1356
www.thelancet.com/infection Vol 16 December 2016
Articles
Association between Zika virus infection and microcephaly
in Brazil, January to May, 2016: preliminary report of a
case-control study
Thalia Velho Barreto de Araújo, Laura Cunha Rodrigues, Ricardo Arraes de Alencar Ximenes, Demócrito de Barros Miranda-Filho,
Ulisses Ramos Montarroyos, Ana Paula Lopes de Melo, Sandra Valongueiro, Maria de Fátima Pessoa Militão de Albuquerque,
Wayner Vieira Souza, Cynthia Braga, Sinval Pinto Brandão Filho, Marli Tenório Cordeiro, Enrique Vazquez, Danielle Di Cavalcanti Souza Cruz,
Cláudio Maierovitch Pessanha Henriques, Luciana Caroline Albuquerque Bezerra, Priscila Mayrelle da Silva Castanha, Rafael Dhalia,
Ernesto Torres Azevedo Marques-Júnior, Celina Maria Turchi Martelli, on behalf of investigators from the Microcephaly Epidemic Research
Group, the Brazilian Ministry of Health, the Pan American Health Organization, Instituto de Medicina Integral Professor Fernando Figueira,
and the State Health Department of Pernambuco*
Summary
Background The microcephaly epidemic, which started in Brazil in 2015, was declared a Public Health Emergency of
International Concern by WHO in 2016. We report the preliminary results of a case-control study investigating the
association between microcephaly and Zika virus infection during pregnancy.
Methods We did this case-control study in eight public hospitals in Recife, Brazil. Cases were neonates with
microcephaly
. Two controls (neonates without microcephaly), matched by expected date of delivery and area of
residence, were selected for each case. Serum samples of cases and controls and cerebrospinal fl
uid samples of cases
were tested for Zika virus-specifi c IgM and by quantitative RT-PCR. Laboratory-confi rmed Zika virus infection during
pregnancy was defi ned as detection of Zika virus-specifi c IgM or a positive RT-PCR result in neonates. Maternal
serum samples were tested by plaque reduction neutralisation assay for Zika virus and dengue virus. We estimated
crude odds ratios (ORs) and 95% CIs using a median unbiased estimator for binary data in an unconditional logistic
regression model. We estimated ORs separately for cases with and without radiological evidence of brain abnormalities.
Findings Between Jan 15, 2016, and May 2, 2016, we prospectively recruited 32 cases and 62 controls. 24 (80%) of
30 mothers of cases had Z
ika virus infection compared with 39 (64%) of 61 mothers of controls (p=0·12). 13 (41%) of
32 cases and none of 62 controls had laboratory-confi rmed Zika virus infection; crude overall OR 55·5 (95% CI
8·6–∞); OR 113·3 (95% CI 14·5–∞) for seven cases with brain abnormalities; and OR 24·7 (95% CI 2·9–∞) for four
cases without brain abnormalities.
Interpretation Our data suggest that the microcephaly epidemic is a result of congenital Zika virus infection. We
await further data from this ongoing study to assess other potential risk factors and to confi rm the strength of
association in a larger sample size.
Funding Brazilian M
inistry of Health, Pan American Health Organization, and Enhancing Research Activity in
Epidemic S
ituations.
Copyright © This is an Open A
ccess article published under the CC BY 3.0 IGO license, which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited. In any use of this article, there
should be no suggestion that WHO endorses any specifi c organisation, products or services. The use of the WHO logo is
not permitted. This notice should be preserved along with the article’s original URL.
Introduction
In August, 2015, an increase in the number of neonates with
microcephaly was detected in Brazil, which by the end of the
year had become a major epidemic. Up to June 25, 2016,
8165 cases had been notifi ed, of which 1638 were confi rmed,
3466 excluded, and 3061 remained under investigation.
1
Microcephaly results from any insult that disturbs early
brain growth, and can be caused by genetic variations,
teratogenic agents, or other congenital infec tions.
2
Because
of the temporal and geographical overlap with an ongoing
outbreak of Zika virus, the hypo thesis was soon formulated
that the micro cephaly epidemic was caused by Zika virus
infection during pregnancy
. In November, 2015, the
Brazilian Ministry of Health declared the situation a national
public health emergency.
3
Human infection with Zika virus was initially limited
to sporadic cases in a small number of countries and
perceived not to cause outbreaks or severe disease.
Outbreaks were fi rst detected in the Pacifi c Islands in
2007 and 2013. Since 2007, transmission has been
detected in 61 countries and territories, most of them
located in the Americas.
4
In February, 2016, WHO
Lancet Infect Dis 2016;
16: 1356–63
Published Online
September 15, 2016
http://dx.doi.org/10.1016/
S1473-3099(16)30318-8
See Comment page 1307
See Correspondence page 1331
*Investigators contributing
on
behalf of these organisations are
listed at
the end of the report
Department of Social Medicine
(T V B Araújo PhD,
S Valongueiro PhD), and
Department of Tropical
Medicine
(Prof R A A Ximenes PhD),
Federal University of
Pernambuco, Recife, Brazil;
Department of Community
Health, Federal University of
Pernambuco, Vitória de Santo
Antão, Brazil (A P L Melo MSc);
Department of Infectious
Disease Epidemiology, London
School of Hygiene & Tropical
Medicine, London, UK
(Prof L C Rodrigues PhD);
University of Pernambuco,
Recife, Brazil
(Prof R A A Ximenes,
D B Miranda-Filho PhD,
U R Montarroyos PhD);
The Research Center Aggeu
Magalhães (CPqAM) and
Oswaldo Cruz Foundation
(Fiocruz), Recife, Brazil
(M F P Militão
de Albuquerque PhD,
W V Souza PhD, C Braga PhD,
S P Brandão Filho PhD,
M T Cordeiro PhD,
P M S Castanha PhD, R Dhalia PhD,
E T A Marques-Júnior PhD,
Prof C M T Martelli PhD);
Pan American Health
Organization, Brasília, Brazil
(E Vazquez PhD); Instituto
Materno Infantil Fernando
Figueira, Recife, Brazil
Articles
www.thelancet.com/infection Vol 16 December 2016
1357
(D D C S Cruz MD); Fiocruz
Brasília, Brasília, Brazil
(C M P Henriques MSc);
Pernambuco State Health
Department, Recife, Brazil
(L C A Bezerra MSc); and
Department of Community
Health, Federal University of
Goiás, Goiânia, Brazil
(Prof C M T Martelli PhD)
Correspondence to:
Dr Thália Velho Barreto de
Araújo, Federal University of
Pernambuco, Recife,
Pernambuco, 50.670-901, Brazil
thalia@ufpe.br
declared, in relation to the epidemic that started in Brazil,
that “the cluster of microcephaly cases and other neuro-
logical disorders constitutes…a Public Health Emergency
of International Concern”.
5
Since the hypothesis that the microcephaly epidemic in
Brazil is caused by congenital Zika virus infection was fi rst
proposed,
6
there has been an accumulation of evidence
supporting the association.
7–17
The relation between Zika
virus and birth defects is strong enough to be deemed
causal, but the argument would be stronger if confi rmed
by at least one case-control study and a cohort study.
18
Indeed, the evidence so far comes from case reports,
16
case
series,
19,20
modelling studies,
17
and the preliminary report of
a cohort study.
7
None of these studies included appropriate
population control groups. We report the preliminary
analysis of a case-control study, requested by the Brazilian
Ministry of Health, to investigate the causes of the
microcephaly epidemic in Brazil; the main hypothesis was
that it is caused by congenital Zika virus infection.
Methods
Study design and participants
This case-control study, with prospective recruitment of
newborn cases and concurrent controls, was done in the
metropolitan region of Recife, Pernambuco State, Brazil.
We enrolled neonates born to women resident in
Pernambuco and delivered in eight public maternity
units. Neonates with anencephaly or encephalocele were
excluded. This preliminary analysis includes neonates
born between Jan 15, 2016, and May 2, 2016.
Cases were neonates with microcephaly, defi ned as
head circumference at least 2 SD smaller than the
mean for sex and gestational age in the Fenton growth
chart.
21
Controls were live neonates without micro-
cephaly, with no brain abnormalities identifi ed by
trans fontanellar ultra sonography and no major birth
defects detected by physical examination by a
neonatologist. For each case, two controls were selected
from the fi rst neonates born from the following
morning in one of the study hospitals, matched by
health region of residence and expected date of delivery
(to ensure cases and controls were conceived at the
same stage of the epidemic). The criteria for matching
for expected date of delivery were specifi c for gestational
age of cases. For cases born at term and post-term
(37 weeks or more), controls were the next eligible
neonates born at 37 weeks’ gestation or more. For early
preterm cases (born at <34 weeks), controls were the
next eligible neonates who were born at less than
34 weeks’ gestation. For preterm cases born between
34 and 36 weeks’ gestation, controls were the next
eligible neonates born at 34–36 weeks’ gestation.
The protocol was approved by the research ethics
committees of the Pan American Health Organization
and Fiocruz Pernambuco. All mothers provided written
informed consent to participate in the study.
Procedures
Gestational age was estimated antenatally by fetal
ultrasonography. If ultrasound measurements were not
available, gestational age was estimated from the date of
the last menstrual period recorded on the antenatal care
card or reported by the mother. In cases in which
antenatal ultrasound had not been done and the mother
Research in context
Evidence before this study
The epidemic of microcephaly, which started in Brazil,
was
declared a Public Health Emergency of International Concern by
WHO in February, 2016. The hypothesis that the disorder is
caused by congenital Zika virus infection was proposed early on,
and the (mostly mechanistic) evidence for this association has
since been accumulating. We searched PubMed and LILACS using
the search term “Zika”. We searched these sources up until
May 30, 2016, including publications in English, Portuguese, and
Spanish. We did not identify any case-control study of Zika virus
infection and microcephaly. In a review of the evidence, published
in May, 2016, Rasmussen and colleagues used diff erent
frameworks to assess whether the available evidence supports
this hypothesis. One of these frameworks, Shepard’s criteria for
teratogenicity, requires at least two epidemiological studies of
high quality confi rming the association. The investigators
concluded that there was suffi cient evidence to accept causation,
despite the paucity of epidemiological studies. Here, we report
the fi rst case-control study of microcephaly and congenital Zika
virus infection, with the aim of adding the missing piece to the
process of defi ning causality.
Added value of this study
To our knowledge, this is the fi rst case-control study to examine
the association between microcephaly and in-utero Zika virus
infection, investigated by molecular and serological methods in
cases of microcephaly and their controls at time of birth. We did
this prospective study in the metropolitan region of Recife in
Pernambuco State, the hotspot of the microcephaly epidemic
in Brazil, between January and May, 2016. We are reporting the
preliminary fi ndings of our study because of the striking
magnitude of the association between microcephaly and Zika
virus infection.
Implications of all the available evidence
We conclude that the microcephaly epidemic is a result of
congenital Zika virus infection. We recommend that the list of
congenital infections normally referred to as TORCH
(toxoplasmosis, others [syphilis, varicella-zoster, parvovirus B19],
rubella, cytomegalovirus, and herpes) is renamed as TORCHZ,
and that we prepare for a global epidemic of microcephaly and
other manifestations of congenital Zika syndrome.
Articles
1358
www.thelancet.com/infection Vol 16 December 2016
did not know the date of last mentruation, the Capurro
method was used to estimate gestational age.
22
Head
circumference was measured in the delivery room with
a non-stretch Tefl on tape. If microcephaly was detected,
cord blood was collected. If microcephaly was confi rmed
by a second head circumference measurement 12–24 h
after birth, the neonate was deemed eligible for
the study.
After the mothers signed the informed consent form,
samples were obtained from mothers and neonates
(cases and controls), mothers were interviewed, and
children referred for brain imaging. Interviews were
done in the hospital by a trained female nurse, using a
structured standardised questionnaire. Radiological
brain imaging was done by CT scan without contrast in
cases and by transfontanellar ultrasonography by
radiologist in controls.
Cerebrospinal fl uid samples were collected from cases
by the study neonatologists. Umbilical cord blood was
collected in the delivery room from cases and controls;
when necessary, peripheral blood was collected before
the neonate left hospital. Blood specimens were sent to
the Virology and Experimental Therapy Department,
Fiocruz Pernambuco (Recife, Brazil), where they were
divided into samples and stored. For neonatal deaths and
stillbirths, macerated tissue material was tested by
RT-PCR.
RNA was extracted from serum of mothers and
neonates (cases and controls) and cerebrospinal fl uid
samples (of cases) and analysed by RT-PCR for detection
of worldwide African and Asian Zika virus genomes
using primers designed by Lanciotti and colleagues.
23
Additionally, serum of mothers and neonates and
cerebrospinal fl uid samples were tested for Zika virus-
specifi c IgM antibodies using a capture ELISA based on
the US Centers for Disease Control and Prevention
(CDC) Emergency Use Authorization protocol, with
reagents from the CDC (Fort Collins, CO, USA).
23
For the
ELISA assay, both Zika virus and dengue virus were
tested in parallel to check for cross-reactivity.
The presence of neutralising antibodies to Zika virus
and dengue virus (serotypes DENV-1–4) was assessed in
mothers and neonates by plaque reduction neutralisation
test (PRNT) in Vero cells, following a protocol described
in detail elsewhere.
24
The cutoff value for PRNT positivity
was defi ned as 50% (PRNT
50
). The PRNT
50
assay was
done using the virus strain ZIKV PE/243 isolated in
Pernambuco, Brazil. Serum samples of mothers and
neonates were tested for IgM and IgG antibodies specifi c
for toxoplasmosis, rubella, and cytomegalovirus (the
main infectious causes of congenital microcephaly
25
).
Laboratory confi rmation of Zika virus infection was
defi ned as a positive RT-PCR result or detection of IgM
antibody against Zika virus. For the purpose of this
analysis, results of brain imaging (CT scan in cases and
ultrasonography in controls) were classifi ed as normal or
abnormal (including calcifi cation, ventriculomegaly,
lissencephaly, and other abnormalities). A neonate was
deemed small for gestational age when birthweight was
lower than the 10th percentile for gestational age and sex
in the Fenton growth chart.
Statistical analysis
The original study aimed to include 200 cases and
400 controls to have 90% power, 95% precision to detect
an association with an odds ratio of 2 or greater, assuming
that 67% of cases were exposed.
We estimated the crude OR and 95% CI for the
association between microcephaly and laboratory con-
fi rmation of Zika virus infection, considering the results
in serum or cerebrospinal fl uid and the results in serum
alone, overall and separately for cases with and without
radiological evidence of brain abnormalities. To deal
with the fact that all controls tested negative for Zika
virus, the OR was calculated using a median unbiased
estimator for binary data in an unconditional logistic
regression model.
26,27
This statistical approach is
appropriate for zero cells, and was applicable to our
situation in which the sample size for this preliminary
analysis is small and data structure sparse. Another
consequence of all controls being Zika virus negative
was that although the design was matched, a conditional,
matched analysis was not needed because matched and
unmatched analyses will give the same result. We
calculated OR adjusted for maternal age and maternal
education (as a proxy of socioeconomic status) for the
overall association. We investigated the agreement
between the Zika virus-specifi c IgM for serum and that
for cerebrospinal fl uid. We also compared the ELISA
value (optical density) in serum and cerebrospinal fl uid.
We used Stata version 14.1 software for the statistical
analyses.
Role of the funding source
The funders of the study were involved in the data
interpretation and writing of the report. The cor-
responding author had full access to all the data in the
study and had fi nal responsibility for the decision to
submit for publication.
Figure: Newborn baby with microcephaly with laboratory-confi rmed Zika virus and abnormalities detected
on CT scan
The neonate shows phenotypic features previously described during the microcephaly epidemic, including
craniofacial disproportion, prominent externa occipital protuberance, and excessive scalp skin (photo, left).
Radiological features found on brain imaging (CT images, centre and right) include reduced volume of cortical
brain parenchyma, cortical and subcortical calcifi cations, simplifi ed gyral pattern, and ventriculomegaly.
Articles
www.thelancet.com/infection Vol 16 December 2016
1359
Results
This preliminary analysis included 32 neonates with
micro cephaly (cases) and 62 neonates without micro-
cephaly (controls). A photograph and cerebral CT image of
one of the cases shows features previously described
during the current microcephaly epidemic (fi gure).
28
Two cases had only one matched control. The
participation rate was 100% for cases and 76% for
controls. No controls were excluded because of birth
defects. Five cases did not have brain imaging: three died
in intensive care before CT scan was done, one was
stillborn, and another was in intensive care at the time of
this analysis. These fi ve cases were included in the
analysis for all cases and excluded from the analysis
stratifi ed by brain imaging fi ndings.
A higher proportion of mothers of controls reported that
they had no rash during pregnancy than did mothers of
cases, although this diff erence was not signifi
cant (table 1;
see appendix pp 2–5 for characteristics of individual cases
and controls). No statistical diff erence was detected in
years of education between mothers of cases and controls.
Mothers of cases had a slightly higher, although non-
signifi cant, frequency of Zika virus infection detected by
PRNT
50
than did mothers of controls (80% vs 64%,
respectively). Zika virus was the primary fl avivirus
infection in nine (10%) of 91 mothers, including the
mothers of fi ve cases. Overall, 54 (59%) of 91 mothers had
more than one fl avivirus infection, with DENV-3 and
DENV-4 the predominant dengue serotypes (appendix
p 6). Results of PRNT
50
for Zika virus and dengue virus
were similar in mother–neonate pairs (data not shown).
11 (34%) of 32 cases had severe microcephaly (head
circumference at least 3 SD smaller than the mean for
sex and gestational age at birth). Signifi cantly higher
proportions of cases than controls were born with low
birthweight and were small for gestational age (table 1).
Of the 27 cases investigated by brain imaging, 11 had one
or more abnormalities: seven had calcifi cations, fi ve had
ventriculomegaly, one had lissencephaly, and six had
other abnormalities (table 1).
Cases
(n=32)
Controls
(n=62)
p value
Mothers
Age, years
15–24 20 (63%) 34 (55%) 0·76
25–34 9 (28%) 20 (32%) ··
≥35 3 (9%) 8 (13%) ··
Number of years in education
<4 3 (9%) 4 (6%) 0·52
5–9 9 (28%) 18 (29%) ··
10–12 16 (50%) 37 (60%) ··
≥13, higher education 4 (13%) 3 (5%) ··
Reported rash during pregnancy
No rash 19 (59%) 46 (74%) 0·09
First trimester 6 (19%) 2 (3%) ··
Second trimester 2 (6%) 4 (6%) ··
Third trimester 3 (9%) 8 (13%) ··
Do not remember 2 (6%) 2 (3%) ··
Specifi c antibodies, PRNT
50
Zika virus positive 24/30
(80%)
39/61
(64%)
0·12
Zika virus negative 6/30
(20%)
22/61
(36%)
··
Not done 2 1 ··
Neonates
Sex
Girls 22 (69%) 32 (52%) 0·11
Boys 10 (31%) 30 (48%) ··
Head circumference, for
gestational age and sex*
Normal 0 62 (100%) <0·0001
Between 2 and 3 SD smaller than
the mean
21 (66%) 0 ··
>3 SD smaller than the mean 11 (34%) 0 ··
Birthweight, g
≥2500 10 (31%) 58 (94%) <0·0001
1500–2499 17 (53%) 4 (6%) ··
<1500 5 (16%) 0 ··
(Table 1 continues in next column)
Cases
(n=32)
Controls
(n=62)
p value
(Continued from previous column)
Weight for gestational age
Normal 5 (16%) 58 (94%) <0·0001
Small for gestational age† 27 (84%) 4 (6%) ··
Gestational age
Term, ≥37 weeks 24 (75%) 53 (85%) 0·21
Premature, 33–36 weeks 8 (25%) 9 (15%) ··
Method of estimation of
gestational age
First trimester fetal
ultrasonography
14 (44%) 27 (44%) ··
Second trimester fetal
ultrasonography
11 (34%) 22 (35%) ··
Third trimester fetal
ultrasonography
3 (9%) 5 (8%) ··
Report of last menstrual period 1 (3%) 5 (8%) ··
Capurro method
22
3 (9%) 3 (5%) ··
Brain imaging fi ndings‡
Abnormal 11/27
(41%)
0··
Normal 16/27
(59%)
62 (100%) ··
Not done 5§ 0 ··
Data are n (%). PRNT=plaque reduction neutralisation test. *For gestational age and
sex in the Fenton preterm growth chart.
29
†Defi ned as birthweight lower than the
10th percentile for gestational age and sex in the Fenton growth chart. ‡Cases
assessed by CT scan and controls by ultrasonography. §One stillbirth, three neonatal
deaths, and one case in intensive care at the time of analysis.
Table 1: Characteristics of mothers and neonates
See
Online for appendix
Articles
1360
www.thelancet.com/infection Vol 16 December 2016
None of the controls had a positive result by either
RT-PCR with serum samples (none of 62 controls tested)
or detection of Zika virus-specifi c IgM in serum (none of
59 tested). 13 (41%) of 32 cases tested positive for Zika
virus (RT-PCR or Zika virus-specifi c IgM) in cerebrospinal
fl uid or serum, and nine (28%) cases tested positive in
serum only (table 2). No cross-reactivity with dengue
virus-specifi c IgM was seen in samples positive for Zika
virus-specifi c IgM.
The overall crude OR for microcephaly and laboratory-
confi rmed Zika virus infection was 55·5 (95% CI 8·6–∞),
and was similar when adjusted for maternal education or
maternal age (table 3). The magnitude of the association
in cases with brain abnormalities detected by CT scan
was very strong: OR 113·3 (95% CI 14·5–∞). In cases
with normal fi ndings on brain imaging the association
was still strong and signifi cant: OR 24·7 (95% CI 2·9–∞).
When results in serum alone were considered, the crude
OR was 31·7 (95% CI 4·7–∞), and similar after
adjustment. The magnitude of the association in cases
with brain abnormalities on CT scan was very strong,
OR 80·9 (95% CI 10·2–∞), but the association in cases
with normal fi ndings on brain imaging was not
signifi cant (OR 3·9, 95% CI 0·1–∞).
Four of 13 cases with laboratory-confi rmed Zika virus
infection had normal fi ndings on brain imaging. Of the
six cases whose mothers were seronegative to Zika virus
(PRNT
50
), fi ve also tested negative (to Zika virus-specifi c
IgM or by RT-PCR in serum or cerebrospinal fl uid
samples), and one had a positive RT-PCR result.
There was good agreement between serum and
cerebrospinal fl uid test results for detection of Zika virus-
specifi c IgM (κ 0·91, 95% CI 0·74–1·00). Zika virus-
specifi c IgM was detected in cerebrospinal fl uid of nine
(36%) of 25 cases (seven cases were not tested) and in
serum of eight (27%) of 30 cases (two were not tested);
the geometric mean titre was 23·62 (95% CI 18·8–29·6)
in cerebrospinal fl uid and 16·8 (95% CI 12·3–23·0) in
serum. No neonates or mothers had IgM antibodies
specifi c for toxoplasma, cytomegalovirus, or rubella in
serum. A high proportion of mothers had IgG in serum
(toxoplasma-specifi c IgG: 17 [53%] of 32 mothers of cases
vs 27 [44%] of 62 mothers of controls; cytomegalovirus-
specifi c IgG: 28 [88%] vs 47 [76%]; rubella-specifi c IgG: 20
[63%] vs 46 [74%]). These proportions did not diff er
signifi cantly between mothers of cases and mothers of
controls (data not shown).
Discussion
This preliminary analysis shows a strong association
between microcephaly and laboratory confi rmation of
Zika virus infection by RT-PCR or Zika virus-specifi c
IgM in cerebrospinal fl uid or serum of neonates. The
risk was high in cases with brain abnormalities detected
by imaging, but was also present in cases without brain
abnormalities. Results of RT-PCR or Zika virus-specifi c
IgM were positive only in neonates with microcephaly
and were negative in serum of all neonates in the control
group. M
ore than half of neonates with microcephaly
had normal fi ndings on brain imaging. There was very
good agreement of Zika virus-specifi c
IgM-positive
results in serum and cerebrospinal fl uid of neonates.
To our knowledge, our study is the fi rst to estimate the
seroprevalence of Zika virus infection in pregnant
women in an epidemic area in Brazil. The high Zika
virus PRNT
50
seropositivity (64%) in mothers of controls
indicates high frequency of Zika virus infection in this
population in Recife. Similar frequencies of Zika virus
infection were reported in the general population in Yap
island
30
and in French Polynesia after the outbreaks in
these regions.
31
We cannot determine with any degree of
Cases (n=32) Controls (n=62) p value
RT-PCR or Zika virus-specifi c IgM (cerebrospinal fl uid or serum)
Positive 13 (41%) 0 <0·0001
Negative* 19 (59%) 62 (100%)
RT-PCR or Zika virus-specifi c IgM (serum)
Positive 9 (28%) 0 <0·0001
Negative 23 (72%) 62 (100%)
Data are n (%). *For one stillbirth and one neonatal death RT-PCR was tested in
macerated tissues.
Table 2: Results based on RT-PCR or specifi c IgM for Zika virus in
cerebrospinal fl uid or serum samples for cases and in serum samples
for controls
Cases Controls Odds ratio
(95% CI)
Serum or cerebrospinal fl uid samples
All cases
Crude 13/32 (41%) 0/62 55·5 (8·6–∞)
Adjusted for maternal
education
·· ·· 59·2 (9·0–∞)
Adjusted for maternal age ·· ·· 55·6 (8·5–∞)
Cases categorised by brain imaging fi ndings*
Abnormal 7/11 (64%) 0/62 113·3 (14·5–∞)
Normal 4/16 (25%) 0/62 24·7 (2·9–∞)
Serum samples only
All cases
Crude 9/32 (28%) 0/62 31·7 (4·7–∞)
Adjusted for maternal
education
·· ·· 38·5 (5·5–∞)
Adjusted for maternal age ·· ·· 30·2 (4·5–∞)
Cases categorised by brain imaging fi ndings*
Abnormal 6/11 (55%) 0/62 80·9 (10·2–∞)
Normal 1/16 (6%) 0/62 3·9 (0·10–∞)
Data are number of cases positive for Zika virus by RT-PCR or Zika virus-specifi c
IgM/total number of patients (%), unless otherwise indicated. *Odds ratios in
these subgroup analyses are crude because of small numbers; brain imaging was
not done in fi ve cases (one stillbirth, three neonatal deaths, and one case in
intensive care at the time of analysis).
Table 3: Association between microcephaly and laboratory confi rmation
of Zika virus infection