Content uploaded by Dorit Lev
Author content
All content in this area was uploaded by Dorit Lev on Nov 22, 2015
Content may be subject to copyright.
Accepted Article
Microcephaly at birth - the accuracy of three references for fetal head
circumference. How can we improve prediction?
Z. Leibovitz*, E. Daniel-Spiegel†, G. Malinger‡§, K. Haratz¶, M. Tamarkin¶, L.
Gindes¶, L. Ben-Sira**, D. Lev††, I. Shapiro*, H. Bakry*, B. Weizman*, A. Zreik*,
S. Egenburg‡‡, A. Arad‡‡, R. Tepper§§, D. Kidron¶¶, T. Lerman-Sagie¶
*Ultrasound unit, Department of Obstetrics and Gynecology, Bnai Zion Medical
Center, Haifa, Israel; †Department of Obstetrics and Gynecology, Haemek Medical
Center, Afula, Israel; ‡Lis Maternity Hospital, Division of OB-GYN Ultrasound, Tel-
Aviv Sourasky Medical Center, Tel Aviv, Israel; §Sackler Faculty of medicine, Tel-
Aviv University, Tel-Aviv, Israel; ¶Unit of Fetal Neurology and Prenatal Diagnosis,
Department of Obstetrics and Gynecology, Wolfson Medical Center, Holon, Israel;
**Department of Radiology, Tel Aviv Medical Center, Tel Aviv, Israel; ††Genetics
Institute, Edith Wolfson Medical Center, Holon, Israel; ‡‡Department of pathology,
Bnai Zion Medical Center, Haifa, Israel; §§Ultrasound unit, obstetrics and
gynecology, Meir Medical Center, Kfar-Saba, Israel; ¶¶ Department of pathology,
Meir Medical Center, Kfar Saba, Israel
Keywords: fetal microcephaly, prenatal diagnosis, head circumference
Corresponding author: Dr Z. Leibovitz, Ob/Gyn department, Bnai-Zion Medical
Center, P.O. Box 4940, Haifa 31048, Israel; (e-mail: zvi_l@yahoo.com,
zvi.leibovitz.dr@gmail.com)
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which
may lead to differences between this version and the Version of Record. Please cite this
article as doi: 10.1002/uog.15801
This article is protected by copyright. All rights reserved.
Accepted Article
Abstract
OBJECTIVE: To evaluate the ability to predict microcephaly at birth (MICB) by
using the conventional and two new prenatal references for fetal head circumference
(HC). To assess whether integrating additional parameters can improve prediction.
METHODS: Microcephaly in utero was defined as a fetal head circumference 3
standard deviations (SD) below the mean for gestational age according to Chervenak
et al's reference (CR). The records of cases with fetal microcephaly (FMIC) were
evaluated for the medical history, imaging findings, biometry, and postnatal
examination/autopsy findings. Microcephaly was confirmed at birth by an
occipitofrontal circumference (OFC) or a brain weight at autopsy 2SD below the
mean for gestational age. The new INTERGROWTH-21
st
Project (I21P)
and a recent
Israeli reference (IR) for fetal growth were applied for evaluation of the FMIC
positive predictive value (PPV) for diagnosis of cases with microcephaly at birth
(MICB). Optimal HC cut-offs were determined for each of the new references aimed
at detection of all MICB cases and minimizing the number of false positive ones
found with a normal head circumference at birth (NHCB). We assessed: the PPV of
the FMIC for MICB diagnosis, the difference between the z-scores of the prenatal HC
and the corresponding OFC at birth, the frequency of growth restriction, decreased
HC/abdominal circumference (AC) and HC/femur length (FL), the percentage of
associated malformations, and family history.
RESULTS: Forty two fetuses were diagnosed as having FMIC according to the CR.
In only 24 microcephaly was confirmed at birth or by autopsy (PPV 57.1%). The
optimal I21P and IR HC cut-offs for MICB diagnosis were the mean-3SD and the
mean-2.3SD, resulting in a statistically non-significant PPV improvement of 61.5%
and 66.7%, respectively. The presence of family history of MIC, IUGR, associated
malformations and application of stricter HC cut-offs resulted in a higher prediction
rate of MICB, although not statistically significant. The deviation of the HC from the
mean, by all references, was significantly larger compared to the actual deviation of
the OFC at birth. The mean differences between the corresponding z-scores of the HC
and OFC were -1.15, -1.95, and -0.74 for the CR, I21P, and IR, respectively. The
estimated weight of the cases with FMIC was below the 10
th
and 3
rd
percentile in
83.3% and 42.9%, correspondingly. A birth weight below the 3
rd
percentile was found
This article is protected by copyright. All rights reserved.
Accepted Article
in 34.8% of MICB and in 11.1% of NHCB (p<0.05). A low HC/AC and a low HC/FL
were found in 37.5% and 50% of MICB, and 27.8% and 55.6% NHCB cases,
respectively. Associated anomalies were found in 58.3% of MICB compared to 27.8%
of NHCB.
CONCLUSIONS: The evaluated references, all result in considerable over diagnosis
of fetal microcephaly. The use of the two new HC references did not significantly
improve MICB prediction compared to Chervenak et al's one. The OFC deviates
significantly less from the mean compared to the HC. By addition of a family history,
associated anomalies, IUGR, and stricter HC cut-offs we could improve the predictive
accuracy. We suggest that addition of the difference between the HC and OFC z-
scores, developed in this study, to a particular HC z-score will enable better
estimation of the actual OFC deviation at birth.
This article is protected by copyright. All rights reserved.
Accepted Article
Introduction
Microcephaly (MIC) is frequently associated with intellectual disability and
neurologic abnormalities. The definition of MIC after birth is non-uniform. Usually it
is defined as an occipitofrontal circumference (OFC) more than 2 standard deviations
(SD)
1
below the mean for age and gender, but some authors put the OFC cut-off at -
3SD
2
. Assuming a normal OFC distribution, 2.3% of children would be defined as
microcephalic based on an OFC of 2SD below the mean. However, published
estimates for this threshold at birth are lower (0.56%
3
and 0.54%
4
). For an OFC of
3SD below the mean only 0.1% of children would be diagnosed with MIC, which
corresponds to the published estimate of 0.14% of neonates
4
. The incidence of
intellectual disability correlates with the number of SDs the OFC deviates below the
mean: 11% and 51% for 2SD and 3SD, correspondingly
4
.
Similarly, the diagnosis of fetal microcephaly (FMIC) also relies on the measurement
of an abnormally small fetal head circumference (HC). There are many published
nomograms for fetal HC growth. Although new HC references have recently been
developed, the established ones are still commonly used, despite being based on
relatively small groups of fetuses and outdated measurement modalities
5,6
.
The yield of the commonly used growth charts for prenatal MIC diagnosis is
considered low. The inaccuracy is primarily related to the inconsistent HC
measurement methodology and the absence of properly designed studies aimed at
optimization of predictive strategies for prenatal diagnosis of MIC. Furthermore, the
sonographic HC z-score (expressed as a number of standard deviations below the
mean for gestational age) is consistently over-estimated relative to the corresponding
postnatal OFC z-score. In a study of fetuses with HC z-scores between -2 and -3, 90%
were found to be normocephalic at birth
7
.
In 1984 Jeanty et al developed HC growth standards based on a longitudinal
assessment of 45 normal pregnancies of medical personnel volunteers in New Haven,
Connecticut
5
. This reference was applied to fetuses with suspected microcephaly in
two studies by Chervenak et al
8,9
. The studies included a total of 40 fetuses. In only 13
MIC was confirmed at birth. Their established cut-off for FMIC diagnosis (HC z-
score ≤ -3SD) was associated with a PPV of 46%
8
and 50%
9
for a false negative rate
of zero and 15%, correspondingly. As a result of the low PPV numerous fetuses may
This article is protected by copyright. All rights reserved.
Accepted Article
be wrongly diagnosed as microcephalic; and in countries where termination of
pregnancy is an option, it may be inadequately offered.
Our study's main goals were to assess the accuracy of the established and new HC
nomograms in prediction of microcephaly at birth (MICB) and to assess whether
integrating additional parameters can improve prediction.
Materials and Methods
The records of cases with FMIC diagnosed between 2007 and 2014 were collected
from the registries of four Israeli medical centers. The inclusion criteria were:
available records of singleton pregnancies with a sonographically documented
gestational age in the first or early second trimester, a HC 3SD below the mean for
gestational age according to Chervenak et al's reference
8
(CR), OFC and weight at
birth or autopsy results.
Examinations were performed with Voluson E8, Voluson 730 Expert, and Voluson
730 Pro (GE Healthcare Ultrasound, Milwaukee, WI, USA) ultrasound machines. All
records were evaluated for the medical history, imaging results, laboratory data, and
postnatal examination or autopsy findings. Microcephaly was confirmed at birth by
an OFC 2SD below the mean
10
or a brain weight of less than 2SD on autopsy
11
.
The positive predictive values (PPV) of FMIC for diagnosis of microcephaly at birth
were calculated using CR and the two new fetal growth nomograms
12,13
.
The new nomograms differed in their methodology. The first, INTERGROWTH-21st
Project (I21P), was intended to produce prescriptive fetal growth standards in
singleton pregnancies by studying a worldwide cohort of young, educated, affluent,
adequately nourished, and clinically healthy women of low risk for adverse pregnancy
outcome
12
. The second one was a large Israeli population based reference (IR) of
singleton gestations that excluded cases of an uncertain gestational age, fetal
malformations, and known fetal syndromes
13
. This reference provided descriptive
growth standards without limitations of maternal age and health risk factors. In both
new references the HC was measured using an electronic ellipse facility outlining the
outer cranial borders, whereas Chervenak et al
8
calculated the HC by using the
biparietal diameter (BPD) as an outer-to-inner cranial measurement and the
This article is protected by copyright. All rights reserved.
Accepted Article
occipitofrontal diameter (OFD) as a distance between the middle of the bone echoes
applying a specific formula: HC=1.62(BPD+OFD)
5
(Figure 1).
To keep with the conventional MIC assessment, the z-score method was used for
evaluation of the fetal head size. Although originally based on the quantile regression
analysis, the IR was adjusted for HC z-score calculation (by developing of HC mean
and SD values according to gestational week). The developed table, formulas for the
HC and SD, and a HC z-score calculator are provided in the supplementary material.
Optimal HC cut-offs were determined for each HC reference aimed at detection of all
MICB cases and minimizing the number of false positive ones found with a normal
head circumference at birth (NHCB).
The categorical parameters for integration with HC cut-offs included: associated
IUGR (defined as an estimated fetal weight (EFW) below the 10
th
percentile), fetal
anomalies, and a family history of MIC.
The biometric parameters were: the EFW calculated according to Hadlock et al
14
(based on BPD, HC, abdominal circumference (AC), and femur length (FL)); fetal
and newborn weight percentiles according to Dolberg et al
15
; and HC/AC and HC/FL
according to Snijders et al
16
.
The accuracy of the HC and integrated criteria for diagnosis of microcephaly at birth
were assessed by the PPV and the false negative rate (FNR) for each of the studied
references. The PPV was calculated as a number of the MICB cases divided by a
number of the MICB and NHCB cases, all fulfilling a specified condition; and the
FNR as the number of MICB cases not fulfilling the specified condition divided by all
MICB cases. Statistical comparison between the PPV and FNR of the studied criteria
was performed by χ
2
test.
The HC z-scores according to the applied references were compared in each study
group (FMIC, MICB, and NHCB) by t-test for paired samples.
A comparison between the HC z-scores of the MICB and NHCB groups was
performed by t-test for unpaired samples.
This article is protected by copyright. All rights reserved.
Accepted Article
In each case the difference between the corresponding HC and OFC z-scores was
calculated according to the applied references and correlated to the time interval
between FMIC diagnosis and OFC assessment at birth.
The data was analyzed by IBM SPSS statistics software (IBM Corporation Software
Group, NY, US).
This study was approved by the Institutional Review Boards.
This article is protected by copyright. All rights reserved.
Accepted Article
Results
Using Chervenak et al's reference
8
a fetal head circumference 3SD below the mean
for gestational age was found in 42 fetuses (FMIC), of them at birth 24 were
diagnosed with microcephaly (MICB) and 18 had a normal HC (NHCB). The mean
gestational week at FMIC diagnosis was 32.6 ± 5.6 (range: 22 to 38). The mean
gestational week did not significantly differ between the MICB and the NHCB
subgroups (33.7 ± 5.3 and 31.0 ± 5.7, correspondingly). The female/male ratio of
FMIC was 1/1.14. A karyotype was obtained in 27 fetuses and was normal in all.
Serology for intrauterine infection was negative in all (including 2 cases of
histologically confirmed CMV fetopathy). Associated fetal anomalies were found in
45.2%, 58.3%, and 27.8% of FMIC, MICB, and NHCB cases, correspondingly.
The definite or possible etiology for microcephaly was determined in 23 of 24 MICB
cases: primary MIC (3), malformation of cortical development (4), pontocerebellar
hypoplasia (2), vermian hypoplasia (1), CMV fetopathy (2), hypoxic/hemorrhagic
brain damage (2), fetal alcohol syndrome (1), placental pathology (4), syndromic MIC
(3), and Aicardi-Goutières syndrome (1).
The etiology for small HC was established in 16 of 18 NHCB cases: Chiari-II
malformation with open spina bifida (4), vertical cranial elongation (5, coronal
craniosynostosis (1) and skull molding (4)), and placental pathology (2).
Microcephaly at birth was confirmed in only 24 fetuses according to Chervenak et al
8
(HC below mean-3SD) resulting in PPV of 57.1%. The optimal HC cut-offs for I21P
and IR were: mean-3SD and mean-2.3SD, correspondingly. These cut-offs resulted in
a PPV of 61.5% and 66.7%, respectively. Application of the 1
st
quantile cut-off of the
IR (the lowest reported quantile in the original paper
13
) provided a PPV of 65.7% with
a FNR of 4.2%. The PPV of the optimal HC cut-offs did not differ significantly
between the references (χ
2
test).
Table 1 summarizes the accuracy of the applied references for MICB diagnosis using
stricter and optimal HC cut-offs with integration of additional parameters (IUGR,
fetal malformations, and family history). Both, CR and IR achieved a PPV of 100%
for MICB at a HC < mean-4SD, whereas I21P did not exceed the PPV of 66.7% even
at a HC< mean-6SD. Integration of the optimal HC cut-offs with the presence of
IUGR, fetal anomalies, or family history had an additive effect on MICB prediction
This article is protected by copyright. All rights reserved.
Accepted Article
for all three references. Adding IUGR resulted in a PPV of 62.9%, 66.7%, and 73.3%
for EFW below the 10
th
percentile and 66.7%, 70.6%, and 75% below the 3
rd
percentile for the CR, I21P, and IR, respectively. Integration of the optimal HC cut-
offs with the presence of fetal anomalies resulted in a MICB PPV of 70%, 73.7%, and
82.4% for CR, I21P, and IR, correspondingly. Addition of a family history of MIC
resulted in a PPV of 100% for each reference. A comparison between the PPVs of the
above mentioned criteria did not reach statistical significance for each of the
references (explained by the small size of the groups). Similarly, the PPVs did not
differ significantly between the references, tested for the corresponding criteria (χ
2
test). As a rule, stricter HC cut-offs and integrated conditions were associated with a
high FNR of MICB for all references, reaching a maximum of 87.5% for the optimal
HC cut-off integrated with family history of MIC (p<0.0001, compared to a FNR of
an optimal HC cut-off alone; χ
2
test).
The mean HC z-scores in FMIC, MICB, and NHCB groups are presented in Table 2.
The I21P reference showed significantly lower HC z-scores compared to the other
two references, tested in each of the study groups (p<0.0001, t-test for paired
samples).
When the HC z-scores of the MICB fetuses were evaluated by CR and IR, they were
significantly lower compared to the NHCB's ones (p<0.04 and p<0.02 for CR and IR,
correspondingly; t-test for unpaired samples).
The HC z-scores of FMIC, MICB, and NHCB groups were significantly lower
(p<0.005, paired t-test) than the corresponding OFC ones for all prenatal references,
except for the MICB group according to IR (Table 3). The mean differences between
the HC and OFC z-scores of FMIC fetuses were: -1.15 (95% CI: -1.51 to -0.79), -1.95
(95% CI: -2.47 to -1.44), and -0.74 (95% CI: -1.07 to -0.40) for CR, I21P, and IR,
respectively.
The time interval between FMIC diagnosis and OFC assessment at birth was 2.4±4.1
weeks ranging between 1day and 17.6 weeks. No significant correlation was found
between the difference of the corresponding HC and OFC z-scores and the time
interval between the two assessments (Pearson's correlation coefficients were -0.05, -
0.3, and -0.15 for CR, I21P, and IR, respectively).
This article is protected by copyright. All rights reserved.
Accepted Article
The EFW of the FMIC cases was below the 10
th
and 3
rd
percentile in 83.3% and
42.9%, respectively. A birth weight below the 10
th
percentile was found in 78.3% and
33.3% of MICB and NHCB cases, and below the 3
rd
percentile in 34.8% and 11.1%,
correspondingly. A low birth weight was significantly more frequent in MICB cases
compared to NHCB (p<0.05).
A HC/AC <5
th
percentile was found in 33.3%, 37.5%, and 27.8% of FMIC, MICB,
and NHCB cases; and a HC/FL <5
th
percentile in 52.4%, 50%, and 55.6%,
respectively.
Termination of pregnancy was performed in 20 FMIC cases upon parental request
(47.6%): 11 and 9 fetuses of the MICB and NHCB groups, correspondingly.
Abnormal CNS finding on autopsy were found in 63.6% of MICB and in 33.3% of
NHCB (all with Chiari-II malformation). Six autopsies of fetuses from the NHCB
group were anatomically normal.
Neurological abnormalities were reported in 84.6% and 11.1% (p=0.01) of MICB and
NHCB live births, correspondingly. Death during the first 7 years occurred in 53.8%,
and 0% of MICB, and NHCB cases, respectively. Eight of the 9 patients from the
NHCB group had normal development.
This article is protected by copyright. All rights reserved.
Accepted Article
Discussion
In a recent update on a classification of malformations of cortical development by
Barkovich et al
17
, microcephaly is categorized according to the onset of the disorder
relative to the glial and neuronal migration phase. The premigrational category
includes multiple genetic disorders that manifest by fetal and neonatal microcephaly
due to reduced proliferation or excessive apoptosis of neuronal and glial cells. The
post-migrational category is related to decreased brain growth during late gestation or
the early postnatal period as a result of ischemia, infection, trauma, inborn metabolic
disorders, teratogens, and genetic factors. The last category develops in the first two
years of life, however, prenatal diagnosis may be possible, if the deceleration in head
growth is progressive and occurs relatively early in the third trimester. Patients with
isolated genetic MIC (primary MIC) are almost always born with a significantly small
OFC, but are not necessarily considered microcephalic at birth. Syndromic MIC
(associated with dysmorphism and/or brain and/or systemic malformations) is usually
diagnosed postnatally
18
.
FMIC, defined as a HC 3SD below the mean
8
, can be expected in 0.1% of
pregnancies, if the HC is normally distributed. However, the actual estimate of such
low HC measurements is probably higher. According to the Israeli registry
13
of the
11169 singleton pregnancies with certain dating and no fetal anatomical
malformations or known syndromes, 23 (0.2%) were found to have FMIC, while only
13 of them were diagnosed as microcephalic postnatally, resulting in a population
estimate of a false positive rate of 0.1%.
The correct prediction of MICB based on prenatal biometry is suboptimal
19
. In our
series, the established reference
8
resulted in over-diagnosis of FMIC in 43% of cases,
leading to erroneous termination of 6 pregnancies with apparently normal fetuses.
The use of the two new HC references
12,13
did not significantly improve MICB
prediction compared to Chervenak et al's one
8
(explained by a relatively small series).
The low predictive accuracy can also be related to the nomogram's statistical
properties: a mean HC that is too high for the actual population or a SD value that is
too small, may cause substantial changes in the z-score calculation. Figure 2
demonstrates that the mean HC growth curves, before 32-33 gestational weeks, are
very similar for all three references, whereas the nomogram used in CR
5
shows a
This article is protected by copyright. All rights reserved.
Accepted Article
progressive HC increment in later gestational weeks, relative to the I21P and IR.
Since 60% of our FMIC cases were diagnosed after 33 weeks, it can partially explain
the higher percentage of false positive cases according to CR.
The I21P data showed significantly lower HC z-scores compared to the other
references (Table 2) resulted by the low SD values of the I21P nomogram (Figure 2).
The extremely deviated HC z-scores were found in 6 NHCB cases (-3.9 to -6.6) and
in 4 MICB ones (-4.3 to -6.4), thus explaining why the I21P nomogram did not
exceed a PPV of 66.7% even with very strict HC cut-offs (Table 1).
A PPV of 100% was achieved for the HC cut-off below mean-4SD using both the CR
and IR nomograms. Adding a family history of MIC to the optimal HC cut-offs also
resulted in a PPV of 100% using all three references. Integrating either the presence
of IUGR or the association of fetal anomalies with the optimal HC cut-off by the IR
gave a maximal PPV of 75% and 82.4%, correspondingly (Table 1). Our series is
relatively small due to the rarity of a measurement of a HC< mean-3SD. Therefore,
the additive prediction of the stricter HC cut-offs and the integration of additional
parameters did not reach statistical significance. However, we suggest that
considering these parameters in every case of FMIC can improve the accuracy of
MICB prediction, although this improvement is associated with significantly high
false negative rates using all studied references (Table 1). This indicates that in many
FMIC cases the absence of integrated parameters will preclude accurate diagnosis of
MICB.
The difference between intrauterine and postnatal measurements of head size prevents
accurate MICB prediction (Figure 1). Sonographic HC measurements are consistently
underestimated relative to postnatal OFC and the difference increased with gestational
age
20
. The probable cause for the discrepancy is that the HC is measured as the
perimeter of the fetal skull while the OFC includes the scalp and hair. However,
differences in the anatomical landmarks for the HC and OFC assessment, molding,
scalp edema, interobserver and intraobserver variability of the HC measurements
21
may also affect the magnitude of the inconsistency. Our study demonstrates that in
addition to the multifactorial discrepancy between the HC and OFC measurements,
the deviation of the HC from the mean was significantly larger compared to the actual
deviation of the OFC at birth for all references (Table 3).
This article is protected by copyright. All rights reserved.
Accepted Article
No information exists on the correlation between the severity of FMIC and the risk for
intellectual disability; the only established risks are based on the degree of the
postnatal OFC deviation
4,22
. Therefore, only achievement of accurate prediction of the
OFC z-score at birth will enable precise prenatal prognostication. Based on
calculations in our study group we can provide confidence interval limits of the
difference between the HC and OFC z-scores (Table 3). These limits can be added to
a particular HC z-score in order to predict the range of the OFC one. The accuracy of
this new approach should be confirmed in future studies.
FMIC frequently presents with IUGR. In a study on congenital microcephaly detected
by prenatal ultrasound, 66% of the MICB cases were small for gestational age
23
.
IUGR is described in multiple disorders of both pre- and postmigrational
microcephaly
17,21,24
. FMIC in such cases can be misinterpreted as part of general fetal
growth restriction without taking into account a small head as an alarming prognostic
factor. Our study also confirms that MICB is strongly associated with IUGR,
indicating that restricted fetal growth cannot be considered as a "simple" explanation
for a small HC. According to our and den Hollander et al's results
23
a normal HC/AC
is found in the majority of FMIC cases, thus indicating that a disproportionately small
HC is not a typical feature of fetal microcephaly.
Recently a novel statistical approach allowing accurate individualized prediction of
late fetal and neonatal growth outcomes was proposed by Deter et al
25,26
. This method
is based on serial sonographic assessments of multiple biometric parameters (starting
from 18 gestational weeks and continuing all through pregnancy at 3-4 week
intervals). Unfortunately such a systematic follow-up is not always available in FMIC
cases which are usually first diagnosed in the third trimester. Moreover, the accuracy
of this approach for prediction of MICB needs to be assessed.
Placental pathology was an infrequent explanation for the association of MICB and
IUGR in our study (16.7%, 4/24). In contrast, Dahlgren et al
27
found that 79% of
microcephalic newborns had abnormal placental findings. However, it is not clear if
this was the sole pathology or part of a more complex disorder.
Our study showed no significant improvement in predicting MICB by application of
the new HC references compared to the established one. By addition of a family
history, associated anomalies, IUGR, and stricter HC cut-offs we could improve the
This article is protected by copyright. All rights reserved.
Accepted Article
predictive accuracy, however this improvement was not statistically significant due to
the small study groups. It is important to realize that in many MIC cases there are no
additional parameters to consider and therefore the ability to accurately predict
postnatal microcephaly remains limited. We found significantly better OFC z-scores
compared to the corresponding HC ones. We suggest that addition of the confidence
interval limits of the difference between the HC and OFC z-scores to a particular HC
z-score will enable an improved estimation of the actual OFC z-score at birth and
more precise counseling in cases with FMIC.
This article is protected by copyright. All rights reserved.
Accepted Article
Study limitations:
Our study was designed to predict a pathologically small OFC at birth without
considering the postnatal clinical outcome. The study was based on a relatively small
cohort. The correlation between the degree of fetal HC smallness and the incidence of
MIC during the postnatal years was not addressed. The correlation between the
severity of FMIC and intellectual disability was not studied.
This article is protected by copyright. All rights reserved.
Accepted Article
Table 1. Accuracy of fetal HC cut-offs and integrated parameters in predicting
microcephaly at birth according to the applied references
Predictive criterion
Positive predictive value
% (cases
#
)
False negative rate
% (cases
†
)
Chervenak et al
8
HC≤mean-3SD 57.1 (24/42) NA*
HC≤mean-3SD and EFW<10th% 62.9 (22/35) 8.3 (2/24)
HC≤mean-3SD and EFW<3rd% 66.7 (12/18) 50.0 (12/24)
HC≤mean-3SD and fetal anomalies 70.0 (14/20) 41.7 (10/24)
HC≤mean-3SD and familial MIC 100.0 (3/3) 87.5 (21/24)
HC≤mean-4SD 100.0 (8/8) 66.7 (16/24)
Papageorghiou et al
12
HC≤mean-3SD 61.5 (24/39) 0.0 (0/24)
HC≤mean-3SD and EFW<10th% 66.7 (22/33) 8.3 (2/24)
HC≤mean-3SD and EFW<3rd% 70.6 (12/17) 50.0 (12/24)
HC≤mean-3SD and fetal anomalies 73.7 (14/19) 41.7 (10/24)
HC≤mean-3SD and familial MIC 100.0 (3/3) 87.5 (21/24)
HC≤mean-4SD 65.4 (17/26) 29.2 (7/24)
HC≤mean-5SD 66.7 (8/12) 66.7 (16/24)
HC≤mean-6SD 66.7 (4/6) 83.3 (20/24)
Daniel-Spiegel et al
13
HC≤mean-2.3SD 66.7 (24/36) 0.0 (0/24)
HC≤mean-2.3SD and EFW<10th% 73.3 (22/30) 8.3 (2/24)
HC≤mean-2.3SD and EFW<3rd% 75.0 (12/16) 50.0 (12/24)
HC≤mean-2.3SD and fetal anomalies 82.4 (14/17) 41.7 (10/24)
HC≤mean-2.3SD and familial MIC 100.0 (3/3) 87.5 (21/24)
HC<1st quantile 65.7 (23/35) 4.2 (1/24)
HC≤mean-3SD 71.4 (15/21) 37.5 (9/24)
HC≤mean-4SD 100.0 (4/4) 83.3 (20/24)
HC - head circumference; SD - standard deviation; EFW - estimated fetal weight
* NA - not applicable (this cut-off was the inclusion criterion).
#
Numbers in parentheses indicate the MICB cases / (MICB+NHCB) cases, all
fulfilling the specified condition.
†
Numbers in parentheses indicate the MICB cases not fulfilling the specified
condition / all MICB cases.
This article is protected by copyright. All rights reserved.
Accepted Article
Table 2. HC z-scores in the study groups according to the applied references
HC - head circumference; SD - standard deviation;
FMIC - fetal microcephaly; MICB - microcephaly at birth; NHCB - normal head
circumference at birth.
The HC z-scores of each study group were compared according to the applied
references using the t-test for paired samples. The HC z-scores by the I21P were
significantly lower compared to ones by the CR and IR (p<0.0001
*
) in all study
groups. No statistically significant difference was found comparing the HC z-scores
according to the CR and IR in each of the groups.
The HC z-scores of the MICB and NHCB cases were compared by t-test for unpaired
samples using each reference. The MICB cases had significantly lower HC z-scores
compared to the NHCB's ones using CR and IR (p<0.04
†
and p<0.02
¶
,
correspondingly), while the z-scores did not significantly differ for the I21P (p=0.3
‡
).
Reference
HC z-score
FMIC (42 cases)
mean (SD)
HC z-score
MICB (24 cases)
mean (SD)
HC z-score
NHCB (18 cases)
mean (SD)
Chervenak et al
8
(CR) -3.56 (0.65) -3.74 (0.75)
†
-3.33 (0.37)
†
Papageorghiou et al
12
(I21P) -4.4 (1.12)
*
-4.56 (1.03)
*‡
-4.19 (1.24)
*‡
Daniel-Spiegel et al
13
(IR) -3.16 (0.9) -3.45 (0.97)
¶
-2.78 (0.65)
¶
This article is protected by copyright. All rights reserved.
Accepted Article
Table 3. Difference between the corresponding HC
*
and OFC
†
z-scores in the study
groups according to the applied references
HC - head circumference; OFC - occipitofrontal circumference; SD - standard
deviation; CI - confidence interval; FMIC - fetal microcephaly; MICB - microcephaly
at birth; NHCB - normal head circumference at birth.
*
Assessed at FMIC diagnosis
†
Assessed at birth
‡
95% confidence intervals of the z-score differences are provided for the suggested
estimation of the actual OFC z-score at birth (see Discussion).
The differences between the corresponding HC and OFC z-scores were examined in
each study group comparing the z-scores according to the prenatal HC
nomograms
8,12,13
and the postnatal OFC reference
10
; (
#
p<0.005,
¶
p=0.31; paired t-test).
Compared references
Z-score difference
FMIC (42 cases)
mean (SD)
Z-score difference
MICB (24 cases)
mean (SD)
Z-score difference
NHCB (18 cases)
mean (SD)
Chervenak et al
8
vs. Fenton et al
10
-1.15 (1.14)
#
(95% CI: -1.51 to -0.79)
‡
-0.52 (1.0)
#
-1.95 (0.75)
#
Papageorghiou et al
12
vs. Fenton et al
10
-1.95 (1.62)
#
(95% CI: -2.47 to -1.44)
‡
-1.24 (1.45)
#
-2.86 (1.37)
#
Daniel-Spiegel et al
13
vs. Fenton et al
10
-0.74 (1.06)
#
(95% CI: -1.07 to -0.40)
‡
-0.18 (0.84)
¶
-1.44 (0.88)
#
This article is protected by copyright. All rights reserved.
Accepted Article
References
1. Opitz JM, Holt MC. Microcephaly: general considerations and aids to
nosology. J Craniofac Genet Dev Biol. 1990;10(2):175-204.
2. Passemard S, Kaindl AM, Verloes A. Microcephaly. Handb Clin Neurol.
2013;111:129-41.
3. Vargas JE, Allred EN, Leviton A, Holmes LB. Congenital microcephaly:
phenotypic features in a consecutive sample of newborn infants. J Pediatr
2001;139:210-4.
4. Dolk H. The predictive value of microcephaly during the first year of life for
mental retardation at seven years. Dev Med Child Neurol 1991;33:974-83.
5. Jeanty P, Cousaert E, Hobbins JC, Tack B, Bracken M, Cantraine F. A
longitudinal study of fetal head biometry. Am J Perinatol. 1984;1:118-28.
6. Hadlock FP, Deter RL, Harrist RB, Park SK. Fetal head circumference:
relation to menstrual age. AJR Am J Roentgenol. 1982;138:649-53.
7. Stoler-Poria S, Lev D, Schweiger A, Lerman-Sagie T, Malinger G.
Developmental outcome of isolated fetal microcephaly. Ultrasound Obstet
Gynecol. 2010;36:154-8.
8. Chervenak FA, Jeanty P, Cantraine F, Chitkara U, Venus I, Berkowitz RL,
Hobbins JC. The diagnosis of fetal microcephaly. Am J Obstet Gynecol.
1984;149:512-7.
9. Chervenak FA, Rosenberg J, Brightman RC, Chitkara U, Jeanty P. A
prospective study of the accuracy of ultrasound in predicting fetal
microcephaly. Obstet Gynecol. 1987;69:908-10.
10. Fenton TR, Kim JH. A systematic review and meta-analysis to revise the
Fenton growth chart for preterm infants. BMC Pediatr. 2013;13:59.
11. Maroun LL, Graem N. Autopsy standards of body parameters and fresh organ
weights in nonmacerated and macerated human fetuses. Pediatr Dev Pathol.
2005;8:204-17.
12. Papageorghiou AT, Ohuma EO, Altman DG, Todros T, Cheikh Ismail L,
Lambert A, Jaffer YA, Bertino E, Gravett MG, Purwar M, Noble JA, Pang R,
Victora CG, Barros FC, Carvalho M, Salomon LJ, Bhutta ZA, Kennedy SH,
Villar J; International Fetal and Newborn Growth Consortium for the 21st
Century (INTERGROWTH-21st). International standards for fetal growth
This article is protected by copyright. All rights reserved.
Accepted Article
based on serial ultrasound measurements: the Fetal Growth Longitudinal
Study of the INTERGROWTH-21st Project. Lancet. 2014;384:869-79.
13. Daniel-Spiegel E, Weiner E, Yarom I, Doveh E, Friedman P, Cohen A, Shalev
E. Establishment of fetal biometric charts using quantile regression analysis. J
Ultrasound Med. 2013;32:23-33.
14. Hadlock FP, Harrist RB, Sharman RS, Deter RL, Park SK. Estimation of fetal
weight with the use of head, body, and femur measurements--a prospective
study. Am J Obstet Gynecol. 1985;151:333-7.
15. Dollberg S, Haklai Z, Mimouni FB, Gorfein I, Gordon ES. Birth weight
standards in the live-born population in Israel. Isr Med Assoc J. 2005;7:311-4.
16. Snijders RJ, Nicolaides KH. Fetal biometry at 14-40 weeks' gestation.
Ultrasound Obstet Gynecol. 1994;4:34-48.
17. Barkovich AJ, Guerrini R, Kuzniecky RI, Jackson GD, Dobyns WB. A
developmental and genetic classification for malformations of cortical
development: update 2012. Brain. 2012;135:1348-69.
18. Seltzer LE, Paciorkowski AR. Genetic disorders associated with postnatal
microcephaly. Am J Med Genet C Semin Med Genet. 2014;166:140-55.
19. Malinger G, Lev D, Lerman-Sagie T. Assessment of fetal intracranial
pathologies first demonstrated late in pregnancy: cell proliferation disorders.
Reprod Biol Endocrinol. 2003;1:110
20. Melamed N, Yogev Y, Danon D, Mashiach R, Meizner I, Ben-Haroush A.
Sonographic estimation of fetal head circumference: how accurate are we?
Ultrasound Obstet Gynecol. 2011;37:65-71.
21. Sarris I, Ioannou C, Chamberlain P, Ohuma E, Roseman F, Hoch L, Altman
DG, Papageorghiou AT; International Fetal and Newborn Growth Consortium
for the 21st Century (INTERGROWTH‐21st). Intra- and interobserver
variability in fetal ultrasound measurements. Ultrasound Obstet Gynecol.
2012;39:266-73.
22. Ashwal S, Michelson D, Plawner L, Dobyns WB; Quality Standards
Subcommittee of the American Academy of Neurology and the Practice
Committee of the Child Neurology Society. Practice parameter: Evaluation of
the child with microcephaly (an evidence-based review): report of the Quality
Standards Subcommittee of the American Academy of Neurology and the
This article is protected by copyright. All rights reserved.
Accepted Article
Practice Committee of the Child Neurology Society. Neurology. 2009;73:887-
97.
23. den Hollander NS, Wessels MW, Los FJ, Ursem NT, Niermeijer MF,
Wladimiroff JW. Congenital microcephaly detected by prenatal ultrasound:
genetic aspects and clinical significance. Ultrasound Obstet Gynecol.
2000;15:282-7.
24. Berkley E. and Abuhamad A.Sonographic dating and standard fetal biometry
(Chapter 8). In: Queenan JT, Spong CY, Lockwood CJ, eds. Queenan's
Management of High-Risk Pregnancy: An Evidence-Based Approach, 6th
Edition. Chichester, MO: Wiley-Blackwell; 2012:63-74.
25. Deter RL, Lee W, Sangi-Haghpeykar H, Tarca AL, Yeo L, Romero R.
Individualized fetal growth assessment: critical evaluation of key concepts in
the specification of third trimester size trajectories. J Matern Fetal Neonatal
Med. 2014;27:543-51
26. Deter RL, Lee W, Sangi-Haghpeykar H, Tarca AL, Yeo L, Romero R. Fetal
growth cessation in late pregnancy: its impact on predicted size parameters
used to classify small for gestational age neonates. J Matern Fetal Neonatal
Med. 2015;28:755-65
27. Dahlgren L, Wilson RD. Prenatally diagnosed microcephaly: a review of
etiologies. Fetal Diagn Ther. 2001;16:323-6
This article is protected by copyright. All rights reserved.
Accepted Article
Fig
u
cir
c
Le
g
Pa
n
ref
e
an
o
dist
by
a
Pa
n
out
e
ref
e
Pa
n
u
re 1. Asse
s
c
umference
g
end
n
el (a) dem
o
e
rence
8
(de
v
o
uter-to-in
n
ance betwe
a
formula:
H
n
el (b) sho
w
e
r cranial b
o
e
rences
12,13
.
n
el (c) illust
r
s
sment of f
e
(OFC)
o
nstrates th
e
v
eloped by
J
n
er cranial
m
en the mid
d
H
C=1.62(B
P
w
s HC meas
u
o
rders. Thi
s
r
ates the po
e
tal head ci
r
e
method o
f
J
eanty et al
5
m
easureme
n
d
le of the b
o
P
D+OFD).
u
rement us
i
s
method o
f
stnatal OF
C
r
cumferenc
f
HC assess
m
5
). The bip
a
n
t and the o
c
o
ne echoes.
i
ng an elect
r
f
HC assess
m
C
measure
m
e (HC) and
m
ent used i
n
a
rietal diam
e
c
cipitofron
t
The head
c
r
onic ellips
m
ent was a
p
m
ent includi
n
neonatal o
c
n
Cherven
a
e
ter (BPD)
t
al diamete
r
c
ircumferen
e
facility o
u
p
plied in b
o
n
g the scal
p
c
cipitofron
t
a
k et al's
is measure
d
r
(OFD) as
a
ce is calcul
u
tlining the
o
th new HC
p
and hair.
t
al
d
as
a
ated
This article is protected by copyright. All rights reserved.
Accepted Article
Fig
u
acc
o
Le
g
CR
,
HC
,
Th
e
ind
i
u
re 2. The
g
o
rding to t
h
g
end
,
Chervena
k
,
head circ
u
e
curves of
t
i
cated line
t
g
rowth cur
v
h
e applied r
e
k
et al
8
; I21
P
u
mference;
S
t
he HC and
ypes.
v
es of mean
e
ferences
P
, Papageo
r
S
D, standa
r
SD of eac
h
head circu
m
r
ghiou et al
1
r
d deviatio
n
h
reference
a
m
ference a
n
1
2
; Daniel-
S
n
.
a
re illustrat
e
n
d standard
S
piegel et a
l
e
d accordi
n
deviation
l
13
.
n
g to the
This article is protected by copyright. All rights reserved.