusof patientswithchronicChagasdisease living in a
vector-controlledarea in Minas
Gerais, Brazil,before and nineyears
afteraetiologicaltreatment.MemInstOswaldo Cruz.2009;104
(8):1139–47. PMID:20140375
48. LacunzaCD, NegretteOS, MoraMC, GarcíaBustosMF,
Basombrío MÁ.Uso de la reacciónencadena
de la polimerasa para el control terapéuticode la infección
crónicaporTrypanosoma cruzi.Rev Patol
Trop. 2015;44(1):21–32. Available:
http://www.revistas.ufg.br/index.php/iptsp/article/view/34815

49. Afonso AM,Ebell MH,Tarleton RL. A SystematicReview of
HighQualityDiagnostic Tests for Chagas
Disease.PLoS NeglTrop Dis. 2012;6(11).Available:
doi:http://journals.plos.org/plosntds/article?id=
10.1371/journal.pntd.0001881
50. GuhlF, RamírezJD. Retrospective molecular
integratedepidemiology of Chagasdisease in Colombia.
Infect GenetEvol. 2013;20:148–54.Available:
doi:http://dx.doi.org/10.1016/j.meegid.2013.08.028
PMID:24012949
51. RamírezJD, MontillaM,CucunubáZM,FlorézAC, ZambranoP,
GuhlF. Molecular Epidemiology of
HumanOralChagasDisease Outbreaksin Colombia. PLoS
NeglTrop Dis. 2013;7(2):1–7.doi: http://
journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0002
041
52. Elias MCQB, Vargas NS, ZingalesB, Schenkman S.
Organization of satelliteDNA in the genomeof
Trypanosoma cruzi.MolBiochemParasitol.2003;129(1):1–9.
PMID:12798501
53. GonzalezA, Prediger E, HuecasME,NogueiratN,
LizardiPM,NogueiraN, et al.Minichromosomal
repetitiveDNA in Trypanosoma cruzi: its use in a high-
sensitivity parasitedetectionassay. ProcNatl
Acad Sci U S A. 1984;81:3356–60. PMID:6427769
54. MartinsC, BaptistaCS, IenneS, Cerqueira GC, Bartholomeu
DC, ZingalesB. Genomicorganization
and transcription analysisof the 195-bpsatelliteDNA in
Trypanosomacruzi.MolBiochemParasitol.
2008;160(1):60–4. Available:
doi:http://linkinghub.elsevier.com/retrieve/pii/S0166685108000

674
PMID:18440654
55. Sales-Campos H, KappelHB, Andrade CP, LimaTP, de
CastilhoA, GiraldoLER,et al. Trypanosoma
cruziDTUTcIIpresentshigherblood parasitism thanDTU TcI in
an experimental modelof mixed infec-
tion.ActaParasitol. 2015;60(3):435–41. Available:
doi:http://www.degruyter.com/view/j/ap.2015.60.
issue-3/ap-2015-0060/ap-2015-0060.xmlPMID:26204180
56. NichollsRS, Cucunubá ZM,KnudsonA, FlórezAC,
MontillaM, PuertaBula CJ, et al.Enfermedadde
Chagasagudaen Colombia,una entidadpoco sospechada.
Informede 10 casos presentados en el
periodo2002a 2005.Biomédica, Rev del Inst Nac
Salud.2007;27(supl.1):8–17.
57. MarciliA, Valente VC, Valente SA, Junqueira ACV, Silva
FMda, PintoAY das N, et al. Trypanosoma
cruzi in BrazilianAmazonia: Lineages TCIand TCIIa in wild
primates, Rhodniusspp. and in humans
withChagasdiseaseassociatedwithoral transmission. IntJ
Parasitol. 2009;39(5):615–23.Available:
doi:http://linkinghub.elsevier.com/retrieve/pii/S0020751908004
086PMID:19041313
58. Valente SA, da CostaValente V, das Neves PintoAY, de
Jesus BarbosaCésarM,dos SantosMP,
MirandaCOS, et al. Analysisof an
acuteChagasdiseaseoutbreakin theBrazilianAmazon:human
cases, triatomines, reservoirmammalsand parasites.Trans R Soc
Trop MedHyg.2009;103(3):291–
7. doi:10.1016/j.trstmh.2008.10.047PMID:19118852
59. BlanchetD, BrenièreSF, SchijmanAG, BisioM,SimonS,

Véron V, et al.First reportof a family out-
breakof Chagasdisease in French Guianaand posttreatment
follow-up. Infect GenetEvol. Elsevier B.
V.; 2014;28:245–50. Available:
doi:http://www.ncbi.nlm.nih.gov/pubmed/25448161PMID:25448
161
60. SotoH, Tibaduiza T, MontillaM,Triana-Chávez O,
SuárezDC, Torres Torres M,et al. Investigación de
vectores y reservorios en brotede Chagasagudoporposible
transmisión oral en Aguachica,Cesar,
Colombia.Cad SaúdePública. 2014;30(4):746–56. Available:
doi:http://www.scielo.br/pdf/csp/v30n4/
0102-311X-csp-30-4-0746.pdfPMID:24896050
61. HernándezC, CucunubáZ, Parra E, Toro G,ZambranoP,
Ramírez JD. Chagasdisease(Trypano-
somacruzi)and HIV co-infection in Colombia.IntJ Infect Dis.
2014;26:146–8.Available: http://
linkinghub.elsevier.com/retrieve/pii/S1201971214014969doi:10.
1016/j.ijid.2014.04.002PMID:
25080354
62. MessengerLA, RamirezJD, Llewellyn MS, GuhlF
MM.Importation of hybrid human-associatedTrypa-
nosomacruzistrains of southernSouthAmerican
origin,Colombia.EmergInfect Dis. 2016;22(8).
63. BianchiF, CucunubáZ, GuhlF, GonzálezNL, Freilij H,
NichollsRS, et al. Follow-up of an Asymptom-
atic ChagasDiseasePopulation of ChildrenafterTreatment
withNifurtimox(Lampit) in a Sylvatic
EndemicTransmission Areaof Colombia. PLoS NeglTrop Dis.
2015;9(2):e0003465.Available: doi:
PMID:25723465

64. RamírezJD, GuhlF, MessengerL a.,Lewis MD,
MontillaM,CucunubaZ, et al. Contemporarycryptic
sexuality in Trypanosoma cruzi.MolEcol. 2012;21(17):4216–26.
doi:10.1111/j.1365-294X.2012.
05699.xPMID:22774844
http://www.revistas.ufg.br/index.php/iptsp/article/view/34815
.0001881
.0001881
.0002041
.0002041
http://linkinghub.elsevier.com/retrieve/pii/S0166685108000674
http://www.degruyter.com/view/j/ap.2015.60.issue-3/ap-2015-
0060/ap-2015-0060.xml
http://www.degruyter.com/view/j/ap.2015.60.issue-3/ap-2015-
0060/ap-2015-0060.xml
http://dx.doi.org/10.1016/j.trstmh.2008.10.047

http://www.scielo.br/pdf/csp/v30n4/0102-311X-csp-30-4-
0746.pdf
http://www.scielo.br/pdf/csp/v30n4/0102-311X-csp-30-4-
0746.pdf
http://dx.doi.org/10.1016/j.ijid.2014.04.002
http://dx.doi.org/10.1111/j.1365-294X.2012.05699.x
http://dx.doi.org/10.1111/j.1365-294X.2012.05699.x
65. GuhlF, RamírezJD. Trypanosomacruzi I diversity:Towards
the needof geneticsubdivision?Acta
Trop. 2011;119(1):1–4. Available:
doi:http://dx.doi.org/10.1016/j.actatropica.2011.04.002 PMID:
21510916
66. RamírezJD, DuqueMC, MontillaM,Cucunubá Z, GuhlF.
Naturaland emergentTrypanosoma cruzi I
genotypesrevealed by mitochondrial (Cytb)and nuclear
(SSUrDNA)geneticmarkers.Exp Parasitol.
2012;132:487–94.
doi:10.1016/j.exppara.2012.09.017PMID:23031800
67. CruzL, Vivas A, MontillaM,HernándezC, FlórezC, Parra E,
et al. Comparative studyof thebiological
propertiesof Trypanosoma cruzi I genotypes in a
murineexperimental model. Infect GenetEvol. 2015;

29:110–7.Available:
doi:10.1016/j.
meegid.2014.11.012PMID:25461848
68. MolinaI, Gómezi Prat J, Salvador F, Treviño B, SulleiroE,
SerreN, et al.RandomizedTrial of Posaco-
nazole and Benznidazole for ChronicChagas’ Disease.N Engl J
Med.2014;370(20):1899–908.Avail-
able:
doi:http://www.nejm.org/doi/abs/10.1056/NEJMoa1313122PMI
D:24827034
http://www.nejm.org/doi/abs/10.1056/NEJMoa1313122
PU34CH02-Power ARI 12 February 2013 19:27
From Developmental Origins
of Adult Disease to Life
Course Research on Adult

Disease and Aging: Insights
from Birth Cohort Studies
Chris Power,1 Diana Kuh,2 and Susan Morton3
1 MRC Center of Epidemiology for Child Health/Center for
Pediatric Epidemiology &
Biostatistics, University College London Institute of Child
Health, London WC1N 1EH,
United Kingdom; email: [email protected]
2 MRC Unit for Lifelong Health and Ageing, London WC1B
5JU, United Kingdom
3 Centre for Longitudinal Research—He Ara ki Mua, University
of Auckland Tamaki
Campus, Glen Innes, Auckland 1743, New Zealand
Annu. Rev. Public Health 2013. 34:7–28
The Annual Review of Public Health is online at
publhealth.annualreviews.org
This article’s doi:
10.1146/annurev-publhealth-031912-114423
Copyright c© 2013 by Annual Reviews.
All rights reserved
Keywords
lifetime socioeconomic position, growth trajectories, cognitive
and
emotional function, adult health, life course conceptual
frameworks,
intergenerational influences
Abstract
Maturation of long-running birth cohort studies has fostered a

life
course approach to adult health, function, and disease and
related to
conceptual frameworks. Using broad concepts of human
development
including physical, cognitive, and emotional function, birth
cohorts
provide insights into the processes across the life course and
between
generations that link to adult outcomes. We discuss findings on
the
determinants and health consequences of lifetime trajectories of
body
size, cognitive and emotional function, and socioeconomic
position.
Findings from the studies suggest that, for some adult health
outcomes,
explanations will be incomplete unless exposures and processes
from
across the life course are taken into account. New birth cohort
studies
are poised to delineate further the nature and timing of life
course
relationships in contemporary generations of children.
7
A
nn
u.
R
ev
. P

ub
li
c
H
ea
lt
h
20
13
.3
4:
7-
28
. D
ow
nl
oa
de
d
fr
om
w
w
w

.a
nn
ua
lr
ev
ie
w
s.
or
g
by
1
44
.8
2.
38
.2
9
on
0
9/
25
/1
4.

F
or
p
er
so
na
l
us
e
on
ly
.
CVD: cardiovascular
disease
INTRODUCTION
Research on the developmental origins of adult
disease is broad in scope, embracing many
scientific disciplines, exposures, and outcomes.
Such research recognizes the potential for
influences on early-life development to lead to
changes that impact on disease risk decades later
in adulthood. This developmental perspective
is long-established within some disciplines, for
example, those concerned with emotional and

cognitive function, but for other disciplines
it is relatively recent. A major impetus to
developmental origins research was provided
in the early 1980s by a series of studies linking
low birth weight, as a proxy for poor prenatal
growth, to increased risk of chronic diseases
in adulthood, including cardiovascular disease
(CVD) and diabetes (the fetal origins hypothe-
sis) (2). Earlier in the past century, investigators
were interested in early environmental in-
fluences on the individual’s constitution that
might affect later mortality risk (52). Nonethe-
less, the fetal origins hypothesis represented a
shift in emphasis for research on adult chronic
disease, which had focused largely on adult
lifestyles. Emerging research over the past three
decades has led to a convergence of evidence on
the wide-ranging effects of early environment
and associated development for later health
outcomes (50, 99) and to establishment of
the International Society for Developmental
Origins of Health and Disease in 2003.
Birth cohort studies established some
decades ago have been well positioned to
TRAJECTORIES
“A trajectory provides a long-term view of one dimension of an
in-
dividual’s life over time. These may be social states (such as
work,
marriage, socioeconomic position), psychological states (such
as depression) or physiological states (such as lung function).
Implicit is the idea of a normative trajectory around which in-
dividuals deviate” (52). For health function, a trajectory may in-

clude a period of gain to a peak followed by a period of decline
(e.g., for lung function).
investigate developmental origins hypotheses,
and in turn, these hypotheses have provided a
stimulus for the establishment of new cohorts.
There are now several birth and infancy
cohorts in Britain, New Zealand, Finland,
and elsewhere, including those established in
the past few years and longer-running studies
with follow-up in some instances of five or
more decades. Basic details of some of these
studies are given in Table 1; the list is far from
comprehensive but illustrates the range of birth
dates, length of follow-up, and original study
purposes. Both younger and older studies alike
have tended to collect information on parental
characteristics and social and family back-
ground, as well as conditions during pregnancy
and in early childhood. In addition, many older
studies have assessed the physical, cognitive,
behavioral, and emotional development of
their participants and collected information on
social destinations, lifestyles, and other putative
influences on later disease risk. In recognition
of the importance of charting early develop-
mental milestones and trajectories (see sidebar
on Trajectories), some younger cohorts have
been instigated during pregnancy rather than at
or soon after birth and also have more frequent
contacts early in childhood than did some older
studies. Longer-running studies have been able
to investigate influences across developmental
domains (physical, cognitive, emotional) in
relation to health in later life, and with infor-
mation collected at different life stages, these

maturing birth cohorts have fostered a life
course approach to adult disease. The objective
of a life course approach in epidemiology is
to establish how social and biological factors
operating at different stages of life and across
generations contribute to the development of
adult health and disease over time (52). With
its consideration of different ages, life course
research seeks to understand influences of
early-life exposures and development on later
disease outcome and the processes occurring
in the intervening years of life that link them.
Thus, life course epidemiology extends the de-
velopmental origins of adult disease perspective
by focusing attention on potentially sensitive
8 Power · Kuh · Morton
A
nn
u.
R
ev
. P
ub
li
c
H
ea
lt

h
20
13
.3
4:
7-
28
. D
ow
nl
oa
de
d
fr
om
w
w
w
.a
nn
ua
lr
ev
ie

w
s.
or
g
by
1
44
.8
2.
38
.2
9
on
0
9/
25
/1
4.
F
or
p
er
so
na

l
us
e
on
ly
.
Table 1 Selected birth (or infancy) cohortsa
Cohort
name/country
Year of
birth Selection/design
N at
baseline
Follow-ups
(timing)
N at last
sweep
(%)
Initial reason
for/focus of study
National Survey of
Health and Devel-

opment/Britain
(54, 108)
1946 Socially stratified
sample of singleton
babies born in one
week, March 1946,
to married women
5,362 Birth, 2, 4, 6, 7, 8, 9,
10, 11, 13, 15, 19,
20, 22, 23, 26, 31,
36, 43, 47(F),
48(F), 49(F), 50(F),
51(F), 52(F), 53,
54(F), 60–64 years
2,661 To consider cost of and
care in pregnancy and
childbirth. Social class
differences in
maternal and child
mortality and
morbidity
National Child
Development
Study/Britain (77)
1958 All born in one
week, March 1958
17,638 Birth, 7, 11, 16, 23,
33, 42, 45, 46,
50 years

9,790 To identify social and
obstetric factors
linked to stillbirth and
neonatal death
Aberdeen Children
of the 1950s
Study/Britain (62)
1950–
1956
All primary-school
children in
Aberdeen in 1962
12,150 Perinatalb (linked),
7, 9, 11,
45–50 years
7,655 To understand
predictors of low
childhood cognition
Northern Finnish
Birth Cohort
Study/Finland (87)
1966 Live born,
European descent,
with expected birth
dates in 1966, Oulu
and Lapland
(northern Finland)
12,058 Pregnancy, birth,

1 year, 14 years,
31 years
8,690 To examine risk factors
for childhood
mortality and
morbidity in a
geographically
defined population
1970 British Birth
Cohort
Study/Britain (20)
1970 All born in one
week, April 1970
16,571 Birth, 5, 10, 16, 26,
30, 34, 42 years
9,656 at
34 years
To examine the social
and biological
characteristics of
mothers in relation to
neonatal morbidity
Dunedin
Multidisciplinary
Health and
Development
Study/New
Zealand (98)

1972 All births in
Dunedin March
1972–April 1973
enrolled at 3 years
1,037 3, 5, 7, 9, 11, 13, 15,
18, 21, 26, 32 years
972 To conduct a
longitudinal
population-based
multidisciplinary
study of child health,
development, and
behavior
Christchurch
Health and
Development
Study/New
Zealand (24)
1977 Children born
April–early August
1977 in
Christchurch
1,265 Annually from birth
to 16, 18, 21, 25,
30 years
934 To conduct a
longitudinal birth
cohort focused on
child health and
development

Avon Longitudinal
Study of Parents
and Children
(ALSPAC)/England
(7)
1991–
1992
Offspring of all
pregnant women in
the county of Avon,
estimated delivery
date April 1,
1991—December
31, 1992
15,247
eligible
preg-
nancies
en-
rolled
68 time points, birth
to 18 years
7,729c To investigate
modifiable influences
on child health and
development
(Continued )
www.annualreviews.org • Life Course Research on Adult

Disease 9
A
nn
u.
R
ev
. P
ub
li
c
H
ea
lt
h
20
13
.3
4:
7-
28
. D
ow
nl
oa

de
d
fr
om
w
w
w
.a
nn
ua
lr
ev
ie
w
s.
or
g
by
1
44
.8
2.
38
.2

9
on
0
9/
25
/1
4.
F
or
p
er
so
na
l
us
e
on
ly
.
Table 1 (Continued )
Cohort

name/country
Year of
birth Selection/design
N at
baseline
Follow-ups
(timing)
N at last
sweep
(%)
Initial reason
for/focus of study
Southampton
Women’s
Survey/England
(44)
1998–
2007
Births to recruited
12,583
prepregnant
women
3,156 Prepregnancy,
pregnancy
(multiple),
6 months, 1, 2, 3,
4, 6, 8 years

1,477 at
6 years
To understand
perinatal and early life
determinants of
children’s growth and
development
Millennium Cohort
Study
(MCS)/United
Kingdom (14)
2001 Nationally
representative
sample across
United Kingdom
recruited during
infancy
19,244 9 months, 3, 5, 7,
11 years
15,590 at
7 years
To understand the
biological and
environmental
determinants of
contemporary child
development
a Note: This list of studies is not exhaustive. Birth cohorts

documented have collected perinatal information and have
existed for long enough to provide
sufficient longitudinal data to inform the areas addressed in this
article. The success of older cohorts has provided the impetus
for many new birth cohort
studies begun around or after the turn of the millennium. Newer
studies include those in the European Child Cohort Network
(EUCCONET), as well as
studies in Australia (Growing Up in Australia—LSAC) and New
Zealand (Pacific Island Families Study and Growing Up in New
Zealand). EUCCONET
includes the Norwegian Birth cohort (MoBa), Danish National
Birth cohort, Generation R Study (The Netherlands), Born in
Bradford (United
Kingdom), Millennium Cohort study (MCS), Growing Up in
Scotland (GUS), Growing Up in Ireland (GUI), ELFE (France)
and will include a newly
planned UK study.
b The Aberdeen Children of the 1950s Study is included here
because it has birth information for the 14,932 children who
were part of this cohort.
However, they were enrolled between ages 5 and 11 years (in
1962), and perinatal data were linked for the cohort at that time.
c Completing ≥1 item during the transition to adulthood.
periods in childhood and adolescence as well
as in the prenatal period. It extends the adult
lifestyle theories of chronic disease by focusing
attention on the early acquisition of lifestyle and
its cumulative effects. It extends social causation
theories of adult chronic disease by drawing
attention to the impact of the socioeconomic
environment in childhood as well as adulthood.
It also extends both developmental origins and
adult theories of disease causation by consider-
ing the joint action of early and later exposures.

With an increasing number of studies and
duration of follow-up, birth cohorts are now
charting new territories. Some recently estab-
lished cohorts are focusing on understanding
early development in the wider neighborhood
and societal context (69). Older birth cohort
studies, often with decades of information, are
examining influences on the population range
of functional capacities and disease risk into
later adulthood, and an agenda on life course
influences on aging is already emerging. Many
birth cohorts now incorporate genetic factors
and are contributing to the discovery of genetic
variants associated with important phenotypes
(e.g., obesity) through consortia for genome-
wide association studies. Moreover, the cohorts
offer the prospect of advancing understanding
of (epi)genetic influences on developmental tra-
jectories and how these relate to the potentially
modifiable environmental context.
At a time when birth cohort studies are
evolving to address a range of social, economic,
and health questions, it is timely to critically
review their contributions and identify future
challenges. This article does not purport to
provide a comprehensive summary of the vast
literature from these cohorts, but it consid-
ers some main themes of relevance to public
health. Specifically, we consider examples of
where birth cohorts have stimulated thinking
about the life course frameworks that might
guide research on pathways to adult disease
and where, in this regard, they have added

new knowledge. Evidence from the studies has
been informative on many important themes,
A
nn
u.
R
ev
. P
ub
li
c
H
ea
lt
h
20
13
.3
4:
7-
28
. D
ow

nl
oa
de
d
fr
om
w
w
w
.a
nn
ua
lr
ev
ie
w
s.
or
g
by
1
44
.8
2.

38
.2
9
on
0
9/
25
/1
4.
F
or
p
er
so
na
l
us
e
on
ly
.

SEP: socioeconomic
position
e.g., on the natural history of some condi-
tions, long-term outcomes of specific preg-
nancy exposures/characteristics, and determi-
nants of healthy lifestyles, but we do not cover
these themes in any detail here. Instead, we
summarize some of the work on the follow-
ing in relation to adult function and disease:
lifetime socioeconomic position (SEP), lifetime
growth trajectories, and cognitive and emo-
tional development at different life stages. For
each of these research areas the birth cohorts
have contributed a substantial body of empir-
ical evidence. With the exception of lifetime
SEP, which to some extent can be investigated
retrospectively, associations for growth, cogni-
tion, and mental health can be assessed only
with prospectively obtained measures at differ-
ent life stages.
CONCEPTUAL FRAMEWORKS
Conceptual frameworks have been developed
to guide research on the life course processes
leading to adult disease. Investigators have pro-
posed various general models that have then
been adapted and applied to different health
outcomes. Although such frameworks have
been used in other contexts (e.g., record link-
age studies), they are particularly relevant in the
investigation of birth cohorts.
Figure 1 provides an example of a general
framework, with main components that are

often considered in life course research. This
simplified representation incorporates inter-
generational factors, developmental domains
(cognitive, emotional, and physical), social
identities and health behaviors, and environ-
mental influences that potentially act at all life
stages to affect later health. Figure 1 is pre-
sented to highlight five main points. First, birth
cohort studies, in general, can focus on deter-
minants of the full spectrum of both health and
disease in a population. Second, life course tra-
jectories for body functions (e.g., muscle func-
tion, lung function) are a dynamic way to study
lifetime influences on health and disease; these
trajectories capture the natural history of bio-
logical systems that grow and develop rapidly
LIFE COURSE MODELS OF ADULT DISEASE
OUTCOMES
The critical period model (also called biological programming
or
latency model) refers to exposures acting during a critical
window
of development that affect the structure or function of organs,
tissues, or body systems and which, in turn, affect later disease
risk. This model underpins the fetal origins of adult disease hy-
pothesis. Sensitive-period models are similar, with exposures
ex-
erting greatest effects during times of rapid development, but
there is greater scope for modification by other influences than
there is with a critical period model.
The accumulation of risk model refers to the adverse effect
on later disease of exposures accumulating over the life course.
It thereby focuses on total burden of insults, i.e., the number,

duration, or severity of a range of health-damaging environmen-
tal, socioeconomic, and behavioral factors.
The chains of risk model (also pathways model) refers to se-
quences of events or exposures, whereby one exposure increases
the likelihood that another will follow, leading to a final expo-
sure(s) that is causally related to later disease. Links are not de-
terministic, and earlier exposures do not affect disease risk but
often lead to a final link in the chain that does affect later
disease
risk. Social, biological, and psychological factors can be part of
chains of risk models, possibly acting as mediating or
modifying
factors.
Sources: adapted from Hertzman et al. (38), Keating &
Hertzman (50), Ben-Shlomo & Kuh (4), and Kuh et al. (52)
during the prenatal, prepubertal, and pubertal
periods, reaching a peak or plateau at maturity
and gradually declining with age (Figure 2).
The progressive, generalized deterioration in
function postmaturity can be thought of as bi-
ologically aging; the generally accepted dispos-
able soma theory of aging suggests this is caused
by increasing molecular and cellular damage
from environmental insults and chance (51).
Third, influences over the life course might
operate in several ways to affect adult function
and disease. Researchers have identified models
for the alternative processes that might be in-
volved; main models including critical period,
accumulation, or chains of risk are shown in the
sidebar, Life Course Models of Adult Disease
Outcomes. Although these life course models

Disease 11
A
nn
u.
R
ev
. P
ub
li
c
H
ea
lt
h
20
13
.3
4:
7-
28
. D
ow
nl

oa
de
d
fr
om
w
w
w
.a
nn
ua
lr
ev
ie
w
s.
or
g
by
1
44
.8
2.
38

.2
9
on
0
9/
25
/1
4.
F
or
p
er
so
na
l
us
e
on
ly
.
In
te

rg
e
n
e
ra
ti
o
n
a
l
in
fl
u
e
n
c
e
s
Cognitive function
Childhood Adulthood
Physical health
Emotional health

Environmental
influences
Emotional health
Physical health
A
d
u
lt
o
u
tc
o
m
e
Cognitive function/
education
Parents/
grandparents
Social (origins) identity/
health behavior
Social (destinations)
identity/health behavior
Health declineHealth gain

Figure 1
Simplified framework linking early-life exposures with adult
outcome.
can be seen as distinct, they are not necessarily
mutually exclusive, and one can envisage varia-
tions of basic models. For example, an exposure
that operates during a critical or sensitive pe-
riod may also accumulate with other exposures
over the life course, possibly through modifica-
tion of earlier factors by those occurring later
in life. Fourth, developmental domains may
Years of life
L
e
v
e
l
o
f
fu
n
c
ti
o
n
0 10 20 30 40 50 60 70 80 90 100

0
10
20
30
40
50
Level below which
limitations may occur
Adult risk factors
Ea
rl
y
lif
e
ri
sk
fa
ct
o
rs
A

B
C
D
Functional/structural
reserve
Figure 2
Life course functional trajectories. Line A, normal development
and decline;
line B, exposure during development reducing functional
reserve at maturity;
line C, exposure acting postmaturity, accelerating age-related
decline; line D,
combination of B and C. Figure modified from A Life-Course
Approach to Chronic
Disease Epidemiology, edited by Diana Kuh and Yoav Ben-
Shlomo, 2nd edition,
2004, chapter 1, page 9, figure 1.2, by permission of Oxford
University Press.
coevolve, e.g., between physical developmental
characteristics such as height and emotional
or cognitive development, whereas during
adulthood, relationships between health and
other factors (e.g., social circumstances, health
behaviors and so forth) may be bidirectional.
Hence, there are numerous links across the
domains represented in Figure 1; birth cohort
studies provide several examples of coevolution
and bidirectional relationships. Fifth, develop-
mental trajectories and subsequent health in
adulthood can be affected by intergenerational
influences. These intergenerational influences

likely represent the effect of both genetic and
environmental influences acting over time (12).
Although not specified in the simplified
Figure 1, environmental influences can
operate at many levels, including individual
(micro), community and neighborhood (meso),
or national and international (macro) levels
(38). For example, smoking in a population can
be influenced at the individual level in terms
of behaviors relating to smoking initiation and
maintenance (addiction), at the community
level in terms of peer relationships and behavior
as well as access and availability of tobacco, and
also at the national level in terms of taxation
or smoke-free legislation to prohibit smoking
A
nn
u.
R
ev
. P
ub
li
c
H
ea
lt

l
us
e
on
ly
.
Grandparent
Parent
Offspring
Time
Household
level
Neighborhood
level
National level
Joint neighborhood
effects
Childhood
cohort effect

Period effect acting on all
three generations
AB
C
Figure 3
Multigenerational schema: influences of hierarchical and life
course exposures on disease risk across three
related individuals. Intergenerational links, which can be
common genetic and/or social influences, are
shown between grandparents, parents, and offspring. Life course
links are represented by exposures
occurring at different life stages and at differing (household,
neighborhood, and national) levels.
Ben-Shlomo & Kuh (4) illustrate influences acting across time
and across individuals with the example of
adverse neighborhood conditions affecting a mother and her
child (A), and with national exposures (e.g.,
war-time rationing) as either specific to a single population
cohort (B) or experienced by all individuals (C).
Modified from Ben-Shlomo Y, Kuh D, 2002, “A life course
approach to chronic disease epidemiology:
conceptual models, empirical challenges and inter-disciplinary
perspectives,” Int. J. Epidemiol. 31(2):285–93,
by permission of Oxford University Press (4). Adapted from
Hertzman et al. (38).
BMI: body mass index
indoors. Influences acting at these different
levels might affect successive generations at
different points in their lives, with consequent
variations in impact: An exposure could occur
at a critical or sensitive period for one genera-

tion with attendant effects over the life course,
but with lesser impact when experienced at a
later life stage for another generation. Figure 3
illustrates the links between generations and
life stages, with influences from different levels
(individual, neighborhood, and national) of the
broader social environment, as successive gen-
erations live through different periods of time.
The general frameworks in Figures 1 and
3, and other such schemas (29, 69), are not
intended to be comprehensive, and when they
are adapted and refined to investigate particular
life course relationships, details of lifetime
exposures and outcomes can be elaborated.
For example, a study of body mass index (BMI)
at different life stages considered exposures
from the prenatal period (maternal age, BMI,
blood pressure, and smoking in pregnancy),
adolescence, and adulthood (physical activity,
diet, smoking, alcohol consumption, and SEP)
and found that life course factors (e.g., physical
activity) had strong associations, separate from
genetic factors, with BMI at age 31 years (49).
In view of the complexity of elaborated models,
many initial life course studies focus on SEP at
different life stages, which, although a broad,
nonspecific, and distal measure, has been used
to indicate when related exposures might be
operating.
Disease 13
A

nn
u.
R
ev
. P
ub
li
c
H
ea
lt
h
20
13
.3
4:
7-
28
. D
ow
nl
oa
de
d

fr
om
w
w
w
.a
nn
ua
lr
ev
ie
w
s.
or
g
by
1
44
.8
2.
38
.2
9
on

0
9/
25
/1
4.
F
or
p
er
so
na
l
us
e
on
ly
.
LIFETIME SOCIOECONOMIC
POSITION AND
ADULT DISEASE
SEP is a long-established determinant of health
and disease, whereby better outcomes are gen-

erally seen with increasing socioeconomic ad-
vantage; i.e., generally, the association is graded
with health benefits for each increase in SEP.
From their earliest days, many birth cohort
studies have had a particular interest in SEP,
and they have shown a typical trend whereby
Hearing at 4 kHz
IgE
FEV1
Fibrinogen
Triglycerides
HDL cholesterol
Total cholesterol
HbA1c
BMI
DBP
SBP
D
is
e
a
s

e
r
is
k
f
a
c
to
r
–0.10 –0.05 0 0.05 0.10
Difference (SD score) per increase in social class
Figure 4
Associations of child and adult social class with disease risk
factors at age
45 years in the 1958 British birth cohort. Disease risk factors at
45 years
include systolic and diastolic blood pressure (SBP and DBP),
body mass index
(BMI), HbA1c, total and HDL cholesterol, triglycerides,
fibrinogen,
one-second forced expiratory volume (FEV1), total
immunoglobulin E (IgE),
hearing threshold at 4 kHz. Risk factors are converted to
standard deviation
(SD) scores to allow comparison of associations for child and
adult social class
across different outcomes. Estimated effects are differences per
unit increase in
social class (on a six-point scale from professional to unskilled

RESEARCH ARTICLEMolecular Diagnosis of Chagas Disease in.docx

RESEARCH ARTICLEMolecular Diagnosis of Chagas Disease in.docx

Recommended

Recommended

More Related Content

Similar to RESEARCH ARTICLEMolecular Diagnosis of Chagas Disease in.docx

Similar to RESEARCH ARTICLEMolecular Diagnosis of Chagas Disease in.docx (20)

More from audeleypearl

More from audeleypearl (20)

Recently uploaded

Recently uploaded (20)

RESEARCH ARTICLEMolecular Diagnosis of Chagas Disease in.docx