Beruflich Dokumente
Kultur Dokumente
, PROQUAD
& ZOSTAVAX
liti-
gation. Paxil
or generic ondansetron.
Appendix A. Supplementary data
Additional further reading sources related to this article can be
found online at http://dx.doi.org/10.1016/j.siny.2013.09.007.
References
[1] Rasmussen SA, Olney RS, Holmes LB, Lin AE, Keppler-Noreuil KM, Moore CA.,
National Birth Defects Prevention Study. Guidelines for case classication for
the National Birth Defects Prevention Study. Birth Defects Res A Clin Mol
Teratol 2003;67:193e201.
[2] Yoon PW, Rasmussen SA, Lynberg MC, Moore CA, Anderka M, Carmichael SL,
et al. The National Birth Defects Prevention Study. Public Health Rep
2001;116(Suppl. 1):32e40.
[3] Anderka M, Mitchell AA, Louik C, Werler MM, Hernndez-Daz S,
Rasmussen SA., National Birth Defects Prevention Study. Medications used to
treat nausea and vomiting of pregnancy and the risk of selected birth defects.
Birth Defects Res A Clin Mol Teratol 2012;94:22e30.
[4] Cogswell ME, Bitsko RH, Anderka M, Caton AR, Feldkamp ML, Hockett
Sherlock SM, et al. Control selection and participation in an ongoing,
A.R. Scialli / Seminars in Fetal & Neonatal Medicine 19 (2014) 170e176 175
population-based, caseecontrol study of birth defects: the National Birth
Defects Prevention Study. Am J Epidemiol 2009;170:975e85.
[5] Hill AB. The environment and disease: association or causation? Proc R Soc
Med 1965;58:295e300.
[6] US Department of Health, Education, and Welfare. Smoking and health.
Report of the advisory committee to the Surgeon General of the Public
Health Service. Publication No. 1103. Washington, DC: US Government
Printing Ofce; 1964.
[7] Shepard TH. Catalog of teratogenic agents. 10th ed. Baltimore: Johns Hopkins
University Press; 2001xxiv.
[8] Brent RL. Bendectin: review of the medical literature of a comprehensively
studied human nonteratogen and the most prevalent tortogen-litigen. Reprod
Toxicol 1995;9:337e49.
[9] Public Affairs Committee. Causation in teratology-related litigation. Birth
Defects Res (Part A) 2005;73:421e3.
[10] Einarson A, Maltepe C, Navioz Y, Kennedy D, Tan MP, Koren G. The safety of
ondansetron for nausea and vomiting of pregnancy: a prospective compara-
tive study. Br J Obstet Gynaecol 2004;111:940e3.
[11] Asker C, Wikner BN, Klln B. Use of antiemetic drugs during pregnancy in
Sweden. Eur J Clin Pharmacol 2005;61:899e906.
[12] Pasternak B, Svanstrom H, Hviid A. Ondansetron in pregnancy and risk of
adverse fetal outcomes. N Engl J Med 2013;368:814e23.
[13] Czeizel AE. Specied critical period of different congenital abnormalities: a
new approach for human teratological studies. Congenit Anom (Kyoto)
2008;48:103e9.
[14] Shimizu M, Ohta T, Kato M, Kobayashi Y, Yamashita Y, Asano M, et al.
Reproduction study (Seg. II) of ondansetron hydrochloride in rats by intra-
venous route. Yakuri To Chiryo 1992;20(Suppl. 4):155e74 [English abstract].
[15] Ezaki H, Yokoyama S, Takahashi N, Takamatsu M, Tokado H, Takeda K.
Reproduction study (seg II) of ondansetron hydrochloride in rabbits by oral
route. Yakuri to Chiryo 1992;20:s1187e96 [English abstract].
[16] Zeitz KP, Guy N, Malmberg AB, Dirajlal S, Martin WJ, Sun L, et al. The 5-HT3
subtype of serotonin receptor contributes to nociceptive processing via a novel
subset of myelinatedandunmyelinatednociceptors. J Neurosci 2002;22:1010e9.
[17] Sullivan CA, Johnson CA, Roach H, Martin RW, Stewart DK, Morrison JC. A pilot
study of intravenous ondansetron for hyperemesis gravidarum. Am J Obstet
Gynecol 1996;174:1565e8.
[18] Nageotte MP, Briggs GG, Towers CV, Asrat T. Droperidol and diphenhydramine
in the management of hyperemesis gravidarum. Am J Obstet Gynecol
1996;174:1801e5.
A.R. Scialli / Seminars in Fetal & Neonatal Medicine 19 (2014) 170e176 176
Review
Birth defects and assisted reproductive technologies
Joe Leigh Simpson
*
March of Dimes, 1275 Mamaroneck Avenue, White Plains, NY 10605, USA
Keywords:
Assisted reproductive technologies (ART)
Birth defects
Hypospadias
Imprinting
Intracytoplasmic sperm injection (ICSI)
In-vitro fertilization (IVF)
s u m m a r y
Assisted reproductive technologies (ART) using in-vitro fertilization (IVF) account for w1% of births in
the USA and as much as 3e4% in Europe or Australia. Initially studies involved infants prospectively
examined in an early cohort of US births, with salutary results. Later studies began to show the frequency
of birth defects to be increased. In meta-analysis, odds ratio was >1.0, with the 95% condence limit not
extending to <1.0. Although ART are associated with a 30% increase in birth defects; subfertile couples
achieving pregnancy without ART show a 20% increase. It thus appears that the increase in birth defects
is due less, if at all, to ART protocols per se than to the biological perturbations that generated the
infertility that necessitated ART to achieve pregnancy. There is consensus that traditional IVF and
intracytoplasmic sperm injection (ICSI)/IVF show the same overall risk notwithstanding increased sex
chromosome abnormalities in both procedures and increased hypospadias in ICSI. No other organ system
seems disproportionately affected. There is no additive risk in ART twins compared with non-ART twins,
nor in embryos having been cryopreserved. The increased risk observed had not appeared to dissuade
couples from attempting to have their own children.
2014 Published by Elsevier Ltd.
1. Introduction
Assisted reproductive technologies (ART) using in-vitro fertil-
ization (IVF) account for about 1% of births in the USA and as much
as 3e4% in Europe or Australia. More than ve million babies are
estimated worldwide to have been born through ART. Scientic and
medical advances are increasing pregnancy rates, and the preva-
lence of ART births can condentially be expected to continue to
increase. Even prior to the pioneering efforts of Edwards and
Steptoe that resulted in the rst success in 1978, concern had been
raised over whether infants born by ART would universally be
abnormal. This fear has largely been mitigated, and attention is
now focused on more nuanced questions.
The initial studies involved infants carefullyexamined inanearly
cohort of US births, withresults salutaryalthoughsample sizes were
small [1]. Population-based studies from Australia by Lancaster [2]
showed a 2.9% frequency, reassuringly that expected for the gen-
eral population. The rst largepopulationstudy was byWestergaard
et al. [3], who compared 2245 ART births in 1994e1995 with 2245
controls. The odds ratio (OR) was 1.04 with the 95% condence in-
terval (CI) 0.78e1.39 (non-signicant). Other studies of this era
generally failed to show deleterious effects (Table 1) [3e5]. Studies
reported in the 2000s beganto showfrequency of birth defects to be
increased (Table 2). The OR in these studies was often >1.0, with the
95% CI not extending across the 1.0 isobar [6e15]. Thus, there
evolved the present consensus that increased birth defects are
indeed positively associated with ART. The major question at pre-
sent is whether this increase is due toART protocols per se or merely
reective of the biological perturbations that generated the infer-
tility that necessitated ART to achieve pregnancy.
In this communication, we shall rst consider the frequency of
congenital anomalies in offspring of ART pregnancies, stratied by
specic subgroups depending on technique. We shall also comment
on the pitfalls that make difcult denitive conclusions concerning
the explanation for anomalies detected.
2. Frequency of anomalies in ART pregnancies
Almost 50 cohort studies have addressed the question of
anomalies in ART pregnancies, as referenced elsewhere [16e19].
Initial studies naturally focused on IVF alone, because not until the
mid-1990s could male infertility be managed by intracytoplasmic
sperm injection (ICSI) followed by traditional IVF. In the 1990s and
early 2000s, one could condently conclude the overall risk was
not, say, a two- or three-fold increase above the accepted popula-
tion baseline of 2e3%. However, more precise statements could not
be made.
Notwithstanding general reassurance, this author and others
pointed out pitfalls that could sway opinions on safety in either
* Tel.: 1 914 997 4610.
E-mail address: jsimpson@marchofdimes.com.
Contents lists available at ScienceDirect
Seminars in Fetal & Neonatal Medicine
j ournal homepage: www. el sevi er. com/ l ocat e/ si ny
1744-165X/$ e see front matter 2014 Published by Elsevier Ltd.
http://dx.doi.org/10.1016/j.siny.2014.01.001
Seminars in Fetal & Neonatal Medicine 19 (2014) 177e182
direction [20,21]. It was in particular appreciated that power to
detect a signicantly increased frequency of anomalies was not
possible due to small sample size.
Table 2 shows later population-based studies on which this
altered conclusion began to be based [6e15]. All were based on
registries that recorded cohorts of births, from which those with
anomalies could be stratied. A caveat is that the denition of birth
defects among registries varies, usually involving International
Classication of Diseases (ICD)-based diagnosis codes not well
suited to distinguish major from minor anomalies. ICD codes at
birth are not robust in detecting subtle anomalies not externally
visible, nor in excluding inherited syndromes or chromosomal
anomalies. Neither is plausibly due to ART. Not always followed is
the pragmatic, accepted, denition of a major birth defect as one
that causes death, functional impairment or (if structural) requires
surgery.
Reviewing several studies will sufce as illustrative. In 2005
Klemetti et al. [7] published Finnish registry of IVF and ICSI cases
delivered in 1996e1999, nding the adjusted OR for all birth de-
fects to be 1.31 (95% CI: 1.10e1.57), when comparing 4459 cases and
27 078 controls. In 2007, Pinborg et al. [9] used Danish registry data
to derive an OR of 1.24 (95% CI: 1.09e1.43). In 2005 Kallen et al.
[6,12] published an analysis of Swedish registries (1982e1999) data
and in 2010 analysed cases for the years 2001e2007. The latter
study showed OR of 1.25 (95% CI: 1.5e1.37) based on 15 570 cases
[12].
Outside of Scandinavia, the region generating greatest attention
is Australia, betting the countrys sentinel role in developing ART.
In 2002 Hansen et al. reported Western Australia cases, reprising
this in 2012 [14,22]. Their 2012 report of Western Australia registry
data involved ART cases delivered 1994e2002, an interval of sig-
nicance given many changes having since occurred in laboratory
methodology. This study was laudatory, however, in trying to take
into account anomalies in pregnancy terminations. Both the 2911
ART cases and the 210 997 non-ART cases were followed for 6
years. A major birth defect was found in 8.7% of ART cases and 5.4
of non-ART cases (OR: 1.53; 95% CI: 1.30e1.79). Rates were not
different in unlike sex twins (obligatory dizygosity) (OR: 1.08; 95%
CI: 0.77e1.51). The 5.4% birth defects in the control group is at odds
with general acceptance, absent comprehensive laboratory as-
sessments such as array comparative genomic hybridization. These
authors concluded also that there has been a decrease in the
prevalence of birth defects compared with their earlier cohort
(1994e1998) [14,23]. One explanation may simply be greater
appreciation between major and minor anomalies. Plausible sci-
entic reasons include changes in culture media, better regulation
of heat and CO
2
in incubators, and differential ovulation stimula-
tion regimes.
A second Australian group extending interval of ascertainment
to 5 years was that of Davies et al. [13], who used registry data of
308 974 births in South Australia (Adelaide). Minor defects were
excluded unless they required treatment or were disguring.
Among 6163 ART offspring were 513 with anomalies (8.4%),
compared with 5.8% in non-ART births. The adjusted OR was 1.28
(95% CI: 1.16e1.41). OR was 1.26 for IVF alone, and 1.77 for ICSI/IVF.
It would be of interest to knowthe temporal sequence in which the
cumulative absolute frequency of birth defects occurred, i.e.
ascertainment at birth versus later.
Of special interest to US readers is the report of Kelley-Quon
et al. [15], who used the California Patient Discharge Linked Birth
Cohort Database that lists anomalies by ICD-9 codes. This study
involved 4795 infants born in 2006e2007 after ART, compared with
46 025 naturally conceived in the same interval. The overall rate of
major congenital abnormalities was 9.0% vs 6.6% (OR: 1.25; 95% CI:
1.12e1.39; P < 0.001). Incidence of birth defects this high bespeaks
inclusion of many minor anomalies. In this authors opinion, this
alone casts doubt on conclusions that ORs were signicantly
increased for certain organ-specic anomalies: eye, head and neck,
heart and genitourinary track.
These limitations notwithstanding, meta-analyses reached
similar arithmetic conclusions: Rimm et al. [16]: 1.29 (95% CI:
1.01e1.67); Hansen et al. [17,19], risk ratio: 1.32 (95% CI: 1.24e
Table 1
Earlier population-based studies of anomalies in assisted reproductive technology pregnancies.
Study Years of sample accrual Adjusted OR (95% CI) Statistical signicance Country
Dhont et al. [4] 1986e2002 1.25 (0.96e1.64) No Belgium
Westergaard et al. [3] 1994e1995 1.04 (0.78e1.39) No Sweden
Anthony et al. [5] 1995e1996 1.03 (0.86e1.23) No Netherlands
OR, odds ratio; CI, condence interval.
The initial population-based studies depend on data in registries. Ascertainment varied between studies with respect to duration of time in which anomalies were sought and
denition of major defects. The interval of accumulated ART registry cases does not in any report correspond to the extant laboratory and ovulation stimulation protocols used
in 2014.
Table 2
Later population-based studies on anomalies in assisted reproductive technology (ART) pregnancies (2005e2013).
Study Years of sample accrual Adjusted OR (95% CI) Statistical signicance Country
Kallen et al. [6] 1982e2001 1.44 (1.32e1.57) Yes Sweden
Davies et al. [13] 1986e2002 1.24 (1.09e1.41) Yes Australia (Adelaide)
Halliday et al. [11] 1991e2004 1.36 (1.19e1.55) Yes Australia (Parkville)
Hansen et al. [14] 1994e2002 1.53 (1.30e1.79) Yes Australia (Perth)
Pinborg et al. [9] 1995e2000 1.24 (0.97e1.58) No Denmark
Klemetti et al. [7] 1996e1999 1.31 (1.10e1.57) Yes Finland
Ombolet et al. [8] 1997e2003 1.11 (0.08e1.58) No Belgium
Kallen et al. [12] 2001e2007 1.25 (1.15e1.37) Yes Sweden
Kelley-Quon et al. [15] 2006e2007 1.25 (1.21e1.39) Yes USA (California)
Fuji et al. [10] 2006 1.17 (0.81e1.69) No Japan
OR, odds ratio; CI, condence interval.
These population-based studies depend on data in registries. Ascertainment varied with respect to duration of time in which anomalies were sought and denition of major
defects. The interval of accumulated ART registry cases does not in any report correspond to the extant laboratory and ovulation stimulation protocols used in 2013.
J.L. Simpson / Seminars in Fetal & Neonatal Medicine 19 (2014) 177e182 178
1.42); Wen et al. [18]: 1.37 (95% CI: 1.26e1.48) (Table 3). Similar
results are not surprising because the same studies were generally
analyzed.
3. Anomalies by organ system
Because of sample size limitations, individual anomalies have
generally been assessed only by pooling anomalies by organ sys-
tem. Yet within a given system there is a plethora of different
anomalies of difcult etiologies egenetic and otherwise. Collapsing
all such anomalies to assess a single organ system is convenient for
generating statistical power, but this does not obviate the meth-
odological ow.
Hypospadias is the only organ system in which OR has been
consistently shown to be increased. An odds ratio of 3.0 (95% CI:
1.09e6.50) was reported by Wennerholmet al. [23] and 1.5 (95% CI:
1.0e2.1) by Ericson and Kallen [24]. Klemetti et al. [7] also showed
increased hypospadias.
A potential explanation for increased hypospadias in ART is
heritable low testosterone, also adversely affecting spermatogen-
esis. Recurrence risk for hypospadias is 2e5% in rst degree rela-
tives (sons). This could lead to transmission of a polygenic trait
(hypospadias). There was no signicant difference in the meta-
analysis by Wen et al. [18] One confounder is that indications for
ICSI are now much less stringent than when ICSI was initially
introduced, including many cases analyzed by Davies (1986e2002)
[13]. ICSI now constitutes half of all ART cases in the USA and even
more in some venues.
The only other organ system seriously implicated is cardiac.
Sala et al. [25] found a signicant increase, but in general other
authors have found no increase. Special pitfalls exist in assessing
cardiac anomalies. A patent ductus arteriosus is normal at birth,
but soon closes in full-term infants; in premature infants persis-
tence is normal, and surgical ligation is routine. Ventricular septal
defects are not usually evident at birth but manifest about one
week after birth when cardiac circulation is adapted for postnatal
life. Taking these temporal changes into account is not facile, and
beyond the charge of most if not all registries recording birth
defects.
4. Methods of data analysis
Determining whether anomalies are increased in ART could be
based on data from a single center or from a registry. It is fash-
ionable for many to use pooled data (systematic reviews or meta-
analysis), and consider this approach superior.
Meta-analysis is based mostly on obvious strength of sample
size compared with single center studies. Center-specic studies
are thus marginalized. However, the latter can utilize checklist
comparisons by a limited pool of examiners, an approach not
possible in population registries. Indeed, it is studies with less
systematic assessments that more often showed no increase in
birth defects. Meta-analysis has other limitations, namely lack of
consensus on inclusion criteria. In meta-analysis the axiom is that
all studies are comparable, and thus suitable for pooling.
Consider, however, the laudatory reports of Hansen et al. [17,19]
The latter exhaustive report excluded non-English reports, re-
ports in which no birth defects were found in either cases or
controls, and reports in which birth defects ascertainment
extended to a later age. Numbers of studies deemed suitable were
45, similar to Wen et al. (n 46) [18]. Yet Rimm et al. [16]
accepted only 19 studies.
Although the validity of meta-analysis in many circumstances
is strong, this is more arguable in ART and birth defects. In ART
virtually no units utilize similar protocols. Yet many different
laboratories contribute to a single registry, to say nothing of
differences between registries. Different authors must reach
their own opinion as to whether preset criteria are met. De-
cisions on inclusion criteria may be discrepant at different points
in time, even by the same author. For example, upon reassess-
ment 8 years later Hansen et al. [17,19] concluded that many
studies accepted as valid in their 2006 study were no longer
acceptable.
One widely used decision tree is to include or exclude series in
which follow-up extended beyond the newborn period. (Separate
analysis would be more appropriate.) Thus excluded is the report
of Davies et al. [13], despite being an excellent population-based
study that extended to 5 years of age and contained ascertain-
ment from multiple sources. Another study excluded for the same
reason has been Bonduelle et al. [27] who showed increased sex
chromosomal abnormalities in ICSI. These cannot be detected in
registries because physical abnormalities are usually not evident at
birth.
In some meta-analyses it is sometimes decided that certain
studies are given higher quality scores and, hence, greater
weight. For example, Hansen et al. [19] assigned a quality 1 score
(best) to 11 of 13 registry-based studies, and only seven
population-based studies were not scored quality 1. Among 32
clinic samples, only one was assigned quality 1. Many of the latter
were indeed of insufcient sample size, but numbers alone should
not be the sole criterion for high quality because quality of clinical
examination is less than ideal in registries. Consistent clinical
acumen is not possible among all examiners, some of whom lack
familiarity in distinguishing major from minor anomalies. It is
difcult but not impossible to form a collaborative network of
clinical centers to assess anomalies consistently among centers,
but not impossible (see the reports from the South American
Registry ECLAMC [27]). Another problemwith registry data is long
interval during which cases were accumulated. Anomalies were
initially probably more vigorously sought in ART cases than con-
trols [20,21]. This bias is probably less so now, but the earlier bias
is still relevant because registry data being used for currently
published meta-analyses are often based on deliveries even
extending as far back as the 1980s. Paradoxically, the larger the
study (greater power), the more ancient and now atypical cases
included.
The above criticisms are not meant to denigrate work pursued to
unravel a difcult problem, but to remind the reader that crisp data
suitable for analysis do not exist.
5. Pitfalls in assessing presence of birth defects
This author has expressed concern on pitfalls precluding data
accumulated and used for comparison of ARToutcomes [20,21]. Box
1 lists common pitfalls so far discussed, a few aspects of which are
highlighted in the following subsections.
Table 3
Reported meta-analyses on birth defects and assisted reproductive technologies.
Study Year Studies accepted OR (95% CI)
Rimm et al. [16] 2004 19 1.29 (1.01e1.67)
Hansen et al. [17] 2005 25
a
1.29 (1.21e1.37)
Wen et al. [18] 2012 46 1.37 (1.26e1.48)
Hansen et al. [19] 2013 45
a
1.32 (1.24e1.42)
OR, odds ratio; CI, condence interval.
a
Many studies considered acceptable in the 2005 review by Hansen et al. [17]
were no longer considered acceptable in their 2013 report [19].
J.L. Simpson / Seminars in Fetal & Neonatal Medicine 19 (2014) 177e182 179
5.1. Denition of anomalies
The major source of confusion is probably inconsistent deni-
tion of an anomaly (birth defect). How precisely does one distin-
guish major from minor anomalies? Should one rely on studies in
which minor anomalies are collapsed into a single category with
major anomalies? In fact, criteria for distinguishing major from
minor anomalies are well known to geneticists and well codied,
but generally not familiar to clinicians outside medical genetics
[28]. Rigor of physical examination is an issue because geneticists
record more minor anomalies than non-genetic pediatricians.
Relying upon pooled data when the primary outcome (birth de-
fects) is not assessed identically is inherently hazardous. Examiners
not surprisingly have variable skills, training and interest; thus,
criteria cannot be assumed to be comparable across centers,
studies, or countries. This becomes immediately evident by noting
variable rates of anomalies in controls, namely signaled by rates
considerably higher than the accepted 2e4% major birth defects at
birth and shown in checklist-based studies [29].
5.2. ART protocols
There is a lack of conformity in ovulation stimulation regimens,
a plethora of different culture media, and multiple embryo incu-
bation techniques. Standardization does not exist, and commer-
cially utilized media list only components and not concentrations.
5.3. Interval of pregnancy ascertainment
Especially egregious is, as noted, assuming that results derived
from methods used long ago are still generalizable to the
contemporary patient, a concern raised earlier. The most contem-
porary study involves 2006e2007 cycles, well behind currently
used techniques like vitrication [15]. The long interval between
pregnancies analyzed and publication data makes generalizability
to women undergoing ART in 2014 arguable.
5.4. Interval for anomaly surveillance after birth
Another problemis the varying length of surveillance after birth.
Birth defects registries traditionally are restricted to anomalies
ascertained in the neonatal period (before hospital discharge).
However, several excellent single-center studies extend to 5 years,
providing information not possible with registries (e.g. sex chro-
mosomal abnormalities) [13,26]. Certain organ-specic ART-related
anomalies could be manifested only in the newborn period and
other only later. Data analyzed in meta-analysis should ideally be
stratied by length of surveillance.
5.5. Lack of true control group
What is the control group? In theory, the proper control should
consist of women who require ART but who become pregnant
without ART, ideally within the time dened as normal to become
pregnant (12 months). Obviously, this group does not exist. Thus,
no gold standard randomized clinical trials or caseecontrol study
can be conducted. Instead, published reports compare ART out-
comes to naturally conceived offspring. The closest to a proper
control group seems to be a subfertile couple (inability to conceive
by 12 months) who conceive naturally while awaiting ART. Other
options include same-sex couples using a surrogate, but sample
sizes are inadequate.
6. Anomalies in singleton versus twin gestation
Consensus exists that overall birth defects are increased in sin-
gletons, but the same cannot be said for multiple gestations of
which twins constitute by far the largest group. When adjustment
was made for zygosity by Hansen et al. [19], OR was 1.26 (95% CI:
0.9e1.60). Anomaly rates were not different between ART and non-
ART twins. In unlike sex twins (obligatory dizygosity), OR was 1.08
(95% CI: 0.77e1.51). Recall that twins naturally conceived have long
been known to show increased birth defects, particularly if
monozygotic. Prior to the era of ART, half of all like-sexed twins
were monozygotic. Now the percentage should be lower. Interest-
ingly, pretermbirth in ARToffspring is considered to be increased in
singleton births but not in twin gestations [30,31]. Perhaps infer-
tility (and, hence, likelihood of embryo implantation) is greater in
women whose embryos are more likely to be abnormal and, hence,
less likely to sustain two embryos. Or, perhaps, singletons had a co-
twin who underwent unrecognized demise (vanishing twin) that
nonetheless exerted a persistent deleterious effect, perhaps
involving the decidua [32].
In conclusion, anomalies in both non-ART twins and in ART
twins are increased compared with singleton controls, but an ad-
ditive risk in ART is surprisingly unproven at present.
7. Anomalies in ICSI/IVF versus IVF alone
Initially only IVF per se was assessed for association with birth
defects. As stated earlier, no statistically signicant association was
reported during the 1990s. With utilization in the 1990s of ICSI (i.e.
ART using ICSI followed by traditional IVF), increased birth defects
began to be reported. With ICSI the frequency of sex chromosomal
abnormalities appears increased in prenatal samples (amniocen-
teses) as well as hypospadias [7,23,24]. Otherwise, anomaly rates
have appeared similar in ICSI and non-ICSI IVF, rendering overall
anomaly rates not signicantly different. The exception was Davies
Box 1
Pitfalls in experimental design and interpretation in assessing
birth defects in assisted reproductive technology (ART) offspring.
A. Ascertainment
Including or not including anomalies in stillborns.
Including or not including anomalies in terminated
fetuses.
Including or not including anomalies detectable on
ultrasound but not externally visible at birth.
Varying duration of surveillance.
Varying aggressiveness of surveillance.
B. Denition of anomalies
Minor anomalies versus major anomalies.
Failure to utilize pre-set denitions and check lists (i.e.
minimizing individual observer variability).
Determining when a minor anomaly becomes major
(e.g. requiring surgery).
C. Methodology
Failure to stratify analyses into isolated anomalies
versus multiple anomalies (syndromic) disorders.
Failure to recognize (exclude) heritable disorders
whose etiology is not plausibly ART-related.
Generalization of conclusions to a contemporary
population when data were derived in a prior era that
used now-outdated protocols.
Application of meta-analysis when studies are of dis-
similar design (unavoidable given lack of standardized
protocols).
J.L. Simpson / Seminars in Fetal & Neonatal Medicine 19 (2014) 177e182 180
et al. [13], showing increased risk only in ICSI offspring (OR: 1.47;
95% CI: 1.15e1.89).
ICSI couples have the confounder of having increased fre-
quencies of balanced translocations, not only in males (2.05%) but
female (1.38%) sex chromosomal abnormalities, and inversions
were also found not infrequently. Thus, two subfertile individuals
presumably constitute a single infertile couple. The frequency of
balanced translocations in men undergoing ICSI is higher in oligo-
spermic (4%) than in azoospermic men (1%). By analogy the same
principle may hold for single mutant genes, certain men under-
going ICSI having a pleiotropic mutation one component of which is
infertility and another predisposition to a structural anomaly.
Chromosomal abnormalities as assessed by preimplantation ge-
netic diagnosis are increased in ICSI-assisted embryos [32].
In conclusion, sex chromosomal abnormalities and hypospadias
are considered increased in ICSI compared with IVF alone.
8. Cryopreservation
Data are limited concerning whether cryopreservation affects
birth defects in liveborns, but at present data are reassuring. Loss of
cell(s) from cleavage stage biopsy is clearly deleterious to embryo
survival, as shown by decreased pregnancy rates for embryos that
had loss of one cell upon thawing [33]. Similarly, removal of more
than two cells fromcleavage stage embryos reduced the pregnancy
rate [34]. These statements are not applicable for trophectoderm
biopsy where a dozen or more cells are removed. Fortunately, loss
of cells at either cleavage stage or blastocyst stage does not result in
liveborn anomalies because the all or none effect exists. Embryos
surviving to live birth are not affected long term because when loss
of a single cell occurs at this stage of embryogenesis the function of
that cell can be subsumed by another because all cells are totipo-
tential. In their own cohort Hansen et al. [22] claimed to observe a
trend favoring fresh rather than frozen singleton IVF; however, this
was not found in twins or ICSI pregnancies. Davies et al. [13] found
no differences at all.
This author counsels that neither benet nor harm exists for
cryopreservation with respect to birth defects.
9. Is imprinting an explanation for increased birth defects?
The biological basis for increased frequency of birth defects in
ART is unclear. Heritable single gene or polygenic factors conferring
infertility could have a pleiotropic effect on offspring, but this
cannot explain the magnitude of the increase.
There also seems little reason to postulate de-novo chromo-
somal arrangements in vitro or accumulation of the requisite
multiple mutations at multiple loci that would constitute a poly-
genic/multifactorial explanation for isolated structural anomalies.
A plethora of variables in ART technology (vicissitudes of culture
media; ovulation stimulation) could, however, perturb embryonic
differentiation through altered gene expression.
Given lack of obvious explanations, the biological mechanism
most often invoked is perturbations of imprinting. A myriad of
growth factors are included, and these alone give any biologists
pause. A disturbance in methylation precluding scheduled tran-
scription and translation would yield differences in gene products.
Some genes would be expressed that were not originally scheduled
to do so, and the converse.
In 2002, de Braun et al. [35] reported that seven cases with the
autosomal dominant BeckwitheWiedeman syndrome (BWS) were
associated with either IVF alone (n 2) or ART (n 5). In BWS the
usual molecular basis lies inoverexpressionof thepaternal allele; the
maternal allele is ordinarily not expressed in BWS or normal chil-
dren. This usuallyoccurs as result of uniparental (paternal) disomy. In
ART-related cases studied by de Braun et al. [35], however, pertur-
bation was usually due not to increased paternal expression due to
uniparental disomy (UPD), but rather to unscheduled expression of
the maternal allele normally quiescent (hypomethylation of LIT1). A
second study from the UK generated not dissimilar ndings [36].
Other data have been reviewed by Vermeiden et al. [37].
Hansen et al. [14] gathered data in a cohort of pooled known
imprinting disorders: BWS, RusseleSilver syndrome, PradereWilli
syndrome, pseudohypoparathroidism, UPD for chromosome 14,
maternal hypomethylation syndrome. There were four occurrences
in ART cases (0.2%) versus 30 in non-ART (0.01%); OR was 8.21 (95%
CI: 2.81e24.00). On the other hand, a study of Danish registries, for
selecting a list of purported imprinting disorders, found no increase
among ART cases [38]. This argues against imprinting as an
omnibus explanation for increased OR in ART offspring.
In conclusion, invoking imprinting due to ART protocols is a
popular and plausible explanation for the imprinting disorders
listed alone. However, imprinting seems an implausible explana-
tion for isolated structural anomalies (e.g. facial clefts, cardiac).
Moreover, if ART perturbs imprinting only in selected disorders, the
absolute increase will be low. Given an OR of 8, there would be little
arithmetic (absolute) risk in ART offspring, compared with the
general population. Thus, if the sole explanation for the increased
rate of birth defects in ART offspring is imprinting, this is either
disturbed in disorders not yet appreciated or the phenomenon acts
in ways not appreciated (e.g. enhancers in non-coding regions).
10. Subfertility, infertility and birth defects
The likely cause of increased frequency of birth defects in ART is
underlying subfertility [39]. The sentinel observation is that the
frequency of birth defects in infertile couples who belatedly (>12
months unprotected intercourse) achieve pregnancy without ART
is almost the same as that in couples requiring ART [40,41]. In the
study by Zhu et al. [40], OR was 1.17 (95% CI: 1.0e1.36) for infertile
couples who conceived naturally while awaiting ART. Criteria
consisted of requiring more than 12 months of unprotected sex to
succeed. Davies et al. [13] likewise reported increased birth defects
in offspring of women who were subfertile but who conceived
without ART (OR: 1.29; 95% CI: 0.99e1.68). Of special note, women
who had prior offspring with ART but who conceived spontane-
ously in a later cycle had offspring with increased birth defects (OR:
1.25; 95% CI: 1.01e1.56).
In conclusion subfertility is a major factor in, if not the sole
explanation for, ART-related birth defects. Rimm et al. [41] believe
that almost all the increased risk observed in ART offspring reects
selection bias of subfertile couples, who, without medical assis-
tance, would never achieve pregnancy. Hansen et al. [19] disagree,
and indeed one must continue to ferret out potential technology-
related causes. However, like Pinborg et al. [9] and Rimm et al.
[41], this writer believes that selection bias seems more likely.
11. Conclusion
Assisted reproductive technology is associated with a small (OR:
1.3) increase in birth defects. This should be communicated to pa-
tients prior to undergoing ART. The counselor may also wish to
communicate concern over pitfalls in data used to derive these
opinions, but must realize that perfection cannot be achieved in
experimental design because the ideal control group cannot be
constructed. It is often instructive to remind all couples contem-
plating pregnancy that the baseline anomaly rate is 2e3%,
compared with 3e4% in ART. The consensus to be communicated is
that both traditional IVF as well as ICSI/IVF show the same
increased risk, except for increased sex chromosome abnormalities
J.L. Simpson / Seminars in Fetal & Neonatal Medicine 19 (2014) 177e182 181
and hypospadias in ICSI. Otherwise, no particular organ system
seems disproportionately affected. No additive risk seems to exist
in ART twins compared with non-ART twins, nor in embryos pre-
viously cryopreserved. Overall, the increased risk observed e irre-
spective of etiology e seems unlikely to dissuade couples from
attempting to have their own children.
Conict of interest statement
None declared.
Funding sources
None.
References
[1] Morin NC, Wirth FH, Johnson DH, Frank LM, Presburg HJ, Van de Water VL,
et al. Congenital malformations and psychosocial development in children
conceived by in vitro fertilization. J Pediatr 1989;115:222e7.
[2] Lancaster PA. Registers of in-vitro fertilization and assisted conception. Hum
Reprod 1996;11(Suppl. 4):89e104. discussion 105e9.
[3] Westergaard HB, Johansen AM, Erb K, Andersen AN. Danish National In-Vitro
Fertilization Registry 1994 and 1995: a controlled study of births, malfor-
mations and cytogenetic ndings. Hum Reprod 1999;14:1896e902.
[4] Dhont M, De Sutter P, Ruyssink G, Martens G, Bekaert A. Perinatal outcome of
pregnancies after assisted reproduction: a caseecontrol study. Am J Obstet
Gynecol 1999;181:688e95.
[5] Anthony S, Buitendijk SE, Dorrepaal CA, Lindner K, Braat DD, den Ouden AL.
Congenital malformations in 4224 children conceived after IVF. Hum Reprod
2002;17:2089e95.
[6] Kallen B, Finnstrom O, Lindam A, Nygren KG, Olausson PO. In vitro fertilization
(IVF) in Sweden: risk for congenital malformations after different IVF
methods. Birth Defects Res A Clin Mol Teratol 2005;73:162e9.
[7] Klemetti R, Gissler M, Sevon T, Koivurova S, Ritvanen A, Hemminki E. Children
born after assisted fertilization have an increased rate of major congenital
anomalies. Fertil Steril 2005;84:1300e7.
[8] Ombolet W, Peeraer K, DeSutter P, Gerris J, Bosmans E, Martens G, et al.
Perinatal outcomes of ICSI pregnancies compared with a matched group of
natural conception pregnancies in Flanders (Belgium): a cohort study. Reprod
Biomed Online 2005;11:244e53.
[9] Pinborg A, Henningsen AK, Malchau SS, Loft A. Congenital anomalies after
assisted reproductive technology. Fertil Steril 2013;99:327e32.
[10] Fujii M, Matsouka R, Bergel E, van der Poel S, Okai T. Perinatal risk in singleton
pregnancies after in vitro fertilization. Fertil Steril 2010;94:2113e7.
[11] Halliday JL, Ukoumunne OC, Baker HW, Brehany S, Jacques AM, Garrett C, et al.
Increased risk of blastogenesis birth defects, arising in the rst 4 weeks of
pregnancy, after assisted reproductive technologies. Hum Reprod 2010;25:
59e65.
[12] Kallen B, Finnstrom O, Lindam A, Nilsson E, Nygren KG, Otterblad PO.
Congenital malformations in infants born after in vitro fertilization in Sweden.
Birth Defects Res A Clin Mol Teratol 2010;88:137e43.
[13] Davies MJ, Moore VM, Willson KJ, Van Essen P, Priest K, Scott H, et al.
Reproductive technologies and the risk of birth defects. N Engl J Med
2012;366:1803e13.
[14] Hansen M, Kurinczuk JJ, de Klerk N, Burton P, Bower C. Assisted reproductive
technology and major birth defects in Western Australia. Obstet Gynecol
2012;120:852e63.
[15] Kelley-Quon LI, Tseng CH, Janzen C, Shew SB. Congenital malformations
associated with assisted reproductive technology: a California statewide
analysis. J Pediatr Surg 2013;48:1218e24.
[16] Rimm AA, Katayama AC, Diaz M, Katayama KP. A meta-analysis of controlled
studies comparing major malformation rates in IVF and ICSI infants with
naturally conceived children. J Assist Reprod Genet 2004;21:437e43.
[17] Hansen M, Bower C, Milne E, de Klerk N, Kurinczuk JJ. Assisted reproductive
technologies and the risk of birth defects e a systematic review. Hum Reprod
2005;20:328e38.
[18] Wen J, Jiang J, Ding C, Dai J, Liu Y, Xia Y, et al. Birth defects in children
conceived by in vitro fertilization and intracytoplasmic sperm injection: a
meta-analysis. Fertil Steril 2012;97:1331e7. e1e4.
[19] Hansen M, Kurinczuk JJ, Milne E, de Klerk N, Bower C. Assisted reproductive
technology and birth defects: a systematic review and meta-analysis. Hum
Reprod Update 2013;19:330e53.
[20] Simpson JL. Registration of congenital anomalies in ART populations: pitfalls.
Hum Reprod 1996;11(Suppl. 4):81e8.
[21] Simpson JL, Liebaers I. Assessing congenital anomalies after preimplantation
genetic diagnosis. J Assist Reprod Genet 1996;13:170e6.
[22] Hansen M, Kurinczuk JJ, Bower C, Webb S. The risk of major birth defects after
intracytoplasmic sperm injection and in vitro fertilization. N Engl J Med
2002;346:725e30.
[23] Wennerholm UB, Bergh C, Hamberger L, Lundin K, Nilsson L, Wikland M, et al.
Incidence of congenital malformations in children born after ICSI. Hum
Reprod 2000;15:944e8.
[24] Ericson A, Kallen B. Congenital malformations in infants born after IVF: a
population based study. Hum Reprod 2001;16:504e9.
[25] Sala P, Ferrero S, Buf D, Pastorino D, Bertoldi S, Vaccari L, et al. Congenital
defects in assisted reproductive technology pregnancies. Minerva Ginecol
2011;63:227e35.
[26] Bonduelle M, Wennerholm UB, Loft A, Tarlatzis BC, Peters C, Henriet S, et al.
A multi-centre cohort study of the physical health of 5-year-old children
conceived after intracytoplasmic sperm injection, in vitro fertilization and
natural conception. Hum Reprod 2005;20:413e9.
[27] Castilla EE, Oriolo IM. ECLAMC: the Latin-American collaborative study of
congenital malformations. Community Genet 2004;7:76e94.
[28] Jones KL. Smiths recognizable patterns of human malformations. 7th ed.
Philadelphia: Saunders; 2013.
[29] Mills J, Knopp RH, Simpson JL, Javanovich-Peterson L, Metzger BE, Holmes LB,
et al. Lack of relation of increased malformation rates in infants of diabetic
mothers to glycemic control during organogenesis. N Engl J Med 1988;318:
671e6.
[30] Helmerhorst FM, Perquin DA, Donker D, Keirse MJ. Perinatal outcomes in
singletons and twins after assisted conception: a systematic review of
controlled studies. BMJ 2004;328:261.
[31] Jackson RA, Gibson KA, Wu YW, Croughan MS. Perinatal outcomes in single-
tons following in vitro fertilization: a meta-analysis. Obstet Gynecol
2004;103:551e63.
[32] Ponjaert-Kristoffersen I, Bonduelle M, Barnes J, Nekkebroeck J, Loft A,
Wennerholm UB, et al. International collaborative study of intracytoplasmic
sperm injection-conceived, in vitro fertilization-conceived, and naturally
conceived 5-year-old child outcomes: cognitive and motor assessments. Pe-
diatrics 2005;115:e283e9.
[33] Cohen J, Wells D, Musse S. Removal of 2 cells from cleavage stage embryos is
likely to reduce the efcacy of chromosomal tests that are used to enhance
implantation rates. Fertil Steril 2008;87:496e503.
[34] De Vos A, Staessen C, De Rycke M, Verpoest W, Haentjens P, Devroey P, et al.
Impact of cleavage-stage embryo biopsy in view of PGD on human blastocyst
implantation: a prospective cohort of single embryo transfer. Hum Reprod
2009;24:2988e96.
[35] DeBaun MR, Niemtz EL, Feinberg AP. Association of in vitro fertilization with
BeckwitheWiedemann syndrome and epigenetic alterations of LIT1 and H19.
Am J Hum Genet 2003;72:156e60.
[36] Maher ER, Brueton LA, Bowdin SC, Luharia A, Cooper W, Cole TR, et al.
BeckwitheWiedemann syndrome and assisted reproduction technology
(ART). J Med Genet 2003;40:62e4.
[37] Vermeiden JP, Bernardus RE. Are imprinting disorders more prevalent after
human in vitro fertilization or intracytoplasmic sperm injection? Fertil Steril
2013;99:642e51.
[38] Lidegaard O, Pinborg A, Andersen AN. Imprinting diseases and IVF: Danish
National IVF cohort study. Hum Reprod 2005;20:950e4.
[39] Zhu JL, Basso O, Obel C, Bille C, Olsen J. Infertility, infertility treatment,
and congenital malformations: Danish national birth cohort. BMJ
2006;333:679.
[40] Jaques AM, Amor DJ, Baker HW, Healy DL, Ukoumunne OC, Breheny S, et al.
Adverse obstetric and perinatal outcomes in subfertile women conceiving
without assisted reproductive technologies. Fertil Steril 2010;94:2674e9.
[41] Rimm AA, Katayama AC, Katayama KP. A meta-analysis of the impact of IVF
and ICSI on major malformations after adjusting for the effect of subfertility.
J Assist Reprod Genet 2011;28:699e705.
J.L. Simpson / Seminars in Fetal & Neonatal Medicine 19 (2014) 177e182 182
Review
A historical and practical review of rst trimester aneuploidy
screening
Melissa L. Russo
*
, Karin J. Blakemore
Maternal Fetal Medicine, McKusickeNathans Institute of Genetic Medicine, Johns Hopkins University, School of Medicine, Baltimore, MD, USA
Keywords:
Aneuploidy
First trimester screening
Free b-human chorionic gonadotropin
Maternal serum screening tests
Nuchal translucency measurement
Pregnancy-associated plasma protein A
s u m m a r y
There have been tremendous advancements over the past three decades in prenatal screening for
aneuploidy and we have changed our practice from screening by maternal age alone to combined rst
trimester screening and circulating cell-free fetal DNA. We currently use the nuchal translucency and
biochemical markers of free b-hCG and PAPP-A to determine the risk of fetal aneuploidy. The primary
goal is to identify higher risk women for fetal aneuploidy early in pregnancy and give them the option to
pursue invasive testing in a timely manner if desired.
2013 Elsevier Ltd. All rights reserved.
1. Introduction
The past 30 years have produced numerous discoveries and
advances in prenatal screening for cytogenetic disorders in the
fetus: ultrasound imaging, maternal serum biochemical markers,
and isolation of cell free fetal DNA in maternal serum. These ad-
vancements are responsible for the options currently available for
prenatal screening for aneuploidy in the rst trimester. The pur-
pose of this review article is historical and practical. Its historical
aspect chronicles the origins of and innovations in rst trimester
screening, demonstrating how current screening for fetal aneu-
ploidy has come into practice. On a practical side, it will highlight
the available options for rst trimester screening, indications for
screening, rst trimester screening in twin pregnancies and discuss
the additional knowledge that can be obtained from a rst
trimester combined screen.
2. Origins of prenatal screening for fetal aneuploidy
Prior to the 1980s, the primary method to identify women at
risk for aneuploidy was based on the concept of increased risk with
advanced maternal age. Dr Lionel Penrose was the rst to recognize
this concept in the 1930s when he observed that there was a sig-
nicant association between increasing maternal age and birth of a
Down syndrome child [1]. With the advent of cytogenetic analysis
on cultured amniocytes in the 1970s, all women aged >36 years
were offered amniocentesis to diagnose potential fetal aneuploidy.
Maternal age was a poor screening test in isolation because it only
identied 25e30% of fetal aneuploidy.
The rst recommendation for prenatal aneuploidy screening in
the general population stemmed fromthe observation by Merkatz
et al. [2] in 1984 that maternal serum a-fetoprotein (MSAFP) in
the second trimester was signicantly lower in a woman with a
trisomy 18 fetus. Simultaneously, MSAFP was suggested by Cuckle
[3] as a screening test for Down syndrome in the general popu-
lation. This study modeled a mathematical algorithm that com-
bined maternal age and MSAFP in the second trimester to detect
40% of cases of Down syndrome with a false-positive rate of 6.8%.
This led to the rst protocols for fetal aneuploidy screening in the
general population across all maternal ages. After observations
with MSAFP, altered maternal serum levels in affected pregnan-
cies were observed in other analytes: free b-human chorionic
gonadotropin (b-hCG), inhibin A and unconjugated estriol (uE
3
)
[4e6]. Wald et al. [7] combined the maternal serum markers of b-
hCG, a-fetoprotein (AFP) and uE
3
with maternal age in the triple
screen in 1988. Overall, the mean levels of these analytes were
expressed as a multiple of the expected value for gestational age
based on a logelinear regression in the controls, multiples of the
median (MoM). The mean for the analytes b-hCG, AFP and uE
3
in a
second trimester pregnancy affected with Down syndrome were
2.1, 0.7 and 0.7 MoM respectively. Looking across all maternal
ages, the triple screen detection rate for Down syndrome in the
second trimester was 60% with a false-positive rate of 5%, essen-
tially doubling the detection rate fromage alone. Haddowet al. [8]
examined dimeric inhibin A as yet another marker to be added to
the second trimester screening prole and this was later dubbed
* Corresponding author. Address: Maternal Fetal Medicine, McKusickeNathans
Institute of Genetic Medicine, Johns Hopkins University, School of Medicine, 600
North Wolfe Street, Phipps Building, Suite 228, Baltimore, MD 21287, USA. Tel.: 1
410 502 9893; fax: 1 410 614 8305.
E-mail address: mrusso5@jhmi.edu (M.L. Russo).
Contents lists available at ScienceDirect
Seminars in Fetal & Neonatal Medicine
j ournal homepage: www. el sevi er. com/ l ocat e/ si ny
1744-165X/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.siny.2013.11.013
Seminars in Fetal & Neonatal Medicine 19 (2014) 183e187
the quadruple screen. Using age and the four maternal serum
analytes, the detection rate for Down syndrome was 75% at a
false-positive rate (FPR) of 5%.
3. First trimester maternal serum markers of free b-hCG and
pregnancy-associated plasma protein A (PAPP-A)
The goals for improving prenatal screening for aneuploidy
following the advent of the triple and quadruple screen were to
achieve a higher detection rate, a lower FPR, reassurance to patients
with a low risk and to offer diagnostic testing as early as possible.
Thus, with the development of a maternal serum screening test for
the general population, a logical next question was the feasibility of
implementing prenatal screening earlier in pregnancy. Chorionic
villus sampling (CVS) had been introduced in the 1980s, an invasive
procedure that allowed for prenatal diagnosis in the rst trimester,
adding impetus to move screening to an earlier gestational age [9].
The clear advantages of a screening test in the rst trimester
included earlier reassurance to low risk patients, more time to
consider options for diagnosis, and the possibility of earlier, safer
termination in affected pregnancies.
Spencer et al. [10] published their initial ndings of free b-
hCG as a marker for trisomies 21 and 18 in the rst trimester
and showed that the MoM value of free b-hCG in trisomy 21 was
signicantly greater and the median value in trisomy 18 was
signicantly lower than that in the unaffected controls. The idea
of using free b-hCG was rst proposed by Bogart et al. [11] who
compared the ability of free versus total hCG to detect chro-
mosomally abnormal fetuses in the second trimester. Based on
Bogarts work, Spencer et al. had also examined free b-hCG in
the second trimester. Compared with total hCG, free b-hCG was
a better marker for detection of trisomy 21. Macri and Spencer
[12] then expanded on their earlier paper using free b-hCG as an
analyte from 9 to 13 weeks of gestation and showed that free b-
hCG was signicantly elevated (2.20 MoM) in trisomy 21 preg-
nancies. They concluded that this maternal serum analyte would
serve as a good marker for Down syndrome in the rst trimester.
Around this time, it was discovered that PAPP-A was lower in
pregnancies with fetal aneuploidy. Brambati et al. [13] were the
rst to publish this nding in 13 cases of trisomy 21 at 8e12 weeks
of gestation. The PAPP-A levels were 5th percentile. Wald et al.
[14] expanded on these ndings, concluding that PAPP-A was a
useful marker for trisomy 21 in the rst trimester (0.23 MoM).
Muller et al. [15] further established PAPP-A as a potential useful
marker for trisomy 21 in the rst trimester with their study
examining PAPP-A levels in blood samples originally collected for
toxoplasmosis testing in France (median PAPP-A value for trisomy
21 was 0.42 MoM).
4. Fetal nuchal translucency: a novel approach to screen for
aneuploidy
The British physician, John Langdon Down, was the rst to
describe Down syndrome in 1866 [16]. He noted: The skin has a
slight dirty yellowish texture and is decient in elasticity, giving the
impression of being too large for the body.
In the 1990s, Nicolaides et al. [17] pondered the observation of
the fullness of the neck in Down syndrome neonates and the as-
sociation of nuchal edema/cystic hygroma on second trimester ul-
trasound in fetuses with chromosomal abnormalities. Nicolaides
hypothesized that increased nuchal thickness on ultrasound in the
rst trimester could be a marker for fetal aneuploidy. In a pro-
spective study, 827 pregnant women who chose to have diagnostic
testing underwent transabdominal ultrasound at 10e14 weeks of
gestation to evaluate the uid behind the neck in their fetuses.
Visualizing a sagittal section of the crownerump length, the
maximum thickness of subcutaneous translucency between the
skin and soft tissue overlying the cervical spine was measured. In
the 51 fetuses with a nuchal translucency thickness of 3e8 mm, the
incidence of chromosomal abnormalities was 35%. In the fetuses
with smaller measurements (n 776), by contrast, 1% had chro-
mosomal abnormalities. He also noted that the abnormal uid
collection had, for the most part, resolved by the second trimester.
In this study, Nicolaides reported that an increased nuchal trans-
lucency was associated with an increased risk for chromosomal
defects. He introduced the term nuchal translucency (NT) into the
vocabulary of genetics counselors, obstetricians, ultrasound tech-
nicians, radiologists, maternal fetal medicine specialists and aca-
demic medicine. This article transformed the landscape for prenatal
screening.
Based on Nicolaides study, another prospective study by
Brambati et al. [18] was performed to evaluate the technical and
practical aspects of a screening program in the general population.
In women undergoing CVS between 8 and 13 weeks of gestation,
the NT was measured for maximum thickness. These investigators
were among the rst to introduce a standardized protocol: (i) two
different observers scrutinize the ultrasonographic images of the
posterior fetal contour in sagittal plane with the requirement to
distinguish between amnion and fetal skin; and (ii) an abnormal
cut-off value of 3 mm. In 70 fetuses with NT above the cut-off,
18.6% had chromosomal disorders versus 1.7% in the normal NT
group. Brambati et al. afrmed that the NT increased with
increasing gestational age and that an increased NT was associated
with chromosomally abnormal fetuses; he also addressed the need
for quality control and standardization with measurement of NT
screening.
Pandyas and Nicolaides group examined 1015 fetuses with NTs
3 mm and found that the incidence of chromosomal abnormal-
ities, namely, trisomy 21, 18 and 13, was signicantly associated
with increased fetal NT and additively with maternal age at 10e14
weeks of gestation [19]. Another study by Pandya et al. [20]
demonstrated that fetal NT increases with increasing crowne
rump length, and that the likelihood of trisomy 21 varies with the
degree by which a given NT deviates from the normal median at a
given crownerump length. Instead of having a specic numerical
cut-off for the NT, they used the cut-off of >95th percentile above
the normal median and detected 77% of fetuses with trisomy 21 and
78% of other chromosomal abnormalities using only maternal age
plus NT. For quality control, a subgroup analysis showed good
reproducibility of NT measurements when ultrasound examina-
tions were performed by different sonographers, all trained in the
same fashion.
Snijders et al. [21] published one of the most comprehensive
studies on fetal aneuploidy screening with maternal age and nuchal
translucency from the Fetal Medicine Foundation First Trimester
Screening Group. This trial, conducted at 22 centers, examined
96 127 cases of rst trimester NT screening with known genetic
outcomes. The median maternal age for the study group was 31
years; there was a preponderance of older-age women in this study.
The NT was above the 95th percentile for a given crownerump
length in 71.8% of 326 trisomy 21 pregnancies, and in 70.5% of 325
other chromosomally abnormal pregnancies versus 4.4% of normal
pregnancies. With the addition of maternal age, a risk cut-off of 1 in
300 or higher was found in 82.2% of trisomy 21 pregnancies, 77.8%
of other chromosomally abnormal pregnancies and 8.3% of normal
pregnancies.
Another important contribution of this study was the estab-
lishment of criteria by the Fetal Medicine Foundation to achieve a
uniform 10e14-week scan among many different operators and
institutions [21]. These criteria include: a proper sagittal view;
M.L. Russo, K.J. Blakemore / Seminars in Fetal & Neonatal Medicine 19 (2014) 183e187 184
magnication (fetus occupies at least 75% of the image); distinction
between fetal skin and amnion; and reporting the maximum
measurement obtained. The Nuchal Translucency Education and
Quality Review program (NTQR) was developed in 2004 with
criteria from the Society for Maternal Fetal Medicine similar to that
of the Fetal Medicine Foundation for nuchal translucency
screening, and these criteria have since been implemented in North
America and Canada.
Another study on quality control by Whitlow et al. [22] exam-
ined howthe head in an extended or exed position distorts the NT
measurement. From this publication came the recommendation to
measure the fetal neck in a neutral position, as this is the most
reproducible.
5. Combination of biochemistry and nuchal translucency in
the rst trimester
Brambati et al. [23] rst reported the combination of free b-hCG
and PAPP-A, attaining a detection rate of 78.9% for trisomy 21 in a
small series of 13 cases of trisomy 21 and 89 unaffected controls
using just these maternal serum analytes, plus maternal age. Noble
et al. [24] were among the rst to combine biochemistry and the NT
plus age, nding that free b-hCG and NT were two independent
parameters; thus their likelihood ratios could be combined. This
combined screening trial had a detection rate for trisomy 21 of 85%
at a false-positive rate (FPR) of 5%.
Two years later, Orlandi et al. [25] published one of the rst
studies that prospectively assessed the combination of maternal
age, biochemical screening tests of free b-hCG and PAPP-A, and NT
to determine risk for fetal aneuploidy in the rst trimester. The
free b-hCG and PAPP-A values were divided by respective gesta-
tional day-specic median levels to determine the MoM for each
analyte. Trisomy 21 pregnancies were associated with elevated
free b-hCG and low PAPP-A, and trisomy 18 pregnancies were
associated with low free b-hCG and PAPP-A. The likelihood ratios
for each analyte, the NT, and the patients prior age-related risk
were multiplied together to determine a nal risk ratio, and high
risk was determined to be 1 in 380. With this combination, the
detection rate for trisomy 21 was 87% with a FPR of 5% whereas
the detection rate for trisomy 18 was 76% with FPR of 1%. Orlandi
et al.s study also demonstrated the utility of dried blood tech-
nology for specimen collection. The use of the dried blood spot
simplied and allowed for greater exibility in collection, shipping
and processing.
Spencer et al. [26] examined the combined rst trimester
screening for trisomy 21 in a retrospective analysis, and expedited
the process to a one-stop clinic employing new technology that
allowed for results of biochemical analysis within 30 min of
obtaining a blood sample. This technology resulted in precise
measurements with good reproducibility and allowed for com-
plete early fetal assessment in one visit. This study also demon-
strated that each marker was independent; when combined they
obtained a detection rate for trisomy 21 of 89% with FPR of 5%.
Spencer et al. re-conrmed their conclusions from earlier studies
that risk algorithms needed to be created that were adjusted for
gestational age because detection rates varied at different gesta-
tional ages.
Krantz et al. [27] reported on the rst trimester combined
screenings ability to detect trisomy 18 in addition to trisomy 21.
They used the dried blood spot technology previously described,
and a standardized protocol outlined by the Fetal Medicine Foun-
dation for nuchal translucency measurements. Krantz et al. showed
a detection rate for trisomy 21 of 87.5% with FPR of 4.5% in women
aged <35 years and detection rate of 92% with FPR of 14.3% in
women aged 35 years. For trisomy 18, there was a detection rate
of 100% with FPR of 0.4e1.4% for women aged <35 and 35 years
respectively.
Wapner et al. [28] published the rst multicenter trial in the USA
for the combined rst trimester screening for trisomies 21 and 18.
This study, the BUN trial, demonstrated that across multiple cen-
ters this screening method was reproducible and appropriate for
use in clinical practice. The detection rate for trisomy 21 was 78.7%
with FPR of 5%. The detection rate for trisomy 18 was 90.9% with
FPR of 2%.
6. Comparison trials of different screening methods
Once it was demonstrated that combined rst trimester
screening was feasible and applicable to the general population,
the next question was the comparison of combined rst
trimester screening to second trimester screening to determine
whether one test was superior. Numerous studies showed that
the detection rates for trisomy 21 in the rst trimester, with a
xed FPR of 5%, ranged from 76% to 91% and this was thought to
be better than the detection rates in the second trimester [25e
29]. In any comparison of rst and second trimester screening,
however, it is important to recognize that the overestimate in
detection rates in the rst trimester is because the incidence of
affected pregnancies is higher and a certain percentage of
pregnancies detected by early prenatal screening will result in
spontaneous fetal loss as gestation advances. A study by Dunstan
and Nix [30], for example, provides a methodology to compare
detection rates between the rst and second trimester screening
tests for trisomy 21 by taking into account fetal loss rates in the
rst trimester and incorporating this information into the com-
parison. Dunstan and Nix summarized that rst trimester
screening could only be determined to be superior to second
trimester screening if the aneuploidy detection rate was at least
8.3% higher.
During this time, the integrated screen was introduced by Wald
et al. [31] and used markers from the rst and second trimester to
provide a single estimate of aneuploidy risk revealed during the
second trimester. The detection rate for trisomy 21 was 85% with a
much lower FPR of 0.9%.
Two large, multicenter, prospective trials in the UK followed:
the Serum, Urine and Ultrasound Screening Study (SURUSS), and
in the USA, the First-Trimester and Second-Trimester Screening or
Both for Down Syndrome (FASTER) trial. These trials sought to
compare all of the different available screening methods and
determine the most effective and safest method for screening for
trisomy 21. In the SURUSS trial, the most effective screening test
was the integrated test with a detection rate of 85% and a FPR of
0.9% [32]. For the same detection rate of 85%, the serum inte-
grated, combined rst trimester screen and quadruple second
trimester screen had FPRs of 3.9%, 4.3% and 6.2% respectively. The
FASTER trial had similar results. At a xed detection rate of 85%,
the FPRs of the rst trimester combined, serum integrated, fully
integrated, and second trimester quadruple screen were, respec-
tively, 4.8%, 4.4%, 0.8% and 7.3% [33]. From these trials, it was
concluded that the choice of screening test is individualized and
depends on the womans gestational age, whether she desires a
denitive diagnosis in the rst trimester versus the best overall
screening test.
7. History of cell-free fetal DNA from maternal blood in
detection of aneuploidy
Non-invasive aneuploidy screening with cell-free fetal DNA
from maternal plasma is a relatively novel clinical methodology;
however, the concept of isolating fetal cells and fetal cell-free DNA
M.L. Russo, K.J. Blakemore / Seminars in Fetal & Neonatal Medicine 19 (2014) 183e187 185
from maternal blood was rst discovered more than 50 years ago
[34]. Preliminary endeavors towards non-invasive prenatal
screening were focused on detection and isolation of intact fetal
cells from the maternal blood using uorescence-activated cell
sorting [35,36]. One study used isolated, nucleated fetal red blood
cells via uorescent in-situ hybridization (FISH) analysis to detect
trisomy 18 [37]. A large, multicenter trial by the National Institute of
Child Health and Human Development (NIFTY) used intact fetal
nucleated red blood cells with cell sorting, FISH and polymerase
chain reaction to detect 74% of cases of fetal aneuploidy with FPR of
0.6e4% [38]. This method of aneuploidy screening was comparable
with prenatal serumscreening, but had higher specicity. Before its
introduction into the clinical arena, further advancement was
deemed necessary.
Cell-free fetal DNA in maternal plasma was reported by Lo
et al. [39] in 1997. This discovery paved the way for subsequent
developments of non-invasive screening for aneuploidy. Circu-
lating cell-free DNA comprises 3e6% of total cell-free maternal
DNA. It is derived primarily from the placenta and is cleared
from maternal blood within hours after delivery. Cell-free fetal
DNA was rst used for gender determination with X-linked
disorders and for fetal Rh(D) genotyping with Rh-negative
pregnant women.
In 2008, massive parallel sequencing revolutionized the detec-
tion of cell-free fetal DNA [40e42]. With this shotgun sequencing-
based approach, the number of cell-free DNA fragments was
quantied and fetal aneuploidy was detectable. This technological
advance led to several large trials to explore the validity and
feasibility of aneuploidy screening with cell-free fetal DNA [43e46].
Each of these clinical trials had slightly different methods for
analysis of cell-free fetal DNA fractions. After these large clinical
trials, cell-free fetal DNA screening was implemented into clinical
practice for women at risk for fetal aneuploidy.
8. Practical application and other aspects of rst trimester
screening for aneuploidy
There are many options currently available for aneuploidy
screening: the combined rst trimester screen, quadruple screen,
integrated screen or sequential screen. Additionally, for pregnan-
cies at high risk for aneuploidy, cell-free fetal DNA testing is now
being offered. It is important to recognize that these tests are not
diagnostic and each test has associated false negatives and false
positives. If a patient wants to have a denitive diagnosis of
aneuploidy, chorionic villus sampling or amniocentesis are
currently the only available options.
A secondary marker in the rst trimester screen for aneu-
ploidy is the presence of the nasal bone. The incorporation of the
nasal bone into the rst trimester screen improves the detection
rate for trisomy 21 with xed FPR, or retains a 90% detection rate
with a simultaneous 10-fold decrease in the FPR from 5% to 0.5%
[47].
First trimester screening in twin pregnancies is a possible and
practical option. There are several factors to take into consider-
ation with twins. The NTs can be used to determine an individual
risk to each fetus; however, serum markers can only be used to
obtain a pregnancy-specic risk. If there is a positive screen
result, then invasive testing is generally offered for each fetus.
In addition to its use as a screening tool for aneuploidy, the
rst trimester combined screen has several other applications.
The rst trimester screen can accurately establish estimated date
of connement and chorionicity in multiple gestations, which is
very important for obstetrical management. The rst trimester
evaluation of the nuchal translucency is, moreover, a valued
examination to screen for congenital cardiac defects,
diaphragmatic hernia, some skeletal dysplasias and other ge-
netic conditions. In the future, rst trimester screening may
evolve to incorporate the newer technology of cell-free fetal DNA
along with maternal serum analytes and NT measurement by
ultrasound.
9. Conclusions
First trimester screening for aneuploidy is a valuable tool for the
obstetrician and indications and methods of screening have
evolved over the last quarter century. The primary goal of rst
trimester screen is to identify higher risk women for fetal aneu-
ploidy and give them the option to pursue diagnostic testing in a
timely manner if desired.
Conict of interest statement
None declared.
Funding sources
None.
References
[1] Penrose LS. The Colchester Survey: a clinical and genetic study of 1280 cases
of mental defect. London: HMSO/Privy Council of Medical Research Council;
1938.
[2] Merkatz IR, Nitowsky HM, Macri JN, Johnson WE. An association between
low maternal serum alpha-fetoprotein and fetal chromosomal abnormalities.
Am J Obstet Gynecol 1984;148:886e94.
[3] Cuckle HS, Wald NJ, Lindenbaum RH. Maternal serum alpha-fetoprotein
measurement: a screening test for Down syndrome. Lancet 1984;323:926e9.
[4] Macri JN, Kasturi RV, Krantz DA, Cook EJ, Moore ND, Young JA, et al. Maternal
serum Down syndrome screening: free beta-protein is a more effective
marker than human chorionic gonadotropin. Am J Obstet Gynecol
1990;163(4 Pt 1):1248e53.
[5] Canick JA, Knight GJ, Palomaki GE, Haddow JE, Cuckle HS, Wald NJ. Low
second trimester maternal serum unconjugated oestriol in pregnancies with
Downs syndrome. Br J Obstet Gynaecol 1988;95:330e3.
[6] Van Lith JM, Pratt JJ, Beekhuis JR, Mantingh A. Second-trimester maternal
serum immunoreactive inhibin as a marker for fetal Downs syndrome.
Prenat Diagn 1992;12:801e6.
*[7] Wald NJ, Cuckle HS, Densem JW, Nanchahal K, Royston P, Chard T, et al.
Maternal serum screening for Downs syndrome in early pregnancy. BMJ
1988;297:883e7.
*[8] Haddow JE, Palomaki GE, Knight GJ, Foster DL, Neveux LM. Second trimester
screening for Downs syndrome using maternal serum dimeric inhibin A.
J Med Screen 1998;5:115e9.
[9] McGowan KD, Blakemore KJ. Amniocentesis and chorionic villus sampling.
Curr Opin Obstet Gynecol 1991;3:221e9.
Practice points
The rst trimester screen for fetal aneuploidy has evolved
continuously over the past 30 years.
The current rst trimester aneuploidy screen includes
maternal serum analytes for free b-hCG and PAPP-A and
NT measurement on ultrasound.
For patients at high risk for fetal aneuploidy, cell-free fetal
DNA in maternal plasma has been incorporated as
another screening test.
First trimester screening has other applications such as
conrmation of dating of a pregnancy, establishment of
chorionicity in multiple gestations, and screening for
congenital heart defects, skeletal dysplasias or other
congenital malformation and conditions by increased
nuchal translucency.
M.L. Russo, K.J. Blakemore / Seminars in Fetal & Neonatal Medicine 19 (2014) 183e187 186
[10] Spencer K, Macri JN, Aitken DA, Connor JM. Free beta-hCG as rst-trimester
marker for fetal trisomy. Lancet 1992;339:1480.
[11] Bogart MH, Pandian MR, Jones OW. Abnormal maternal serum chorionic
gonadotropin levels in pregnancies with fetal chromosome abnormalities.
Prenat Diagn 1987;7:623e30.
[12] Macri JN, Spencer K, Aitken D, Garver K, Buchanan PD, Muller F, et al. First-
trimester free beta (hCG) screening for Down syndrome. Prenat Diagn
1993;13:557e62.
[13] Brambati B, Lanzani A, Tului L. Ultrasound and biochemical assessment of
rst trimester of pregnancy. In: The embryo: normal and abnormal devel-
opment and growth. London: SpringereVerlag; 1990. pp. 181e94.
[14] Wald N, Stone R, Cuckle HS, Grudzinskas JG, Barkai G, Brambati B, et al. First
trimester concentrations of pregnancy associated plasma protein A and
placental protein 14 in Downs syndrome. BMJ 1992;305:28.
[15] Muller F, Cuckle H, Teisner B, Grudzinskas JG. Serum PAPP-A levels are
depressed in women with fetal Down syndrome in early pregnancy. Prenat
Diagn 1993;13:633e6.
[16] Down JL. Observations on an ethnic classication of idiots. Clinical Lecture
Reports, London Hospital 1866;3:259e62.
*[17] Nicolaides KH, Azar G, Byrne D, Mansur C, Marks K. Fetal nuchal trans-
lucency: ultrasound screening for chromosomal defects in rst trimester of
pregnancy. BMJ 1992;304:867e9.
[18] Brambati B, Cislaghi C, Tului L, Alberti E, Amidani M, Colombo U, et al. First-
trimester Downs syndrome screening using nuchal translucency: a pro-
spective study in patients undergoing chorionic villus sampling. Ultrasound
Obstet Gynecol 1995;5:9e14.
[19] Pandya PP, Kondylios A, Hilbert L, Snijders RJ, Nicolaides KH. Chromosomal
defects and outcome in 1015 fetuses with increased nuchal translucency.
Ultrasound Obstet Gynecol 1995;5:15e9.
[20] Pandya PP, Snijders RJ, Johnson SP, De Lourdes Brizot M, Nicolaides KH.
Screening for fetal trisomies by maternal age and fetal nuchal translucency
thickness at 10 to 14 weeks of gestation. Br J Obstet Gynaecol 1995;102:
957e62.
*[21] Snijders RJ, Noble P, Sebire N, Souka A, Nicolaides KH. UK multicentre
project on assessment of risk of trisomy 21 by maternal age and fetal
nuchal-translucency thickness at 10e14 weeks of gestation. Fetal Med-
icine Foundation First Trimester Screening Group. Lancet 1998;352:
343e6.
[22] Whitlow BJ, Chatzipapas IK, Economides DL. The effect of fetal neck position
on nuchal translucency measurement. Br J Obstet Gynaecol 1998;105:872e6.
[23] Brambati B, Tului L, Bonacchi I, Shrimanker K, Suzuki Y, Grudzinskas JG.
Serum PAPP-A and free beta-hCG are rst-trimester screening markers for
Down syndrome. Prenat Diagn 1994;14:1043e7.
[24] Noble PL, Abraha HD, Snijders RJ, Sherwood R, Nicolaides KH. Screening for
fetal trisomy 21 in the rst trimester of pregnancy: maternal serum free
beta-hCG and fetal nuchal translucency thickness. Ultrasound Obstet Gyne-
col 1995;6:390e5.
[25] Orlandi F, Damiani G, Hallahan TW, Krantz DA, Macri JN. First-trimester
screening for fetal aneuploidy: biochemistry and nuchal translucency. Ul-
trasound Obstet Gynecol 1997;10:381e6.
[26] Spencer K, Souter V, Tul N, Snijders R, Nicolaides KH. A screening program for
trisomy 21 at 10e14 weeks using fetal nuchal translucency, maternal serum
free beta-human chorionic gonadotropin and pregnancy-associated plasma
protein-A. Ultrasound Obstet Gynecol 1999;13:231e7.
*[27] Krantz DA, Hallahan TW, Orlandi F, Buchanan P, Larsen Jr JW, Macri JN. First-
trimester Down syndrome screening using dried blood biochemistry and
nuchal translucency. Obstet Gynecol 2000;96:207e13.
*[28] Wapner R, Thom E, Simpson JL, Pergament E, Silver R, Filkins K, et al., First
Trimester Maternal Serum Biochemistry and Fetal Nuchal Translucency
Screening (BUN) Study Group. First-trimester screening for trisomies 21 and
18. N Engl Med J 2003;349:1405e13.
[29] Wald NJ, Hackshaw AK. Combining ultrasound and biochemistry in rst-
trimester screening for Downs syndrome. Prenat Diagn 1997;17:821e9.
[30] Dunstan FD, Nix AB. Screening for Downs syndrome: the effect of test date
on the detection rate. Ann Clin Biochem 1998;35(Pt 1):57e61.
[31] Wald NJ, Watt HC, Hackshaw AK. Integrated screening for Downs syndrome
on the basis of tests performed during the rst and second trimesters. N Engl
Med J 1999;341:461e7.
*[32] Wald NJ, Rodeck C, Hackshaw AK, Walters J, Chitty L, Mackinson AM. First
and second trimester antenatal screening for Downs syndrome: the results
of the Serum, Urine and Ultrasound Screening Study (SURUSS). J Med Screen
2003;10:56e104.
*[33] Malone FD, Canick JA, Ball RH, Nyberg DA, Comstock CH, Bukowski R, et al.
First-trimester or second-trimester screening, or both, for Downs syndrome.
N Engl Med J 2005;353:2001e11.
[34] Kleihauer E, Braun H, Betke K. Demonstration of fetal hemoglobin in eryth-
rocytes of a blood smear (German). Klin Wochenschr 1957;35:637e8.
[35] Herzenberg LA, Bianchi DW, Schrder J, Cann HM, Iverson GM. Fetal cells in
the blood of pregnant women: detection and enrichment by uorescence-
activated cell sorting. Proc Natl Acad Sci USA 1997;76:1453e5.
[36] Iverson GM, Bianchi DW, Cann HM, Herzenberg LA. Detection and isolation of
fetal cells from maternal blood using the uorescence-activated cell sorter
(FACS). Prenat Diagn 1981;1:61e73.
[37] Price JO, Elias S, Wachtel SS, Klinger K, Dockter M, Tharapel A, et al. Prenatal
diagnosis with fetal cells isolated from maternal blood by multiparameter
ow cytometry. Am J Obstet Gynecol 1991;165(6 Pt 1):1731e7.
[38] Bianchi DW, Simpson JL, Jackson LG, Elias S, Holzgreve W, Evans MI, et al.
Fetal gender and aneuploidy detection using fetal cells in maternal blood:
analysis of NIFTY I data. National Institute of Child Health and Development
Fetal Cell Isolation Study. Prenat Diagn 2002;22:609e15.
[39] Lo YM, Corbetta N, Chamberlain PF, Rai V, Sargent IL, Redman CW, et al.
Presence of fetal DNA in maternal plasma and serum. Lancet 1997;350:
485e7.
[40] Fan HC, Blumenfeld YJ, Chitkara U, Hudgins L, Quake SR. Noninvasive diag-
nosis of fetal aneuploidy by shotgun sequencing DNA from maternal blood.
Proc Natl Acad Sci USA 2008;105:16266e71.
[41] Chiu RW, Chan KC, Gao Y, Lau VY, Zheng W, Leung TY, et al. Noninvasive
prenatal diagnosis of fetal chromosomal aneuploidy by massively parallel
genomic sequencing of DNA in maternal plasma. Proc Natl Acad Sci USA
2008;105:20458e63.
[42] Chiu RW, Akolekar R, Zheng YW, Leung TY, Sun H, Chan KC, et al. Non-
invasive prenatal assessment of trisomy 21 by multiplexed maternal plasma
DNA sequencing: large scale validity study. BMJ 2011;342:c7401.
[43] Palomaki GE, Kloza EM, Lambert-Messerlian GM, Haddow JE, Neveux LM,
Ehrich M, et al. DNA sequencing of maternal plasma to detect Down syn-
drome: an international clinical validation study. Genet Med 2011;13:913e
20.
[44] Palomaki GE, Deciu C, Kloza EM, Lambert-Messerlian GM, Haddow JE,
Neveux LM, et al. DNA sequencing of maternal plasma reliably identies
trisomy 18 and trisomy 13 as well as Down syndrome: an international
collaborative study. Genet Med 2012;14:296e305.
[45] Bianchi DW, Platt LD, Goldberg JD, Abuhamad AZ, Sehnert AJ, Rava RP,
MatErnal BLood IS Source to Accurately diagnose fetal aneuploidy (MELISSA)
Study Group. Genome-wide fetal aneuploidy detection by maternal plasma
DNA sequencing. Obstet Gynecol 2012;119:890e901.
[46] Norton ME, Brar H, Weiss J, Karimi A, Laurent LC, Caughey AB, et al. Non-
Invasive Chromosomal Evaluation (NICE) Study: results of a multicenter
prospective cohort study for detection of fetal trisomy 21 and trisomy 18. Am
J Obstet Gynecol 2012;207:137. e1e8.
[47] Cicero S, Bindra R, Rembouskos G, Spencer K, Nicolaides KH. Integrated ul-
trasound and biochemical screening for trisomy 21 using fetal nuchal
translucency, absent fetal nasal bone, free beta-hCG and PAPP-A at 11 to 14
weeks. Prenat Diagn 2003;23:306e10.
M.L. Russo, K.J. Blakemore / Seminars in Fetal & Neonatal Medicine 19 (2014) 183e187 187
Review
Selective serotonin reuptake inhibitors in human pregnancy: On the
way to resolving the controversy
Asher Ornoy
a, b,
*
, Gideon Koren
b, c
a
Hebrew University Hadassah Medical School, Jerusalem, Israel
b
Israeli Teratology Information Service, Israel Ministry of Health, Jerusalem, Israel
c
Motherisk Program, Division of Clinical Pharmacology/Toxicology, Department of Pediatrics, Hospital for Sick Children, Toronto, Canada
Keywords:
Major anomalies
Neonatal effects
Persistent pulmonary hypertension
Neurodevelopmental effects
Selective serotonin reuptake inhibitors
(SSRIs)
s u m m a r y
There has been an increase in the use of selective serotonin reuptake inhibitors (SSRIs) during pregnancy.
However, in the last 10 years, in spite of a vast literature regarding use in pregnancy there seems to be
some confusion as to the possible risk of these drugs, especially related to cardiovascular anomalies. In
addition, there are data on developmental follow-up studies that raise the question of possible slight
developmental and neurobehavioral problems. The purpose of the present review is therefore to criti-
cally summarize the current evidence for the risk/benet analysis of SSRI use in human pregnancy.
Although most studies have not shown an increase in the overall risk of major malformations, several
have suggested that the use of SSRIs may be associated with a small increased risk for cardiovascular
malformations. However, newcompelling evidence shows that this apparent increased risk occurs also in
women with untreated depression, highlighting the probable ascertainment bias involved in many of
these studies. Persistent pulmonary hypertension of the newborn (PPHN) has also been described with
an absolute risk of <1%; however, here too, higher rates were described among offspring of women with
untreated depression. Poor neonatal adaptation has been described in up to 30% of neonates exposed to
SSRIs late in pregnancy. Of the few postnatal developmental follow-up studies, there are no signicant
developmental problems. The literature on SSRIs in pregnancy is somewhat confusing but when ana-
lysing all prospective cohort data there seems to be no demonstrable increase in the rate of major
anomalies or developmental disorders. When evaluating the risk/benet ratio of SSRI treatment in
pregnancy, the risk associated with treatment discontinuation e e.g. higher frequency of relapse,
increased risk of preterm delivery and postpartum depression e appear to outweigh the potential, un-
proven risks of treatment. Moreover, maternal depression may negatively affect the childs development,
emphasizing the importance of prevention by appropriate treatment during pregnancy with the least
minimal effective dose.
2014 Elsevier Ltd. All rights reserved.
1. Introduction
Selective serotonin reuptake inhibitors (SSRIs) are widely pre-
scribed for the treatment of depression, anxiety and other disor-
ders. The prevalence rates of depression during pregnancy have
been estimated as 7.4%, 12.8% and 12.0%, for the rst, second and
third trimesters, respectively, and up to 20% of women may have a
depressive episode in the rst postpartum year [1,2]. Indeed, the
frequency of the reported use of antidepressants in pregnancy
increased from 2.5% in 1998 to >8% in 2006, as reported by the
American National Birth Defects Study [2,3]. One important group
of drugs comprises the SSRIs, rst introduced for clinical use in the
1980s. They have better tolerability and safety compared with rst
generation antidepressants, e.g. tricyclic antidepressants, and are
safer in overdose exposures. They exert their effects by inhibiting
the presynaptic plasma membrane serotonin transporter which
mediates the reuptake of serotonin into the presynaptic terminal.
Thus, treatment with an SSRI initially blocks reuptake and results in
enhanced and prolonged serotonergic neurotransmission due to an
increase in synaptic dopamine concentrations [4,5]. All SSRIs,
including those with an additional effect on norepinephrine reup-
take (SNRIs), share a similar mechanism of action despite having
different chemical structures. SSRI use in pregnancy has increased
over the years and in the USA is w6% [4,5]. SSRIs readily cross the
* Corresponding author. Address: Israeli Teratology Information Service, Israel
Ministry of Health and Laboratory of Teratology, Department of Anatomy and Cell
Biology, Hebrew University Hadassah Medical School, PO Box 12272, Jerusalem
91120, Israel. Tel./fax: 972 2675 8430.
E-mail address: asher.ornoy@mail.huji.ac.il (A. Ornoy).
Contents lists available at ScienceDirect
Seminars in Fetal & Neonatal Medicine
j ournal homepage: www. el sevi er. com/ l ocat e/ si ny
1744-165X/$ e see front matter 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.siny.2013.11.007
Seminars in Fetal & Neonatal Medicine 19 (2014) 188e194
human placenta [6,7] and are also secreted, although in relatively
small amounts, in human milk. SSRIs are the most studied anti-
depressants in pregnancy with >33 000 pregnancies reported in
different studies [4]. Nevertheless there are still conicting views
on the risks of these drugs during pregnancy. When judging the use
of these drugs in pregnancy one should consider the possible
overall effects on the fetus in relation to risks of untreated
depression or anxiety for the mother and offspring.
The purpose of the present review is to summarize and evaluate
the risk/benet analysis of SSRI use during human pregnancy,
including neonatal effects and neurodevelopmental outcome with
special emphasis on prospective cohort studies.
2. Human studies in pregnancy: Major congenital anomalies
and cardiac defects
There are numerous prospective, retrospective and caseecon-
trol studies on the possible effects of SSRI exposure in pregnancy
and in the neonatal period. There seems to be no specic syndrome
or any specic type of malformation related to any of the SSRIs.
However, a possible association between the use of several SSRIs in
pregnancy and cardiac anomalies [8e15] as well as a fewrare major
malformations (neural tube defects, craniosynostosis, omphalocele,
cystic kidney or hypospadias) [15e18] have been described in some
studies. In the positive studies, there was some dominance of
ventricular septal defects and right ventricular outow tract
obstruction defects. Paroxetine was the drug involved with most of
these associations. Many of the prospective studies were under-
powered for associations between exposure and specic malfor-
mations, and retrospective studies had serious potential biases. The
relative risk in positive studies was generally <2 and the lower
range of the 95% condence interval was often close to one. Many of
the studies with positive ndings on SSRIs in pregnancy have been
prescription registries, and women may not have actually taken the
drugs, as shown clearly in the Swedish registry [19]. Exact timing of
exposure during sensitive periods is often problematic, although
exposure preceded the outcome. The ndings in regard to the type
of cardiac malformations are inconsistent although ventricular
septal defect and related anomalies prevail. Results even differ in
studies from the same database published at different time-points.
On the other hand, many other studies have not shown an in-
crease in the overall risk of major malformations [20e28] or an
association with cardiac anomalies. Although the number of
negative studies seem to be higher with larger numbers of exposed
women, it may be difcult to judge, at least with regards to the
possible increased rate of cardiac anomalies. One should also
remember that in most cases the possible contribution to the un-
derlying maternal disease was not controlled for [4,29]. Hence,
rather than reporting the data of individual studies it may be more
accurate to describe the results of the meta-analyses and of
different reviews published on this topic.
2.1. Meta-analyses
Several meta-analyses on paroxetine exposure in pregnancy and
cardiovascular malformations have demonstrated an increased rate
of cardiac anomalies but their signicance is not simple [30e32]. In
the meta-analysis published by Bar-Oz et al. [30] summarizing
studies from 1985 to 2006, rst trimester paroxetine exposure was
associated with a signicant increase in the risk of cardiac anom-
alies. The authors compared the outcome of 2575 pregnancies us-
ing paroxetine with 14 026 non-exposed pregnancies and found an
overall odds ratio of 1.72. However, they also found that women
using antidepressants in pregnancy had signicantly more ultra-
sound examinations in pregnancy and postnatal echocardiograms
compared with women who did not. Hence, detection bias was
suggested as a contributing factor to the observed risk of cardio-
vascular malformations with paroxetine. It is therefore difcult to
conclude from this publication that SSRIs increased the rate of
congenital anomalies. Wurst et al. [31] in their meta-analysis found
an increased risk for combined cardiac defects and aggregated
congenital defects with rst trimester paroxetine use, analysing 20
studies that met the inclusion criteria. They based some of their
results on the data previously published by Bar Oz et al. in their
meta-analysis. The overall prevalence odds ratio was 1.46 for car-
diac defects and 1.24 for aggregated congenital anomalies. Hence,
the risk seems to be low. Moreover, opposing commentaries on this
meta-analysis were recently published stating that the meta-
analysis did not include several important, negative studies [32,33].
In contrast to these publications, there are several meta-
analyses where no increase in cardiac anomalies was observed
following exposure to different SSRIs including paroxetine [34e37].
In the meta-analysis of three caseecontrol studies and six cohort
studies including a large number of paroxetine exposures pub-
lished by OBrien et al. [34], no increased risk of congenital mal-
formations was associated with paroxetine. Cardiac malformation
rates were similar and within population norms. A meta-analysis
discussing the effects of uoxetine in pregnancy did not show any
increase in major congenital anomalies [35]; in another meta-
analysis of seven studies on the effects of older and newer SSRIs
the relative risk for major malformation was 1.01 [36]. A similar
meta-analysis of different SSRIs examining the studies published
from 1990 to 2005 also did not nd any increased risk of major
congenital anomalies [37].
The denition of cardiovascular malformations varied among
studies, some including small septal defects, whereas others
excluded them. The inconsistency across these studies may be
explained by differences in study design, by confounding factors,
e.g. maternal underlying psychiatric disorder, co-administered
medications, lifestyle factors (smoking, drinking), or maternal
obesity and diabetes.
3. New evidence towards resolution of the controversy
A recent, very large national population-based study from
Denmark seems to have largely resolved this controversy [38]. The
authors analysed the association between SSRI use and major
congenital malformations, with focus on cardiac defects. Impor-
tantly, unlike scores of other studies previously mentioned, the
authors also identied a group of women with depression who
avoided taking their SSRIs in pregnancy. Using the Danish Medical
Birth Registry, the authors identied 848 786 pregnancies, 4183
having been exposed to an SSRI throughout the rst trimester, and
806 pregnancies of depressed women who had avoided use of
these drugs during pregnancy. Risks of cardiac malformations were
similar for pregnancies exposed to an SSRI throughout the rst
trimester [adjusted odds ratio (OR): 2.01; 95% condence interval
(CI): 1.60e2.53], and for pregnancies with paused SSRI treatment
during pregnancy (1.85; 1.07e3.20) (P 0.94). The authors found
similar increased risks of specic cardiac malformations for the
individual SSRIs. Just as important, the authors found no dosee
response association [38].
This study has documented that the apparent association be-
tween SSRI use and cardiac malformations may be confounded by
indications. Specically, the authors quoted research published in
2007 showing that women with depression and anxiety have a 2e
3-fold higher likelihood of having ultrasound and echocardiogram
in their babies than healthy women, and, hence, much higher
likelihood of detecting cardiac anomalies [29]. Moreover, women
with depression and anxiety are signicantly more likely than
A. Ornoy, G. Koren / Seminars in Fetal & Neonatal Medicine 19 (2014) 188e194 189
healthy women to attend emergency departments with their in-
fants, where there may be more opportunity to detect a heart
murmur and follow it up [29]. This source of ascertainment bias
has not been sought or addressed in any previous study. Moreover,
this bias is further augmented by the widely publicized news of
SSRIs causing malformations, facilitating more babies of women
on SSRI to be examined but with no similar urge to test children of
healthy women.
These new data have major implications for pregnant women
with depression, as large numbers of them discontinue much-
needed antidepressants, with risk of exacerbation of symptoms,
hospitalization, suicide attempts and postpartum depression.
4. Animal studies
Reproductive and teratogenicity studies with SSRIs were carried
out on mice, rats, rabbits, or dogs administered uoxetine paroxe-
tine or sertraline during organogenesis [39e41]. These studies did
not show an increased risk of teratogenesis even with high doses
that caused maternal toxicity, although there was some decrease in
neonatal survival and growth [41]. This is in contrast with effects
shown in rat and mouse whole embryo cultures [42e44] or cardiac
myocyte cultures [45]. For example, sertraline, uoxetine and
amitriptyline produced in cultured mouse embryos craniofacial
malformations [42]. Paroxetine (1 mM) was shown to decrease
serotonin-mediated proliferation of dissociated rat embryonic
cardiac myocytes [44]. Fluoxetine was found to adversely affect cell
viability and differentiation to cardiomyocytes at higher concen-
trations than achieved clinically in a dose-dependent manner using
mouse embryonic stem cell system [46]. In contrast with these in-
vitro studies, in-vivo animal studies have not supported an asso-
ciation between in-utero exposure to SSRIs and major anomalies
[39e41,46] or any interference with fetal hormonal status or
circadian rhythms [47]. Moreover, the lack of animal teratogenicity
of SSRIs in spite of the fact that serotonin plays an important role in
the prenatal development of the brain additionally supports the
notion that SSRIs are not human teratogens.
In summary, despite having more than 33 000 reported preg-
nancy outcomes after prenatal exposure to various SSRIs the
overall scientic evidence has not fullled the criteria for proof of
human teratogenicity of SSRIs, especially the seven criteria set by
Shepard [48]. Whereas the differences in the design of these
studies and the conicting results are confusing, new data strongly
suggest that the over-representation of cardiac anomalies may
reect an uncontrolled ascertainment bias. Additional studies have
merit only if they can include a comparison group of untreated
depressed women. In our view, the current data do not support
teratogenicity of SSRIs.
4.1. Clinical practice
There seems to be no increase in the rate of major congenital
anomalies in offspring of mothers using SSRIs during pregnancy.
Hence, there does not seem to be a clear indication to perform fetal
echocardiography in search for congenital heart defects.
5. Miscarriage, intrauterine growth restriction (IUGR),
preterm delivery, and stillbirth
Although most studies reported on the rate of major anomalies
following SSRI exposure, other possible pregnancy complications
were also presented. There was a slight increase in the miscarriage
risk in two meta-analyses [49e51]. However, in the prospective
cohort studies that were reviewed, corrections were not performed
for the effect of earlier gestational age at contact e which is an
important confounder e and the possible effects of the underlying
maternal disease were not ascertained [4,50]. SSRIs alone,
serotonin-norepinephrine reuptake inhibitors alone, and combined
use of antidepressants were all associated with an increase in
miscarriage, and the SSRIs paroxetine and venlafaxine were mainly
associated with increased miscarriage risk.
6. Prematurity and reduced birth weight
In a Finnish study, there was no increase in the rate of preterm
delivery, small for gestational age (SGA) or low birth weight [25].
The risk of both low birth weight and preterm delivery was
increased in infants who were born to mothers who had received
SSRI therapy [27]. Infants exposed to SSRIs had shorter gestations
and lower birth weights than non-exposed infants [51]. The
increased risk of low birth weight remained signicant, even when
maternal illness severity was accounted for. The adjusted OR for
preterm delivery was doubled in SSRI-exposed women compared
with two groups of women who had not used SSRIs during preg-
nancy, one with psychiatric history, and another without [52]. In
another study, the risk of preterm delivery was not signicantly
increased among SSRI users, but the risk of SGA offspring was
increased among women who continued SSRI use beyond the rst
trimester [53]. In a study from the Qubec Pregnancy Registry, no
association was found between SSRIs and the risk of SGA regardless
of trimester of exposure [54]. In other studies, there was an
increased risk for preterm delivery only among women exposed to
SSRIs in the second or third trimesters [15,54,55], or to benzodi-
azepines [56] with no increased risk for lowbirth weight or SGA. As
stated previously, the underlying psychiatric disorder is a potential
confounder in most of these studies.
In summary, associations were found in some studies between
the use of SSRIs during pregnancy and risk of miscarriage, IUGR or
preterm delivery. Most of these studies are potentially confounded
by the gestational age at initial contact and the underlying psy-
chiatric disorder.
6.1. Clinical practice
Due to the possibility that SSRIs may increase the rate of
spontaneous abortions, prematurity or IUGR, fetal growth and
well-being should be carefully monitored in the second half of
pregnancy.
7. Neonatal effects
Poor neonatal adaptation syndrome has been described initially
following prenatal exposure to various SSRIs, in the third trimester
of pregnancy in up to 30% of pregnancies [57e63]. Neonatal with-
drawal syndromes associated with SSRIs are characterized by irri-
tability, abnormal crying, tremor, respiratory distress, tachypnoea,
jitteriness, lethargy, poor tone or colour and rarely convulsions.
Most symptoms are mild and transient. These symptoms are
similar to those observed following late exposure to many psy-
chotropic drugs, opioids and alcohol.
8. Persistent pulmonary hypertension of the newborn
Several studies have suggested an association between maternal
use of SSRIs late in pregnancy and an increased risk of persistent
pulmonary hypertension of the newborn (PPHN) [15,64e67]. The
absolute risk of PPHN was <1%. None of the studies to date
described neonatal death from PPHN associated with SSRI. This is
different fromthe 10e15% mortality rate of PPHNfromother causes
(diaphragmatic hernia, cardiac malformations, aspiration). Other
A. Ornoy, G. Koren / Seminars in Fetal & Neonatal Medicine 19 (2014) 188e194 190
studies, possibly underpowered, did not nd such an association
[66] and in a recent study it was found that PPHN was associated
with the mode of delivery, specically caesarean delivery prior to
the onset of labour, but not with SSRI use in the second half of
pregnancy [67]. A recent Norwegian study found that untreated
depression increased the risk of PPHN.
In summary, considering the conicting results it is obvious that
if the use of SSRIs in pregnancy is associated with PPHN, the ab-
solute risk is <1% for PPHN, and should not be a reason to dis-
continue SSRI in late pregnancy.
8.1. Clinical practice
The newborn infant of mothers using SSRIs during the second
half of pregnancy should be followed up in the rst several days
post delivery to diagnose withdrawal symptoms or pulmonary
hypertension. Mothers should be told to look for clinical symptoms
that may appear even several days after their release from hospital.
9. Neurodevelopmental effects
Since SSRIs cross the human placenta [6,7], they are expected
to be able to reach the fetal brain. Indeed, as with other psycho-
tropic drugs, a relatively high percentage of the newborns of SSRI-
treated mothers demonstrate some degree of withdrawal symp-
toms. Hence, it may be expected that intrauterine exposure to
SSRIs and other psychotropic drugs might have long-term neu-
robehavioral and neurodevelopmental consequences. However,
since the postnatal environment where the child is raised has a
major effect on its development [68], one has to be careful not to
automatically ascribe such effects to in-utero exposure. In that
context it is important to note that even when the pregnant
mother is heroin dependent, the cognitive ability of the offspring
depends on the postnatal environment; in adopted, heroin-
exposed children that are raised in a favorable environment the
cognitive ability up to the age of 6 years is normal [69]. In contrast,
unexposed children raised in an unfavorable environment have
neurodevelopmental delay [69]. However, heroin-exposed chil-
dren developed a high rate of attention decit hyperactivity dis-
order at school age with specic behavioural consequences, in
spite of being raised in a good environment [70]. Hence, in order
to fully assess the effects on long-term development, children
must be followed up to adolescence and perhaps even to adult-
hood. Such long-term studies are missing for most psychotropic
drugs. Studies carried out on the neurodevelopmental effects of
SSRIs should also take into account the possible effects of maternal
mental illness on the development of the child.
The existing data regarding neurodevelopmental outcome are
conicting, as there seem to be positive and negative ndings.
A recent prospective study comparing the fetal motor behav-
iour of 96 women who took SSRIs, 37 women who discontinued
their SSRIs at the beginning of pregnancy and 130 controls
demonstrated that SSRIs during pregnancy affect the neuro-
behavioral development of the human fetus [71]. Fetuses exposed
to SSRIs exhibited dose-related increased motor activity in the
second and third trimester of pregnancy and disrupted sleep
especially in the third trimester. Whether these changes might
affect postnatal behaviour is not clear.
In contrast, several negative studies have been published. A
cohort of children of mothers exposed in pregnancy to uoxetine
(55 children), to tricyclic antidepressants (80 children) or unex-
posed were compared with an unexposed control group of 84
children and examined at 16 and 86 months of age. No differences
were found among the groups in their cognitive ability, language
development or behavior [72e74]. When the results were adjusted
to maternal IQ and socio-economic status, a negative correlation
was found between the childs IQ and language ability with the
duration of maternal depression and the number of depression
episodes [73]. No signicant differences in neurobehavioral scores
were found between children whose mothers were taking uoxe-
tine during pregnancy and non-exposed children [73]. In a later
study the same group also examined children of depressed mothers
who received no treatment or who were treated with venlafaxine
(an SNRI); they again found that the degree of maternal disease was
an important factor in the future development of the child [74].
Normal development was also observed in a small group of 11
children exposed in pregnancy to citalopram and followed up to 1
year [75]. In another prospective study developmental assessment
was performed at 2 and 8 months of age on 46 infants exposed
prenatally to SSRIs compared with 23 controls. No signicant dif-
ferences between SSRI-exposed and unexposed infants were found,
and there was no difference in the exposed infants between those
with poor neonatal adaptation (30%) and with normal neonatal
adaptation. Abnormal internalizing or externalizing behaviours
were observed in children of mothers with depression or anxiety
whether exposed or unexposed prenatally to SSRIs [76,77].
Klinger et al. [78] compared the neurodevelopmental outcome of
30 children born to mothers who used SSRIs in pregnancy and who
exhibited neonatal abstinence syndrome (NAS) with 52 children
who did not exhibit abstinence symptoms; no difference was found
between the two groups except a higher rate of social problems in
those with NAS. Although the intelligence of those without NAS was
higher than in those with NAS, the difference was not statistically
signicant. The authors concluded that there was no evidence of
cognitive impairment following prenatal exposure to SSRI.
Although the majority of studies showed no neuro-
developmental sequelae in children prenatally exposed to SSRIs,
there are some studies that showed slight developmental delay,
mainly in their motor function.
Mental developmental indices were similar in children whose
mothers were diagnosed with major depressive disorder treated or
untreated in pregnancy. However, children exposed to SSRIs scored
lower on the psychomotor developmental indices and the motor
quality factor of the behavioral rating scale compared with unex-
posed children [79]. In a follow-up study using a psychomotor
developmental test (Boel), abnormal test was more frequent in
children prenatally exposed to antidepressants compared with
unexposed [80]. In another neurobehavioral assessment study,
newborns prenatally exposed to SSRIs had abnormal outcomes
including increased motor activity, fewer changes in behavioral
state, and abnormal sleep patterns [81].
Childrens developmental milestones were assessed using a
questionnaire at 6 and 19 months of age. Second or third trimester
exposure to antidepressants was associated with delayed gross
motor developmental milestones, though still within normal range,
compared with unexposed children [82]. These children also closed
the developmental gaps with increasing age.
In summary, in most studies that focused on the possible neu-
rodevelopmental effects of prenatal SSRI exposure, there was no
conclusive evidence for an increased risk of adverse long-term ef-
fects. This can best be demonstrated by the study of Gentile [83]
who summarized data on negative and positive studies related to
the neurodevelopmental outcome of exposed children. Up to 2005
there were 11 studies of 306 children that found no effect of the
drugs and two studies with 81 children describing mainly slightly
abnormal motor development. We should, however, remember
that all neurodevelopmental studies related to SSRIs in pregnancy
report ndings in young children. Hence effects on learning abili-
ties, attention span and future mental illness may still be found in
spite of relatively normal early development.
A. Ornoy, G. Koren / Seminars in Fetal & Neonatal Medicine 19 (2014) 188e194 191
10. Possible association with autism spectrum disorders
(ASD)
Serotonin has been related to ASD for many years. Elevated
levels of serotonin in the platelets in about one-third of children
with ASD have been described by many investigators. Since sero-
tonin is an important neurotransmitter involved in fetal brain
development, the question whether prenatal exposure to SSRIs is
associated with ASDis pertinent [84]. However, there are only a few
studies showing such a possible association. In a recent population-
based caseecontrol study, a two-fold increased risk of ASD was
found with prenatal exposure to SSRIs [85]. The authors compared
the rate of prenatal exposure to antidepressants in 298 children
with ASD and in 1507 control children, nding that 6.7% of mothers
used antidepressants in the ASD group compared with only 3.3% in
the controls. No increased risk was found when mothers with
depression were not treated. In addition, a very recent study
showed that a history of maternal depression was associated with
an increased rate of ASD [86]. A similar increase was found
following the use of SSRIs or other antidepressants. The increase
was observed only in cases of high-functioning ASD. It seems from
these few studies that further studies are needed to verify the
suggested association which may well be because of the underlying
maternal disease.
10.1. Clinical practice
It is advisable that paediatricians of children exposed in utero to
SSRIs will keep in mind the possibility for neurodevelopmental
problems in these children.
11. Is there a reason to discontinue SSRIs during pregnancy?
11.1. Risk of treatment discontinuation
When evaluating the risk/benet ratio of SSRI treatment in
pregnancy, the risks associated with treatment discontinuation
should also be considered. Abrupt discontinuation of psychotropic
drugs in pregnancy may be associated with physical and psycho-
logical adverse effects [87]. SSRI treatment discontinuation during
pregnancy was also associated with a higher frequency of relapse
[88]. Depression is associated with an increased risk for preterm
delivery, and the risk of preterm delivery increases with increasing
severity of depression [88e91]. Treated women have lower
depressive symptom scores and better functioning. These risks
should be a factor in the decision-making with regard to treatment
continuation during pregnancy. Moreover, many of the develop-
mental studies have demonstrated the negative impact of maternal
depression on infant development, irrespective of the mode of
treatment during pregnancy.
12. Conclusions
Clinicians are faced with the difcult risk/benet consideration
of either making a recommendation to treat or not to treat maternal
depression or anxiety with SSRIs in pregnancy. In the eld of
teratology, decisions on new medications during pregnancy often
need to be made with insufcient experience on their safety in
human pregnancy. In the case of SSRIs in pregnancy, despite
extensive available studies on their use, quality is more important
than quantity, and data are still not conclusive.
It is clear that fetal well-being should be an important consid-
erationwhen decision for treatment is made. However, the decision
must also depend on the benet to the pregnant mother of the
treatment. It is also important to take into account the signicance
of maternal disease for the neurobehavioral development of the
child, because there is no question as to the importance of good
maternal physical and mental health to the childs development.
Hence, treatment of mothers with mental illness treated with SSRIs
should continue throughout pregnancy with frequent maintenance
evaluation of the mothers mental status, with proper adjustment
to the dose that should be the minimally effective dose.
In summary, most studies on the use of SSRIs during pregnancy
support the viewthat they are not major human teratogens. A small
increased risk for cardiovascular anomalies, especially with par-
oxetine, cannot be excluded. There appears to be a small increased
risk for miscarriages, which may be associated with the underlying
maternal condition. Neonates of mothers treated with SSRIs should
be closely followed up after delivery as there is an increased risk of
transitory neonatal effects. There is no conclusive evidence for
adverse long-term neurodevelopmental effects of prenatal SSRI
exposure. Discontinuation of treatment may pose risks, e.g. higher
frequency of relapse and increased risk of preterm delivery. Hence,
the general benet of treatment seems to outweigh the potential
small risk of untoward effects on the embryo, fetus or mother.
Conict of interest statement
None declared.
Funding sources
None.
References
[1] Gavin NI, Gaynes BN, Lohr KN, Melzer-Brody S, Gartlehner G, Swinson T.
Perinatal depression; a systematic review of prevalence and incidence. Obstet
Gynecol 2005;106:1071e83.
[2] Bennett HA, Einarson A, Taddio A, Koren G, Einarson TR. Prevalence of
depression during pregnancy: systematic review. Obstet Gynecol 2004;103:
698e709.
[3] Evans J, Heron J, Francomb H, Oke S, Golding J. Cohort study of depressed
mood during pregnancy and after childbirth. BMJ 2001;323:257e60.
[4] Diav-Citrin O, Ornoy A. Selective serotonin reuptake inhibitors in human
pregnancy: to treat or not to treat. Obstet Gynecol Int 2012;2012:698947.
[5] Cooper WO, Willy ME, Pont SJ, Ray WA. Increasing use of antidepressants in
pregnancy. Am J Obstet Gynecol 2007;196(544):e1e5.
[6] Hendrick V, Stowe ZN, Altshuler LL, Hwang S, Lee E, Haynes D. Placental
passage of antidepressant medications. Am J Psychiatry 2003;160:993e6.
[7] Rampono J, Simmer K, Ilett KF, Hackett LP, Doherty DA, Elliot R, et al. Placental
transfer of SSRI and SNRI antidepressants and effects on the neonate. Pharma-
copsychiatry 2009;42:95e100.
[8] Diav-Citrin O, Shechtman S, Weinbaum D, Wajnberg R, Avgil M,
Di Gianantonio E, et al. Paroxetine and uoxetine in pregnancy: a prospective
multicentre, controlled, observational study. Br J Clin Pharmacol 2008;66:
695e705.
[9] Klln B, Otterblad-Olausson P. Antidepressant drugs during pregnancy and
infant congenital heart defects. Reprod Toxicol 2006;21:221e2.
[10] Klln B, Otterblad-Olausson P. Maternal use of selective serotonin re-uptake
inhibitors in early pregnancy and infant congenital malformations. Birth De-
fects Res A Clin Mol Teratol 2007;79:301e8.
[11] Oberlander TF, Warburton W, Misri S, Riggs W, Aghajanian J, Hertzman C.
Major congenital malformations following prenatal exposure to serotonin
reuptake inhibitors and benzodiazepines using population-based health data.
Birth Defects Res B Dev Reprod Toxicol 2008;83:68e76.
[12] Pedersen LH, Henriksen TB, Westergaard M, Olsen J, Bech BH. Selective se-
rotonin reuptake inhibitors in pregnancy and congenital malformations:
population based cohort study. BMJ 2009;339:b3569.
[13] Merlob P, Birk E, Sirota L, Linder N, Berant M, Stahl B, et al. Are selective se-
rotonin reuptake inhibitors cardiac teratogens? Echocardiographic screening
of newborns with persistent heart murmur. Birth Defects Res A Clin Mol
Teratol 2009;85:837e41.
[14] Kornum JB, Nielsen RB, Pedersen L, Mortensen PB, Nrgaard M. Use of se-
lective serotonin-reuptake inhibitors during early pregnancy and risk of
congenital malformations: updated analysis. Clin Epidemiol 2010;2:29e36.
[15] Bakker MK, Kerstjens-Frederikse WS, Buys CH, de Walle HE, de Jong-van den
Berg LT. First trimester use of paroxetine and congenital heart defects: a
population-based caseecontrol study. Birth Defects Res A Clin Mol Teratol
2010;88:94e100.
A. Ornoy, G. Koren / Seminars in Fetal & Neonatal Medicine 19 (2014) 188e194 192
[16] Alwan S, Reefhuis J, Rassmussen SA, Olney RS, Friedman JM. Use of selective
serotonin-reuptake inhibitors in pregnancy and risk of birth defects. N Engl J
Med 2007;356:2684e92.
[17] Louik C, Lin AE, Werler MM, Hernndez-Daz S, Mitchell AA. First-trimester
use of selective serotonin-reuptake inhibitors and risk of birth defects. N Engl
J Med 2007;356:2675e83.
[18] Malm H, Artama M, Gissler M, Ritvanen A. Selective serotonin reuptake in-
hibitors and risk for major congenital anomalies. Obstet Gynecol 2011;118:
111e20.
[19] Klln B, Nilsson E, Olausson PO. Antidepressant use during pregnancy:
comparison of data obtained from a prescription register and from antenatal
care records. Eur J Clin Pharmacol 2011;67:839e45.
[20] Ericson A, Klln B, Wiholm B. Delivery outcome after the use of antide-
pressants in early pregnancy. Eur J Clin Pharmacol 1999;55:503e8.
[21] Chambers CD, Johnson KA, Dick LM, Felix RJ, Jones KL. Birth outcomes in
pregnant women taking uoxetine. N Engl J Med 1996;335:1010e5.
[22] McElhatton PR, Garbis HM, Elfant E, Vial T, Bellemin B, Mastroiacovo P, et al.
The outcome of pregnancy in 689 women exposed to therapeutic doses of
antidepressants. A collaborative study of the European Network of Teratology
Information Services (ENTIS). Reprod Toxicol 1996;10:285e94.
[23] Kulin NA, Pastuszak A, Sage SR, Schick-Boschetto B, Spivey G, Feldkamp M,
et al. Pregnancy outcome following maternal use of the new selective sero-
tonin reuptake inhibitors: a prospective controlled multicenter study. JAMA
1998;279:609e10.
[24] Hendrick V, Smith LM, Suri R, Hwang S, Haynes D, Altshuler L. Birth outcomes
after prenatal exposure to antidepressant medication. Am J Obstet Gynecol
2003;188:812e5.
[25] Malm H, Klaukka T, Neuvonen PJ. Risks associated with selective serotonin
reuptake inhibitors in pregnancy. Obstet Gynecol 2005;106:1289e96.
[26] Wogelius P, Nrgaard M, Gislum M, Pedersen L, Munk E, Mortensen PB, et al.
Maternal use of selective serotonin reuptake inhibitors and risk of congenital
malformations. Epidemiology 2006;17:701e4.
[27] Wen SW, Yang Q, Garner P, Fraser W, Olatunbosun O, Nimrod C, et al. Se-
lective serotonin reuptake inhibitors and adverse pregnancy outcome. Am J
Obstet Gynecol 2006;194:961e6.
[28] Wichman CL, Moore KM, Lang TR, St Sauver JL, Heise Jr RH, Watson WJ.
Congenital heart disease associated with selective serotonin reuptake inhib-
itor use during pregnancy. Mayo Clin Proc 2009;84:23e40.
[29] Byatt N, Deligiannidis KM. Freeman MP Antidepressants use in pregnancy: a
critical review focused on risks and controversies. Acta Psychiatr Scand
2013;127:94e114.
[30] Bar-Oz B, Einarson T, Einarson A, Boskovic R, OBrien L, Malm H, et al. Par-
oxetine and congenital malformations: meta-analysis and consideration of
potential confounding factors. Clin Ther 2007;29:918e26.
[31] Wurst KE, Poole C, Ephross SA, Olshan AF. Birth Defects Res A Clin Mol Teratol
2010;88:159e70.
[32] Einarson A, Koren G. First trimester exposure to paroxetine and prevalence of
cardiac defects: meta-analysis of the literature: unfortunately incomplete.
Birth Defects Res Part A Clin Mol Teratol 2010;88:588. author reply 589.
[33] Scialli AR. Paroxetine exposure and during pregnancy and cardiac malfor-
mations. Birth Defects Res A Clin Mol Teratol 2010;88:175e7.
[34] OBrien L, Einarson TR, Sarkar M, Einarson A, Koren G. Does paroxetine cause
cardiac malformations? J Obstet Gynaecol Can 2008;30:696e701.
[35] Addis A, Koren G. Safety of uoxetine during the rst trimester of pregnancy;
a meta-analytical review of epidemiological studies. Psychol Med 2000;30:
89e94.
[36] Einarson TR, Einarson A. Newer antidepressants and rates of major malfor-
mations; a meta-analysis of prospective comparative studies. Pharmacoepi-
demiol Drug Saf 2005;14:823e7.
[37] Rahimi R, Nikfar S, Abdollahi M. Pregnancy outcomes following exposure to
serotonin reuptake inhibitors: a meta-analysis of clinical trials. Reprod Toxicol
2006;22:571e5.
[38] Jimenez-Solem E, Andersen JT, Petersen M, Broedbaek K, Jensen JK, Afzal S,
et al. Exposure to selective serotonin reuptake inhibitors and the risk of
congenital malformations: a nationwide cohort study. BMJ Open 2012;2(3).
pii: e001148.
[39] Baldwin JA, Davidson EJ, Pritchard AL, Ridings JE. The reproductive toxicology
of paroxetine. Acta Psyciatr Scand 1989;350(Suppl.):37e9.
[40] Byrd RA, Markham JK. Developmental toxicology studies of uoxetine hy-
drochloride administered orally to rats and rabbits. Fundam Appl Toxicol
1994;22:511e8.
[41] Davies TS, Kluwe WM. Preclinical toxicological evaluation of sertraline hy-
drochloride. Drug Chem Toxicol 1998;21:521e37.
[42] Shuey DL, Sadler TW, Lauder JM. Serotonin as a regulator of craniofacial
morphogenesis: site specic malformations following exposure to serotonin
uptake inhibitors. Teratology 1992;46:267e78.
[43] Yavarone MS, Shuey DL, Tamir H, Sadler TW, Lauder JM. Serotonin and cardiac
morphogenesis in the mouse embryo. Teratology 1993;47:573e84.
[44] Sari Y, Zhou FC. Serotonin and its transporter on proliferation of fetal heart
cells. Int J Dev Neurosci 2003;21:417e24.
[45] Kusakawa S, Yamauchi J, Miyamoto Y, Sanbe A, Tanoue A. Estimation of
embryotoxic effect of uoxetine using embryonic stem cell differentiation
system. Life Sci 2008;83:871e7.
[46] Bengel D, Murphy DL, Andrews AM, Wichems CH, Feltner D, Heils A, et al.
Altered brain serotonin homeostasis and locomotor insensitivity to 3,4-
methylenedioxymethamphetamine (Ecstasy) in serotonin transporter-
decient mice. Mol Pharmacol 1998;53:649e55.
[47] Morrison JL, Rurak DW, Chien C, Kennaway DJ, Gruber N, McMillen IC, et al.
Maternal uoxetine infusion does not alter fetal endocrine and biophysical
circadian rhythms in pregnant sheep. J Soc Gynecol Investig 2005;12:356e64.
[48] Shepard T. Proof of human teratogenicity. Teratology 1994;50:97e8.
[49] Hemels MEH, Einarson A, Koren G, Lanctt KL, Einarson TR. Antidepressant
use during pregnancy and the rates of spontaneous abortions: a meta-anal-
ysis. Ann Pharmacother 2005;39:803e9.
[50] Einarson A, Choi J, Einarson TR, Koren G. Rates of spontaneous and therapeutic
abortions following use of antidepressants in pregnancy: results from a large
prospective database. J Obstet Gynaecol Can 2009;31:452e6.
[51] Nakhai-Pour HR, Broy P, Brard A. Use of antidepressants during pregnancy
and the risk of spontaneous abortion. Can Med Assoc J 2010;182:1031e7.
[52] Oberlander TF, Warburton W, Misri S, Aghajanian J, Herzman C. Neonatal
outcomes after prenatal exposure to selective serotonin reuptake inhibitor
antidepressants and maternal depression using population-based linked
health data. Arch Gen Psychiatry 2006;63:898e906.
[53] Lund N, Pedersen LH, Henriksen TB. Selective serotonin reuptake inhibitor
exposureandpregnancyoutcomes. ArchPediatr Adolesc Med2009;163:949e54.
[54] Toh S, Mitchell AA, Louik C, Werler MM, Chambers CD, Hernndez-Daz S.
Antidepressant use during pregnancy and the risk of preterm delivery and
fetal growth restriction. J Clin Psychopharmacol 2009;29:555e60.
[55] Ramos , St-Andr M, Brard A. Association between antidepressant use
during pregnancy and infants born small for gestational age. Can J Psychiatry
2010;55:643e52.
[56] Calderon-Margalit R, Qiu C, Ornoy A, Siscovick DS, Williams MA. Risks of
preterm delivery and other adverse perinatal outcomes in relation to
maternal use of psychotropic medications during pregnancy. Am J Obstet
Gynecol 2009;201(579):e1e8.
[57] Zajecka J, Tracy KA, Mitchell S. Discontinuation symptoms after treatment
with serotonin reuptake inhibitors: a literature review. J Clin Psychiatry
1997;58:291e7.
[58] Stahl MM, Lindquist M, Pettersson M, Edwards IR, Sanderson JH, Taylor NF,
et al. Withdrawal reactions with selective serotonin re-uptake inhibitors as
reported to the WHO system. Eur J Clin Pharmacol 1997;53:163e9.
[59] Stiskal JA, Kulin N, Koren G, Ho T, Ito S. Neonatal paroxetine withdrawal
syndrome. Arch Dis Child Fetal Neonatal Ed 2001;84:F134e5.
[60] Costei AM, Kozer E, Ho T, Ito S, Koren G. Perinatal outcome following third
trimester exposure to paroxetine. Arch Pediatr Adolesc Med 2002;156:
1129e32.
[61] Klln B. Neonate characteristics after maternal use of antidepressants in late
pregnancy. Arch Pediatr Adolesc Med 2004;158:312e6.
[62] Sanz EJ, De-las-Cuevas C, Kiuru A, Bate A, Edwards R. Selective serotonin re-
uptake inhibitors in pregnant women and neonatal withdrawal syndrome: a
database analysis. Lancet 2005;365:482e7.
[63] Levinson-Castiel R, Merlob P, Linder N, Sirota L, Klinger G. Neonatal abstinence
syndrome after in utero exposure to selective serotonin reuptake inhibitors in
term infants. Arch Pediatr Adolesc Med 2006;160:173e6.
[64] Chambers CD, Hernandez-Diaz S, Van Marter LJ, Werler MM, Louik C, Jones KL,
et al. Selective serotonin-reuptake inhibitors and risk of persistent pulmonary
hypertension of the newborn. N Engl J Med 2006;354:579e87.
[65] Klln B, Otterblad-Olausson P. Maternal use of selective serotonin re-uptake
inhibitors and persistent pulmonary hypertension of the newborn. Pharma-
coepidemiol Drug Saf 2008;17:801e6.
[66] Andrade SE, McPhillips H, Loren D, Raebel MA, Lane K, Livingston J, et al.
Antidepressant medication use and risk of persistent pulmonary hypertension
of the newborn. Pharmacoepidemiol Drug Saf 2009;18:246e52.
[67] Wilson KL, Zelig CM, Harvey JP, Cunningham BS, Dolinsky BM, Napolitano PG.
Persistent pulmonary hypertension of the newborn is associated with mode
of delivery and not with maternal use of selective serotonin reuptake in-
hibitors. Am J Perinatol 2011;28:19e24.
[68] Ornoy A. The impact of intrauterine exposure versus postnatal environment
in neurodevelopment toxicity: long term neurobehavioral studies in children
at risk for developmental disorders. Toxicol Lett 2003;140/141:171e81.
[69] Ornoy A, Michailevskaya V, Lukashov I, Bar-Hamburger R, Harel S. The
developmental outcome of children born to heroin dependent mothers, raised
at home or adopted. Child Abuse Negl 1996;20:385e96.
[70] Ornoy A, Segal J, Bar-Hamburger R, Greenbaum C. The developmental
outcome of school age children born to heroin-dependent mothers: impor-
tance of environmental factors. Dev Med Child Neurol 2001;43:668e75.
[71] Mulder EJH, Ververs FFT, de Heus R, Visser GHA. Selective serotonin reuptake
inhibitors affect neurobehavioral development in the human fetus. Neuro-
psychopharmacology 2011;36:1961e71.
[72] Nulman I, Rovet J, Stewart DE, Wolpin J, Gardner HA, Theis JG, et al. Neuro-
development of children exposed in utero to antidepressants drugs. N Eng J
Med 1997;336:258e62.
[73] Nulman I, Rovet J, Stewart DE, Wolpin J, Pace-Asciak P, Shuhaiber S, et al. Child
development following exposure to tricyclic antidepressants or uoxetine
throughout fetal life: a prospective controlled study. Am J Psychiatry
2002;159:1889e95.
[74] Nulman I, Koren G, Rovet J, Barrera M, Pulver A, Streiner D, et al. Neuro-
development of children following prenatal exposure to venlafaxine, selected
serotonin reuptake inhibitors, or untreated maternal depression. Am J Psy-
chiatry 2012;169:1165e74.
A. Ornoy, G. Koren / Seminars in Fetal & Neonatal Medicine 19 (2014) 188e194 193
[75] Heikinnen T, Ekblad U, Kero P, Ekblad S, Laine K. Citalopram in pregnancy and
lactation. Clin Pharmacol Ther 2002;72:184e91.
[76] Misri S, Reebye P, Kendrick K, Carter D, Ryan D, Grunau RE, et al. Internalizing
behaviors in 4-year-old children exposed in-utero to psychotropic medica-
tions. Am J Psychiatry 2006;163:1026e32.
[77] Oberlander TF, Reebye P, Misri S, Rapsdorf M, Kim J, Grunau RE. Externalizing
and attentional behaviors in children of depressed mothers treated with a
selective serotonin reuptake inhibitor antidepressant during pregnancy. Arch
Pediatr Adolesc Med 2007;161:22e9.
[78] Klinger G, Frankenthal D, Merlob P, Diamond G, Sirota L, Levinson-Castiel R,
et al. Long-term outcome following selective serotonin reuptake inhibitor
induced neonatal abstinence syndrome. J Perinatol 2011;31:615e20.
[79] Casper RC, Fleisher BE, Lee-Ancajas JC, Gilles A, Gaylor E, DeBattista A, et al.
Follow up of children of depressed mothers exposed or not exposed to anti-
depressants drugs during pregnancy. J Pediatr 2003;142:402e8.
[80] Mortensen JT, Olsen J, Larsen H, Bendsen J, Obel C, Srensen HT. Psychomotor
development in children exposed in utero to benzodiazepines, antidepres-
sants, neuroleptics, and anti-epileptics. Eur J Epidemiol 2003;18:769e71.
[81] Zeskind PS, Stephens L. Maternal selective serotonin inhibitor use during
pregnancy and newborn neurobehavior. Pediatrics 2004;113:368e75.
[82] Pedersen LH, Henriksen TB, Olsen J. Fetal exposure to antidepressants and
normal milestone development at 6 and 19 months of age. Pediatrics
2010;125:e600e8.
[83] Gentile S. SSRIs in pregnancy and lactation: emphasis on neurodevelopmental
outcome. CNS Drugs 2005;19:623e33.
[84] Harrington RA, Lee LC, Crum RM, Zimmerman AW, Hertz-Piccioto I. Serotonin
hypothesis of autism: implications for selective serotonin reuptake inhibitor
use during pregnancy. Autism Res 2013;6:149e63.
[85] Croen LA, Grether JK, Yoshida CK, Odouli R, Hendrick V. Antidepressant use
during pregnancy and childhood autism spectrum disorders. Arch Gen Psy-
chiatry 2011;68:1104e12.
[86] Rai D, Lee BK, Dalman C, Golding J, Lewis G, Magnusson C. Parental depression,
maternal antidepressant use during pregnancy and risk of autism spectrum
disorder; population based caseecontrol study. BMJ 2013;346:f2059.
[87] Einarson A, Selby P, Koren G. Abrupt discontinuation of psychotropic drugs
during pregnancy: fear of teratogenic risk and impact of counselling.
J Psychiatry Neurosci 2001;26:44e8.
[88] Cohen LS, Altshuler LL, Harlow BL, Nonacs R, Newport DJ, Viguera AC, et al.
Relapse of major depression during pregnancy in women who maintain or
discontinue antidepressant treatment. JAMA 2006;295:499e507.
[89] Dayan J, Creveuil C, Marks MN, Conroy S, Herlicoviez M, Dreyfus M, et al.
Prenatal depression, prenatal anxiety, and spontaneous preterm birth: a
prospective cohort study among women with early and regular care. Psy-
chosom Med 2006;68:938e46.
[90] Li D, Liu L, Odouli R. Presence of depressive symptoms during early pregnancy
and the risk of preterm delivery: a prospective cohort study. Hum Reprod
2009;24:146e53.
[91] Wisner KL, Sit DK, Hanusa BH, Moses-Kolko EL, Bogen DL, Hunker DF, et al.
Major depression and antidepressant treatment: impact on pregnancy and
neonatal outcomes. Am J Psychiatry 2009;166:557e66.
A. Ornoy, G. Koren / Seminars in Fetal & Neonatal Medicine 19 (2014) 188e194 194
Review
Abbreviated assessment of bisphenol A toxicology literature
Rochelle W. Tyl
*
RTI International, Hermann 124, 3040 Cornwallis Road, Research Triangle Park, NC 27709-2194, USA
Keywords:
Basic research studies
Bisphenol A
Good laboratory practices (GLPs)
Guideline studies
Non-monotonic doseeresponse (NMDR)
curves
Premature/neonatal sick infants and BPA
Risk assessment
Route-dependent BPA exposure/metabolism
s u m m a r y
Bisphenol A (BPA), synthesized in 1891, is produced in quantities of >2 million metric tons annually for
polycarbonate plastics, epoxy resins and food contact applications. BPA can be a weak estrogen mimic,
and is ubiquitous in humans (in 93% US population; in 96% US pregnant women). European/US food/drug
agencies conclude that current BPA levels present no risk to the general population (some include in-
fants/children); basic endocrine disruption (ED) researchers state that entire populations are at risk from
these levels. The US Food and Drug Administration banned BPA in baby bottles in 2012 not based on
safety concerns; the US Environmental Protection Agency and its UK counterpart concurred. Basic ED
researchers report reproductive/developmental effects from perinatal BPA exposure in mice at very low
doses, e.g. 2 ng/g body weight (0.002 mg/kg body weight), with non-monotonic doseeresponse (NMDR)
curves, using few animals per group and few groups; contract research organizations, in good laboratory
practice- and guideline-compliant large studies in rats and mice, report no low-dose effects or NMDR
curves. The argument rages!
2014 Elsevier Ltd. All rights reserved.
1. Introduction
Bisphenol A (BPA) was synthesized in 1891, recognized as an
articial estrogen in the 1930s, and was used briey to enhance
cattle/poultry growth and for estrogen replacement in women.
World BPA production was >2.2 million metric tons in 2009. US
consumption in 2003 was 856 000 tons: 72% polycarbonate plas-
tics, 21% epoxy resins, and 5% food-contact applications. BPAs
chemical name is 4,4
0
-dihydroxy-2,2-diphenyl propane, its formula
is (CH
3
)
2
C(C
6
H
4
OH)
2
(or C
15
H
16
O
2
), its CAS No. is 80-05-7, and its
molecular structure is shown in Fig. 1.
BPA, with hydroxyphenol groups, can act as a weak estrogen
mimic, and is considered by some to be an endocrine disruptor
(ED). ScienceDaily [1] reported detectable BPA in 96% of US preg-
nant women; a US Centers for Disease Control and Prevention
survey, 2003e2004, indicated that 92.6% of the US population aged
6 years had detectable urinary BPA [2].
2. Health effects and regulatory activities
European/US food/drug agencies conclude that current BPA
levels present no risk to the general population. Basic ED re-
searchers state that entire populations may suffer adverse health
effects from current BPA levels. The Endocrine Society [3] cited
adverse effects of endocrine-disrupting chemicals (EDCs), espe-
cially BPA. In 2012, the FDA banned BPA in baby bottles, stating that
this action was not based on safety concerns and that the agency
continues to support the safety of BPA for use in products that hold
food. In 2011, the Chief Scientist of the UKs Food Standards Agency
cited a 2011 EPA study with 20 human volunteers consuming three
meals of canned food (BPA lining), and tested hourly for urinary BPA
for 24 h, concluding: This [study] corroborates other independent
studies and adds to the evidence that BPA is rapidly absorbed,
detoxied (to non-estrogenic metabolites), and eliminated from
humans e therefore is not a health concern. EPA concurred that
BPA is not a health concern.
In 1998, vom Saal et al. [4] orally administered 2 or 20 ng BPA/g
body weight to pregnant mice on gestational days 11e17; 2 ng BPA/
g increased male offspring preputial gland size, but reduced
epididymal size (these organs developed fromdifferent embryonic
tissues, proffered as explanation), with no effects at 20 ng BPA/g
(non-monotonic doseeresponse: NMDR). Twenty nanograms BPA/
g body weight decreased efciency of daily spermproduction (DSP/
g testis) 20% in offspring males (but not DSP). Howdeshell et al. [5]
(vom Saals laboratory) reported that BPA accelerated mouse pu-
berty; this has yet to be conrmed.
Ryan et al. [6] (Grays EPA laboratory) reported that after peri-
natal exposures ethinyl estradiol (but not BPA) altered sexually
dimorphic behavior, puberty, fertility and anatomy of offspring
female LongeEvans (LE) rats. vom Saal et al. [7] responded that LE
* Tel.: 1 919 541 5972; fax: 1 919 541 6499.
E-mail address: rwt@rti.org.
Contents lists available at ScienceDirect
Seminars in Fetal & Neonatal Medicine
j ournal homepage: www. el sevi er. com/ l ocat e/ si ny
1744-165X/$ e see front matter 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.siny.2013.11.010
Seminars in Fetal & Neonatal Medicine 19 (2014) 195e202
rat was insensitive, and awed experimental design reveals the
need for guidelines requiring appropriate positive controls in
endocrine disruption research. Gray et al. [8] in a categorical
rebuttal stated that these rat strains are excellent models to study
potential xenoestrogens, and regulatory agencies consider their
studies useful for risk assessment, whereas many of the positive
low-dose studies were inadequate or . results have not been
replicated ..
Myers et al. [9] criticized good laboratory practices (GLPs) and
guideline-compliant studies, with BPA as case study, and argued
that GLPs cannot be used as criteria for selecting data for risk
assessment. Tyl [10] responded, using her three-generation BPA
study to compare strengths/weaknesses of basic exploratory
research vs guideline-compliant studies for hazard evaluation/risk
assessment. Tyl et al. [11] used Charles River (CR) SpragueeDawley
rats, with seven groups, 30/sex/group, dietary BPA at 0, 0.015e
7500 ppm(0, 0.001e500 mg/kg/day) to F0eF3 adults, exceeding US
EPA OPPTS 837.3800 test guidelines. Adult F0/F1/F2 systemic no
observed adverse effect level (NOAEL) was 75 ppm (5 mg/kg/day);
reproductive/postnatal NOAELs were 750 ppm (50 mg/kg/day), no
treatment-related effects in low-dose region (0.001e5 mg/kg/day),
and no NMDRs for any parameter, across generations, in either sex.
Since BPA exposure caused reproductive effects only in presence of
profound systemic toxicity at 7500 ppm (500 mg/kg/day), BPA was
not considered a selective rat reproductive toxicant.
Regulators appeared pleased with this study and used it in de-
liberations. A German EPA contingent visited Tyls laboratory in
2001 examining study data/documentation, reporting no ndings
on study protocol, performance, results or conclusions. Basic re-
searchers responded that CD rat was the wrong animal model,
insensitive to EDCs, and no concurrent positive control was used.
We chose CR/CD-1 mice (used by basic researchers) for our
follow-on BPA reproductive toxicity evaluation, performing one-
[12] and two-generation [13] dietary E
2
studies (six doses) under
OECD guidelines 415/416, to determine systemic, reproductive/
developmental parameters/estrogen responsivity in CD-1 mice. We
then performed a dietary BPA GLP/guideline-compliant CD-1
mouse two-generation study, with seven groups at 0, 0.018
(0.003 mg/kg/day, with reported low-dose effects) to 3500 ppm
BPA (600 mg/kg/day, with anticipated adult systemic toxicity), and
dietary E
2
(0.5 ppm; 0.08 mg/kg/day) as concurrent positive con-
trol, all with 28/sex/group.
E
2
conrmed CD-1 mouse estrogen sensitivity, increasing F0/F1
adult female uterus cervix vagina weights, accelerating pre-
weaning vaginal patency acquisition, decreasing normal/increasing
abnormal F1 female estrous cycles, and increasing absolute and/or
relative pituitary, thyroid (F1 only), liver, and kidney weights in F1/
F1 E
2
males. At 0.018e300 ppm BPA there were no effects on F0/F1/
F2 reproductive/developmental parameters; no NMDRs at any
dose. Systemic NOAEL was 30 ppm BPA (w5 mg/kg/day); repro-
ductive/developmental NOAELs were 300 ppm (w50 mg/kg/day);
effects only at 600 mg/kg/day (with profound systemic toxicity),
like rats. Nor was BPA considered a selective mouse reproductive/
developmental toxicant [14].
This time, basic researchers did not criticize our animal model,
but stated that something in study design (guideline-compliant
EPA/OECD design) and/or study execution affected study animals
(in Tyls opinion, a patently ridiculous accusation). Our rat/mouse
studies were audited in 2009 by the FDA, again with no ndings on
design, conduct, results or conclusions of either study.
In 2010, Stump et al. [15] reported a guideline-compliant
developmental neurotoxicity study of dietary BPA on gestational
day 0epostnatal day (PND) 21 in rats, at 0, 0.15e2250 ppm. F1
offspring exhibited no treatment-related neurobehavioral effects,
neuropathology, or altered brain morphometry, at any dose. Based
on maternal/offspring body weight decrements at 750 and
2250 ppm, systemic toxicity NOAEL was 75 ppm. There were no
treatment-related effects <75 ppm and no NMDRs at any dose.
Developmental neurotoxicity NOAEL was 2250 ppm; with no evi-
dence that BPA is a developmental neurotoxicant in rats.
Newborn mouse route-dependent BPA exposure/metabolism
was conrmed by Draganov et al. [16], dosing PND 3 female mice
once with [H
3
]BPA, 400 mg/kg, by gavage or subcutaneous injection,
50 mice/route. Blood was collected fromve pups/route/time-point
at 5, 10, 20, 30 min, 1 through 4, 6 or 24 h post dosing, and plasma
analysed for free BPA, BPA glucuronide (BPA-G) and BPA sulfate
(BPA-S). Neonatal female mice metabolized BPA to BPA-G, BPA-S
and other inactive metabolites, after both routes, but plasma free-
BPA was three-fold lower following oral dosing, with faster BPA
transport into/out of plasma; BPA-G was major component at every
time-point through 24 h; after subcutaneous injection, free BPA
was the major component at the earliest time-point; free BPA was
three-fold higher and total BPA C
max
four-fold higher vs oral dosing.
Doerge et al. [17] at the National Center for Toxicological
Research measured BPA serum pharmacokinetics in adult female
CD-1 mice following intravenous injection. Deuterated BPA
(100 mg/kg body weight) distinguished administered from envi-
ronmental BPA. BPA pharmacokinetics indicated rapid distribution
of unconjugated BPA from circulation (t
1/2
0.2 h) and terminal
elimination (t
1/2
0.8 h). They concluded that BPA serum/plasma/
tissue kinetics were consistent with rapid equilibrium and un-
derscore . non-persistent nature of BPA [17], and urinary BPA is
not appropriate for internal estimates, since BPA is rapidly conju-
gated, excreted, and rarely found in serum/plasma, although it is a
reasonable surrogate for ingested BPA [18,19].
vom Saal et al. published small studies in mice with low-dose
BPA (<50 mg/kg/body weight), reporting male reproductive organ
changes, including increased prostate weights [4,20]. The ACC
Bisphenol A Task Group and European Chemistry Industry Council
responded by funding robust studies to replicate vomSaals work in
rats/mice. Neither study found low-dose/NMDR BPA effects in
systemic/reproductive organs, including no increased prostate
weights, in rats or mice [11,14]. Basic researchers expressed
concern that funding might have affected study outcomes!
In 2006, the National Institute of Environmental Health Sciences
(NIEHS) sponsored a meeting of thirty-eight of the worlds leading
scientic experts on bisphenol A
0
(all basic researchers), who
assessedcondence levels for BPAndings inmore than700studies;
assessments were integrated into a consensus statement [21].
Also in 2006, the National Toxicology Program (NTP, NIEHS)
began a BPA review within CERHR (Center for the Evaluation of
Risks to Human Reproduction). CERHR convened scientists, inten-
tionally not BPA experts, who applied selection criteria for scientic
literature to include, which excluded some academic studies; aca-
demic scientists accused them of bias in . selection procedures.
CERHR [22] issued draft conclusions in August 2007, reporting
reasons for concern about BPAs impact on neurodevelopment in
fetuses/newborns at estimated current human exposure range, but
little concern about reproductive effects. Academic BPA experts
responded [23]. The nal NTP-CERHR Expert Panel Report on BPA
was published in 2008 by Chapin et al. [24] e an excellent, thor-
ough review.
Fig. 1. Molecular structure of bisphenol A.
R.W. Tyl / Seminars in Fetal & Neonatal Medicine 19 (2014) 195e202 196
In 2003, Hunt et al. [25] reported increased meiotic distur-
bances, including aneuploidy, in mouse oocytes, after exposure to
damaged caging material from inadvertent use of harsh detergent.
They repeated results by intentionally damaging cages and water
bottles, and later administering BPA daily, to conrm that low BPA
levels (as released from damaged cages, etc.) disrupted female
meiosis. They reported that effects were dose dependent, induced
by environmentally relevant BPA doses, and suggested that BPA is a
potent meiotic aneugen.
In 2008, Pacchierotti et al. [26] re-evaluated BPA aneugenic ef-
fects on mouse germ/bone-marrow cells after acute (one BPA dose,
six or seven daily administrations), subchronic (7-week BPA
drinking water), or chronic oral/dietary exposure. They reported no
signicant induction of hyperploidy/polyploidy . in oocytes and
zygotes at any treatment condition; only increased metaphase II
oocytes with prematurely separated chromatids after chronic
exposure, with no irreversible consequence upon . delity of
chromosome segregation during . second meiotic division. Male
mice exhibited no delay of meiotic divisions and no indication of
hyperploidy and polyploidy after 6 exposure days; daily oral BPA
did not increase micronucleus frequencies in mouse bone-marrow
polychromatic erythrocytes. They concluded that our results do
not add evidence to .suspected aneugenic activity of BPA .other
factors or co-factors should be considered to explain the unex-
pected burst of meiotic abnormalities previously attributed to
accidental BPA exposure.
In 2009, Hunt et al. [27] published a mini-review: Bisphenol A
experience, including aneuploidy, disrupted mouse oogenesis,
male/female reproductive tract decits, abnormal fetal oocytes,
pregnancy complications, mammary/prostate gland morphologic
changes, increased adult malignancies, and altered brain sexual
development at very low BPA doses, <50 mg/kg/day, the EFSA TDI,
FDA ADI, and EPA reference dose; none was subsequently
conrmed.
Many multi-generation studies have reported no low-dose BPA
adverse reproductive effects. Hunt et al. [27] states: These multi-
generation tests have the advantage of being conducted under
strict quality control guidelines for Good Laboratory Practice (GLP),
enabling regulators to document data quality in courts of law.
However, adherence to GLP guidelines .provides no assurance of
the soundness of the data analysis and interpretation. Importantly,
for EDC testing, standard multi-generation test protocols have well-
recognized limitations, using methodologies developed long ago.
Effects detectable by contemporary techniques, e.g. immunouo-
rescence, reverse transcriptaseepolymerase chain reaction and
methylation assays, would likely go unnoticed.
The present author was an invited member of the EPA Endocrine
Disruptor Screening and Testing Advisory Committee (EDSTAC),
1996e1998, which led to the EDS Program; (http://epa.gov/endo/
index/htm), and revised/new test guidelines and protocols. Revised
multi-generation protocols include more endocrine-sensitive out-
comes (e.g. estrous cyclicity, pubertal landmarks, sperm measures,
more detailed histopathologic reproductive assessments, hormone
analyses, etc.) [US Environmental Protection Agency, Health Effects
Test Guidelines, OPPTS 870.3800, Reproduction and Fertility Effects,
August, 1998 (http://www.epa.gov/opptsfrs/publications/OPPTS_
Harmonized/870_Health_Effects_TestGuidelines/Series/870-3800.
pdf), and OECD 416].
In 2008, Calafat et al. [2] examined urine samples from 41 pre-
mature sick infants in two hospital neonatal intensive care units
(NICUs) for free/conjugated BPA. BPA was detected in all urine
samples (LOD 0.4 mg/L), >95% conjugated in >75% premature in-
fants, with no differences in free/conjugated BPA concentrations
from breastfed (or fed breast milk) (n 4) vs formula-fed (n 14)
infants.
Teeguarden and Hanson-Drury [28] at the DOE Pacic North-
west National Laboratory examined 150 human exposure studies
(>30 000 individuals in 19 countries, including women/infants),
and 130 toxicity studies which reported use of BPA low doses
(most 10e1000-fold higher than human exposures), concluding
that biologically active BPA is at such low blood concentrations
(parts per trillion) that it is beneath toxicologists current ability
to detect it (which raises concerns about possible sample
contamination in studies reporting high BPA blood levels), and that
BPA exposure [is] too low to cause problems.
Concerns that BPA contributes to estrogenic compound mixture
effects were also considered unfounded, since human BPA expo-
sures are very low and BPA is completely/almost completely/
rapidly conjugated in gut/liver before access to general circulation
and other organs [29], even in sick NICU newborns [30].
Some proponents of inverted U-shaped NMDR curves believe
that BPA effects occur at very low/very high doses (absent/dimin-
ished at intermediate doses), which contribute to non-repeat-
ability of some studies; and that small original research studies
were correct and many, more robust, repeat studies were wrong,
for a variety of design reasons [9,31].
Sharpes [32] comments are germane: Fundamental, repetitive
work on bisphenol A has sucked in tens, probably hundreds, of
millions of dollars from government bodies and industry which, at
a time when research money is thin ., looks increasingly like an
investment with a nil return .I hope that [this opinion piece] will
remind us all of the central importance . attached to . repeat-
ability of experiments and how we should react when a study
proves to be unrepeatable. As scientists, we all like our ideas and
hypotheses to be . correct; yet, there is equal merit in being .
wrong. This is the .trusted way via which scientic understanding
moves onward . our own convictions and presumptions cannot
stand in its way.
Newbold et al. [33] reported that subcutaneous injections to CD-
1 mice on PND 1e5 at 0e1000 mg BPA/kg/day resulted in cystic
ovaries in all groups; increased incidence only at 100 mg BPA/kg/
day. In 2009, Newbold et al. [34] used the same study design and
mouse strain, but administered BPA on gestational days 9e16 at 0e
1000 mg/kg/day, reporting offspring cystic ovaries in all groups,
increased incidence only at 1 mg/kg/day (P < 0.05).
In 2010, Beronius et al. [35] examined 10 BPA risk assessments,
concluding that differences in conclusions regarding human health
risks of BPA between risk assessments are largely due to differing
opinions concerning the reliability and relevance of studies
reporting effects of low-doses. Governmental agencies stated that
low-dose data were not reliable or relevant; Chapel Hill experts
concluded that (their) low-dose data were reliable and relevant.
European Chemicals Bureau (2003), European Commission Health
& Consumer Protection Directorate-General (2002), Opinion of
Scientic Committee on Food on BPA (2002) and NTP CERHR BPA
Monograph (2008) concluded that, although these data were not
sufciently reliable or relevant, they cannot be dismissed or
deemed irrelevant for human health assessment (!). Beronius et al.
[35] concluded only that differences in risk assessment policy and
process expert judgment resulted in drastically different conclu-
sions e from risk to the entire population to no risk to any part of
the population e neither likely nor reasonable.
In a critical evaluation of key evidence on . human health
hazards of exposure to bisphenol A, the Advisory Committee of the
German Society of Toxicology [36] stated that criticism of Tyl rat
and mouse multi-generation studies was scientically not justied
and recently published additional data further support the reli-
ability of the two- and three-generation studies demonstrating lack
of estrogen-dependent effects at/belowdoses on which current TDI
is based. They stated positive results from some explorative
R.W. Tyl / Seminars in Fetal & Neonatal Medicine 19 (2014) 195e202 197
studies have not been conrmed in subsequent studies with higher
numbers of animals or a priori dened hypotheses, and rodent
data can well be used as a basis for human risk evaluation. They
concluded: available evidence indicates that BPA exposure repre-
sents no noteworthy risk to the health of the human population,
including newborns and babies [36].
Viguie et al. [37] evaluated fetal BPA exposure and consequences
on mother/newborn thyroid functions in sheep (gestation/thyroid
regulation similar to humans). Pregnant ewes were injected sub-
cutaneously daily with 5 mg/kg/day BPA, gestational day 28 till
term. Unconjugated BPA was low/similar in pregnant ewes, new-
borns and mothers. BPA glucuronide was w1300-fold higher than
BPA in amniotic uid/cord blood. Total thyroxine (T
4
) was
decreased 30% in BPA-exposed ewes and newborn cord/jugular
blood, with free plasma T
4
reduced in newborn jugular vein blood;
therefore, BPA prenatal exposure is associated with newborn hy-
pothyroidism. Viguie et al. suggest that this effect, if conrmed
would constitute a major issue for BPA risk assessment. This
author could nd no subsequent publications at other doses, no
control/conrmatory data, no information on offspring conse-
quences, etc.
3. Epigenetics
Epigenetics examines gene expression changes/cellular pheno-
type, by mechanisms other than DNA sequence changes (muta-
tions), controlling gene activation/inactivation. Epigenetic changes
occur normally during cellular differentiation as initially totipotent
cells become pluripotent and then specialized/differentiated, some
stable and transmitted to subsequent generations.
Since thyroid hormone (TH) is closely associated with brain
development/function and BPAwas reported to affect TH, Yaoi et al.
[38] examined low-dose BPA effects on mouse forebrain develop-
ment, performing genome-wide analysis study (GWAS) of epi-
genomic alterations on embryonic day 12.5/14.5 brains, after
maternal daily 20 mg BPA/kg/day subcutaneous injections starting
day E0, using telencephalons from 15 (genomic DNA) and six fe-
tuses (RNA preparation) from three dams/group (treated/not
treated), and restriction landmark genomic scanning to detect DNA
methylation changes of cytosine residues in CPG sites of epigenetic
modications. They scanned CpG methylation at 2500 Notl loci
[cleavage sites in CpG islands (CGI)], representing 48 (de)methyl-
ated unique loci: reporting methylation status at most loci was
primarily developmental-stage specic; mRNA expression of two
functionally related genes changed with development, and with
prenatal BPA exposure. In both genes, changes in transcription
correlated well with changes in Notl methylation status. The au-
thors concluded that epigenetic alterations in promoter-associated
CGI after BPA exposure may underlie some effects on brain
development.
Epigenetic modications (acetylation/methylation histone
marks, DNA cytosine nucleotide methylation) correlate with (con-
trol?) transcriptional status of many (all?) genes [39]. Skinner et al.
[40] stated that most environmental toxicants do not alter DNA
sequences, but do alter the epigenome. If environmental toxicants
(e.g. EDCs) alter the somatic cell epigenome, this may cause/pro-
mote disease in exposed individual(s), but not transmitted to next
generation(s). But if toxicants modify germ-line epigenome, then
altered epigenome, and promoted disease, can/may be transmitted
to subsequent offspring generations.
Many diseases cannot be explained by Mendelian genetics.
Identical twins (with essentially identical genotypes) can have
different disease frequencies. The percentage of diseases from ge-
netic abnormalities is small; e.g. 5% of breast cancers are from
known genetic mutations. Most cancers have no known genetic
cause, and almost half of mutated tumor-suppressor genes causing
familial cancers are inactivated by promoter-region epigenetic
hypermethylation. Recent GWAS reported <1% disease incidence
from specic genetic polymorphisms. Environmental exposures
must be considered in conjunction with epigenetic mechanisms.
Other non-Mendelian diseases appearing familial, but not involving
classic genetic transmission, include breast/prostate cancer, dia-
betes, autism, neurological disorders, and metabolic disease.
Nutrition and environmental toxicants, especially endocrine dis-
ruptors can affect/promote disease [40]. Disruption of normal DNA
methylation patterns and/or post-translational covalent histone
modications are implicated in various cancers, from arsenic
exposure, and, putatively, BPA-induced epigenetic modications
resulting in mouse developmental alterations (see below).
Dolinoy et al. [41] suggested that fetal origins of adult disease
include epigenetic modications from developmental exposures,
which may/can inuence adult disease susceptibility; since peri-
natal BPA exposure is associated with adult obesity, breast/pros-
tate cancer and altered reproductive function, perhaps epigenetic
modications are responsible e especially since epigenome is
particularly susceptible to dysregulation during embryogenesis
when normal developmental DNA methylation/chromatin
patterning occur. Therefore, they examined maternal BPA effects on
fetal epigenome by feeding female a/a mice a phytoestrogen-free
diet 50 mg BPA/kg for 2 weeks premating, mating with A
vy
/a
males, through gestation and lactation; 16 control and 17 BPA-
exposed litters. Agouti gene in viable A
vy
mice is a metastable epi-
allele, epigenetically affecting black eumelanin (a) or yellow
phaeomelanin (A) production; epigenetic cytosine methylation at
CpG sites varies among genetically identical A
vy
/a mice. Maternal
dietary BPA did not affect litter size, pup survival, weaning weight,
genotype or sex ratio, but it shifted coat color distribution of
genetically identical PND 22 A
vy
/a offspring toward yellow coat
color phenotype (P 0.007); 21% BPA-exposed offspring were
classied as yellow vs 10% controls; 9.6% BPA-exposed offspring
were classied as pseudo-agouti vs 18.3% controls. Authors
measured DNA methylation at CpG sites in promoter region of A
vy
IAP offspring; BPA decreased (P 0.004) mean percentage
methylation at these sites. They concluded that methylation at A
vy
IAP . mediates BPA effect(s) on A
vy
/a coat color; prenatal BPA
promotes offspring DNA hypomethylation at multiple loci with
metastable epialleles. To evaluate effects of maternal nutritional
supplements on offspring hypomethylated-epigenome from BPA,
additional pregnant a/a mice, exposed to dietary 50 mg BPA/kg,
were fed diets supplemented with methyl donors folic acid, B
12
,
betazine, and choline (14 litters), or genistein (13 litters) during
prematingeperinatal BPA exposures. Neither supplemented diet
affected litter size, pup survival, weanweight, genotypic ratio or sex
ratio, but both supplemented diets restored coat color distribution
of BPA-exposed offspring to that in control litters. Maternal nutri-
tional supplementation negated BPA-induced hypomethylation;
CpG methylation at A
vy
IAP of BPA-exposed/methyl donor-supple-
mented offspring was equivalent to control offspring methylation
pattern (P 0.25). Although genistein is not a methyl donor at
mouse levels comparable with humans on high soy diets, it
counteracted BPA-induced hypomethylation (P 0.46), presum-
ably by hypermethylating offspring genome.
Salian et al. [42] reasoned that e since steroid hormones/re-
ceptors are critical to spermatogenesis, steroid receptor co-
regulators play major roles in steroid receptor functioning, and
BPA is reported to impair spermatogenesis e co-regulator expres-
sion changes should be associated with impaired spermatogenesis.
They investigated whether perinatal BPA exposure changed
testicular steroid receptor/coregulator expression proles. Preg-
nant F0 female rats (eight per group) were gavaged daily at 0, 1.2 or
R.W. Tyl / Seminars in Fetal & Neonatal Medicine 19 (2014) 195e202 198
2.4 mg BPA/kg/day, on gestational age 12ePND 21; adult F1eF3
offspring testes were evaluated by immunohistochemical locali-
zation of testicular receptors co-activator-1 (SRC-1), G-receptor
integrating protein-1 (GRIP-1), p300/CBP/co-integrator-associated
protein (p/CIP) and nuclear co-repressor (NCoR). Authors re-
ported reduced SRC-1 and NCoR and increased p/CIP and GRIP-1
expression in F1 testes (exposed prenatally) and a similar
pattern in F2/F3 offspring testes (not exposed). Any consequences
were not reported. They concluded that perinatal exposure of male
rats to BPA leads to transgenerational perturbations in . expres-
sion prole of testicular steroid receptor coregulators.
Wolstenholme et al. [43] characterized F1 offspring social be-
haviors fromF0 mice exposed to dietary 5 mg BPA/kg/day from7 to
10 days pre-mating till gestational day 18.5, when six F0 dams were
terminated (three control/three BPA-exposed) and their F1 fetuses
extracted, frozen, F1 brains and later F4 fetal (four control and four
BPA litters) brains collected. Plasma BPA levels in exposed F0 dams
were similar to human exposure ranges. Remaining pregnant F0
BPA dams delivered; F1 offspring were fostered to just-delivered
control dams, and maintained on control diet through 4th
offspring generation. The authors evaluated BPA-exposed F1 and
unexposed F4 offspring for social interactions, elevated plus maze
and social preference tests. F1 prenatally BPA-exposed pups
decreased social interactions (social, investigative and play solicit-
ing), increased elevated plus maze activities, and social preference
tests, but unexposed F4 offspring (from BPA-exposed F0 dams)
increased social behavior, decreased non-social behavior and
exhibited no effects on play-soliciting/investigative behaviors. F1
BPA-exposed embryonic brains on gestational day 18.5 had reduced
transcript levels for oxytocin and vasopressin; decreased vaso-
pressin transcript levels persisted to both F4 sexes; oxytocin tran-
script was reduced in F4 males only, with no signicantly altered
transcripts for interaction between diet and sex. The authors
concluded that doses of BPA, only during gestation, had immediate
and long-lasting transgenerational effects on mRNA in brain and
social behaviors. In this authors opinion, small sample sizes, dia-
metrically opposite behavioral responses (social/non-social) in
exposed F1 versus unexposed F4 offspring, and no signicant sex/
diet transcript changes, do not provide convincing evidence of
BPAs supposed transgenerational effects (although persistent
transcript changes are interesting).
The NTP CERHR BPA monograph [24] and FDA Update [44] indi-
cated some concern about the potential effects of BPA on the brain,
behavior and prostate gland in fetuses, infants and young children.
Wolstenholme et al. [45] discussed role of Bisphenol A in shaping
the brain, epigenome and behavior, reiterating work that prenatal
BPA exposures advanced puberty, increased prostatic growth,
altered pubertal mammary gland development, and permanently
changed morphology/functionality of mouse female reproductive
tract and ovaries, although none has been conrmed. BPA was sug-
gested to compromise brain sexual development [46,47].
Kundakovic and Champagne [48] suggested that molecular
mechanisms underlying long-lasting effects from perinatal BPA
exposure, likely involve disruption of epigenetic programming .
of gene expression during development, resulting in altered brain
development, sexual differentiation (sexual dimorphism), neuro-
biological behaviors (anxiety, response to novelty, learning and
memory), and immune function; effects could extend to subse-
quent generations (see above). They suggested that since proposed
molecular mechanisms may be altered by maternal gestational
BPA exposure, resulting in altered expression of specic offspring
genes, it is important to know whether BPA epigenetic effects (not
yet conrmed) are related to its estrogenic activity, and to deter-
mine which downstream effector proteins could mediate DNA
methylation changes by BPA (not yet conrmed), highlighting basic
research studies indicating effects of prenatal BPA exposure on
offspring brain, behavior and immune system, role of epigenetic
pathways in these effects, and transgenerational implications of
EDC exposures (none yet conrmed). This author concurs only that
more work should be done on the role(s) of epigenetic program-
ming and disruption consequence(s).
Basic researchers state that parental BPA exposure can impact
the next generation e a transgenerational effect. However, if F0
dams are exposed during their pregnancies, F1 offspring are
exposed in utero, so F1 effects are not transgenerational; to conrm
transgenerational effects the effect(s) would have to occur in F2 or
later generations with no direct/indirect offspring BPA exposures.
The NIH scientic panel concluded that, due to lack of repro-
ducibility of low-dose experiments, they had minimal concern for
BPA reproductive effects, but did have higher concern for neural
and behavioral BPA effects [24].
In-vitro/small basic research studies report adverse outcomes,
many exhibiting NMDRs in studies with more than two doses, non-
relevant subcutaneous/intravenous injection, and doses 100e10
000-fold those of human exposures. Studies reporting higher than
expected BPA levels in animal blood, urine, tissues, etc., may have
contamination from damaged plasticware: BPA, released into food/
water, taken up by animals, and rapidly glucuronidated (and/or
sulfated) in gastrointestinal tract/liver, before systemic exposure.
Even in sick premature NICU newborns, over 95% of the BPA
detected in urine was glucuronidated in most of the babies [2].
Perinatal BPA reportedly altered DNA methylation status and
subsequent malignancy development in offspring breast and
prostate; Ho et al. [49] considered that DNA methylation may be
one mechanism for reprogramming early prostate; hypo-
methylation of phosphodiesterase type IV variant 4 (PDE4D4)
promotor, in rat prostate after neonatal BPA at an environmentally
relevant dose, was associated with increased susceptibility to
prostate carcinogenesis. Chen and Tang [50] concluded that of
particular importance will be the exploration of molecular and
behavioral changes that occur in response to environmentally
relevant low doses of BPA in animals and humans.
Yeo et al. [51] evaluated early development of mammalian
cortical neurons, when high intracellular chloride levels are pre-
sent. As neurons mature, chloride levels drop w75% because po-
tassium chloride co-transporter protein, Kcc2, removes chloride
ions from cells. If chloride remains high, this can damage neural
circuits and compromise ability of developing nerve cells to migrate
to proper brain positions. When cortical neurons were exposed
in vitro to 100 nM BPA, chloride was reduced by only w50%, with
greater effects in female neurons. The authors proposed that BPA
shuts down Kcc2 gene, reducing Kcc2 protein, delaying chloride
removal. In addition, MECP2, a gene which makes MECP2 protein
important for normal brain function, is also BPA responsive. When
exposed to BPA in vitro, neural MECP2 was higher and bound to
Kcc2 gene at a higher rate, which might . shut Kcc2 down. BPA
also increased chloride in migrating neurons (female neurons more
affected). Authors concluded that BPA disrupts this gene expres-
sion by epigenetic mechanisms ., which could . play a role in
pathogenesis of human neurodevelopmental disorders. Their
ndings raised the question (at least to ScienceDaily [1]) whether
BPA contributes to neural developmental disorders such as Rett
syndrome, a severe autism spectrum disorder in girls, with MECP2
gene mutations [52]. Experts noted (Round-Up [53]), this study
was in vitro with BPA administered directly to neurons, at one dose,
hundreds to thousands of times higher than humans would be
exposed to, and therefore would not provide any linking/sup-
porting evidence because of lack of credibility with the dose.
The European Food Safety Authority [54] stated that human BPA
risks during perinatal period are minor because rapid metabolic
R.W. Tyl / Seminars in Fetal & Neonatal Medicine 19 (2014) 195e202 199
clearance by rst-pass glucuronide metabolism minimizes human
internal exposure to free BPA vs rodents in which low-dose effects
have been observed. However, Ginsberg and Rice [55] argued that
BPA glucuronide deconjugation to active BPA occurs prenatally by
placental (and other tissue) b-glucuronidase, and BPA sulfate
deconjugation by arylsulfatase C to active BPA starts early post
partum in liver/gut; biomonitoring studies and laboratory experi-
ments document free BPA in rat and human maternal, placental and
fetal tissues indicating human BPA exposure is not negligible; risk
assessors should dene human BPA doseeresponse with pharma-
cokinetic models including deconjugation.
There is growing concern that prenatal BPA exposure may affect
brain development, behavior and emotions. Studies have suggested
that rodent fetal and/or lactational BPA exposure alters adult
offspring behavior [56,57], possibly caused by neurotransmitter
changes [58,59].
4. BPA exposures during pregnancy in children
Braun et al. [60] investigated the impact of gestational/child-
hood BPA on behavior/executive function of 3-year-olds, using a
prospective Cincinnati cohort of 244 mothers and their children.
They characterized gestational/childhood BPA using maternal BPA
urinary concentrations during gestation, at delivery, and in 1-, 2-
and 3-year-olds. (Urinary BPA is a reasonable surrogate for ingested
BPA, but not for internal exposure [18].) BPA was detected in >97%
of maternal gestational/childhood urine samples. Each 10-fold in-
crease in maternal gestational urinary BPA concentration was
associated with more anxious and depressed behavior in daugh-
ters; null/negative changes in sons. Interestingly, the authors
concluded that associations between childhood BPA and neuro-
behavior were largely null and not modied by child gender.
Selgrade et al. [61] identied one study relating maternal uri-
nary BPA during pregnancy and delivery, and childhood wheeze
frombirth to 3 years [62]. Maternal gestational urinary BPAwas not
associated with child wheeze at birth to 3 years. Whenprenatal BPA
was separated into above and below the median, BPA above the
median was positively associated with offspring wheeze at 6
months, but not at 3 years. Maternal prenatal urinary BPA at 16
weeks had borderline association with offspring wheeze, but
offspring wheeze was not associated with maternal urinary BPA at
26 weeks of gestation or at delivery. Selgrade et al. considered these
effects to be borderline and transient at best.
CHAMACOS (Center for Health Assessment of Mothers and
Children of Salinas) longitudinal birth cohort study examined
environmental exposures/health among women/children of Sali-
nas, California, beginning in 1999e2000. In 2013, Chevrier et al.
[63] reported on maternal pregnancy urinary BPA and maternal/
neonatal thyroid function, measuring maternal urinary BPA during
rst/second half of pregnancy, and free thyroxine (T
4
), total T
4
, and
thyroid stimulating hormone (TSH) in pregnant women and
offspring. They reported no statistically signicant association
between the mean of two pregnancy BPA measurements and
maternal thyroid levels, but BPA measurement closer to thyroxine
measurement was signicantly associated with reduced maternal
T
4
. Maternal BPA was associated with reduced TSH in sons (not
daughters) especially when BPA was measured in third trimester.
The authors concluded that exposure to BPA during pregnancy .
reduced total T
4
in pregnant women and decreased TSH in male
neonates, a nding which may have implications for fetal and
neonatal development. (T
4
was reduced only in mothers, and is
usually associated with increased TSH which stimulates T
4
syn-
thesis e not with decreased TSH).
A second CHAMACOS cohort paper [64] evaluated effects of
prenatal BPA on childrens body mass. They reported that offspring
girls, age 9 years, exposed to highest prenatal BPA concentrations,
had lower body weights for their heights and were less likely to be
obese, vs girls with lowest prenatal BPA exposures; boys were
unaffected. Nine-year-olds showed a positive association between
current BPA urinary concentrations and current obesity. The au-
thors suggested that these obese children may have higher urinary
BPA levels because they eat more packaged foods (packaging con-
taining BPA) and/or they ingest more calories, so higher BPA levels
could be a result of dietary habits resulting in obesity, rather than
being a cause of obesity. The authors will follow these children
through puberty.
Columbia Center for Childrens Environmental Health recruited
568 pregnant women; mothers during third trimester pregnancy
and their children at ages 3, 5, and 7 years provided spot urine
samples. Wheeze in previous year was assessed by questionnaires
at 5e7 years; asthma determined by a physician once at 5e12 years.
FENO (fractional exhaled nitric oxide, measuring airway inam-
mation) was measured at 7e11 years. Maternal pregnancy urinary
BPA concentrations were associated inversely with offspring
wheeze at 5 years; offspring urinary BPA concentrations at 3 years
were associated positively with wheeze at 5 and 6 years; urinary
BPA concentrations were associated with wheeze, both at 7 years,
and increased FENO values; and urinary BPA concentrations at 3, 5
and 7 years were associated with asthma. The authors stated that
this is the rst report of an association between postnatal urinary
BPA concentrations and asthma in children [65]; an association
perhaps, but no apparent causal relationship. (Did the authors look
for other adverse agent[s] in urine spots? Were there any blood/
serum analyses? Were there BPA-unexposed controls?) Of the 11
authors, seven disclosed potential conict of interest.
Environmental Health Perspectives Journal Online recently re-
ported that FDA has decided to ban bisphenol A (BPA)-based epoxy
resins as coatings in infant formula packaging, saying that . use
has largely been abandoned by manufacturers anyway (EHP Noon
News: Med page today, 18th July 2013).
In welcome newsteps, basic researchers are using more animals
per group and more test groups [66], and regulatory researchers
are adding more cutting-edge endpoints to their large, robust
studies.
5. Conclusions
ED presence/effects are major societal concerns. Sound repro-
ducible science by academic, governmental, regulatory, industrial
and other researchers must address concerns including environ-
mental EDC levels, hazard/risk, better testing protocols, mecha-
nisms of EDC action, and epigenetic role(s) in normal/abnormal
development.
Sharpe [32] argues that scientic facts usually evolve mean-
ing/importance over time, which is important for scientic prog-
ress, and that BPA research has been trying to go through this
process but has become literally bogged down in the mire of con-
troversy, much of which stems from.earliest ndings and seems
to have little to do with . current state of the science. Sharpe
contrasts large, robust BPA studies in rats/mice using oral exposure
with no low-dose effects, some with positive controls, vs small
basic research studies with low-dose BPA effects, noting the
problem of low-dose effects and inverted U-shaped NMDR curves
. in these studies.
Unfounded/wild accusations/criticisms by academic researchers
toward regulatory/industrial studies/authors must stop. We must
work collegially to resolve scientic concerns, potential, real (and
unreal) ED risks in general, and for specic EDCs (e.g. BPA),
especially to humans, other mammals, other vertebrates, in-
vertebrates (and plants?) in our earthly biome.
R.W. Tyl / Seminars in Fetal & Neonatal Medicine 19 (2014) 195e202 200
Conict of interest statement
There is no conict of interest.
Funding sources
The BPA and E
2
studies performed at RTI were guideline- and
GLP-compliant works for hire (RTI is a Contract Research Organi-
zation) funded by the American Plastics Council of the American
Chemistry Council, which played no role in this producing manu-
script. The time to write this manuscript was provided by the RTI
Fellows Program (Dr Tyl is a Distinguished Fellow in DART).
Acknowledgments
The author wishes to thank RTI Internationals Fellows Program
for the time to write this review, the senior and technical staff in the
Pharmacology and Toxicology and Animal Research Facility groups
for their extraordinary dedication and expertise on the BPA and E
2
studies discussed, and Ms Kathy Ancheta and Ms Rebecca Hipp for
their valued assistance in the production of this review, including
their patient typing and retyping of this manuscript.
References
[1] ScienceDaily. 99% of pregnant women in US test positive for multiple chem-
icals including banned ones, study suggests. University of California e San
Francisco; 2011. http://www.sciencedaily.com/releases/2011/01/
110114081653.htm.
[2] Calafat AM, Ye X, Wong LY, Reidy JA, Needham LL. Exposure of the U.S.
population to bisphenol A and 4-tertiary-octylphenol: 2003e2004. Environ
Health Perspect 2008;116:39e44.
[3] Diamanti-Kandarakis E, Bourguignon JP, Giudice LC, Hauser R, Prins GS,
Soto AM, et al. Endocrine-disrupting chemicals: an Endocrine Society scien-
tic statement. Endocr Rev 2009;30:293e342.
[4] vom Saal FS, Cooke PS, Buchanan DL, Palanza P, Thayer KA, Nagel SC, et al.
A physiologically based approach to the study of bisphenol A and other es-
trogenic chemicals on the size of reproductive organs, daily sperm production,
and behavior. Toxicol Ind Health 1998;14:239e60.
[5] Howdeshell KL, Hotchkiss AK, Thayer KA, Vandenbergh JG, vom Saal FS.
Exposure to bisphenol A advances puberty. Nature 1999;401(6755):763e4.
[6] Ryan BC, Hotchkiss AK, Crofton KM, Gray Jr LE. In utero and lactational
exposure to bisphenol A, in contrast to ethinyl estradiol, does not alter
sexually dimorphic behavior, puberty, fertility, and anatomy of female LE rats.
Toxicol Sci 2010;114:133e48.
[7] vom Saal FS, Akiingbemi BT, Belcher SM, Crain DA, Crews D, Guidice LC, et al.
Flawed experimental design reveals the need for guidelines requiring
appropriate positive controls in endocrine disruption research. Toxicol Sci
2010;115:612e3. author reply 614e20.
[8] Gray Jr LE, Ryan BC, Hotchkiss AK, Crofton KM. Letter to the editor: rebuttal of
Flawed experimental design reveals the need for guidelines requiring
appropriate positive controls in endocrine disruption research by vom Saal.
Toxicol Sci 2010;115:614e20.
[9] Myers JP, vom Saal FS, Akingbemi BT, Arizono K, Belcher S, Colborn T, et al.
Why public health agencies cannot depend on good laboratory practices as a
criterion for selecting data: the case of bisphenol A. Environ Health Perspect
2009;117:309e15.
[10] Tyl RW. Basic exploratory research versus guideline-compliant studies used
for hazard evaluation and risk assessment: bisphenol A as a case study. En-
viron Health Perspect 2009;117:1644e51.
[11] Tyl RW, Myers CB, Marr MC, Thomas BF, Keimowiitz AR, Brine DR, et al. Three-
generation reproductive toxicity study of dietary bisphenol A in CD Spraguee
Dawley rats. Toxicol Sci 2002;68:121e46.
[12] Tyl RW, Myers CB, Marr MC, Castillo NP, Veselica MM, Joiner RL, et al. One-
generation reproductive toxicity study of dietary 17beta-estradiol (E2; CAS
No. 50-28-2) in CD-1 (Swiss) mice. Reprod Toxicol 2008;25:144e60.
[13] Tyl RW, Myers CB, Marr MC, Sloan CS, Castillo NP, Veselica MM, et al. Two-
generation reproductive toxicity evaluation of dietary 17beta-estradiol (E2;
CAS No. 50-28-2) in CD-1 (Swiss) mice. Toxicol Sci 2008;102:392e412.
[14] Tyl RW, Myers CB, Marr MC, Sloan CS, Castillo NP, Veselica MM, et al. Two-
generation reproductive toxicity study of dietary bisphenol A in CD-1 (Swiss)
mice. Toxicol Sci 2008;104:362e84.
[15] Stump DG, Beck MJ, Radovsky A, Garman RH, Freshwater LL, Sheets LP, et al.
Developmental neurotoxicity study of dietary bisphenol A in SpragueeDaw-
ley rats. Toxicol Sci 2010;115:167e82.
[16] Draganov Markham D, Beyer D, Moore J, Riehl T, Waechter J, et al. Investi-
gation of route-dependent exposure and metabolism of bisphenol A (BPA) in
neonatal mice following oral and subcutaneous administration of
3
H-BPA.
Society of Toxicology Annual Meeting, San Antonio, Texas, USA; 2013.
[17] Doerge DR, Twaddle NC, Vanlandingham M, Fisher JW. Pharmacokinetics of
bisphenol A in serum and adipose tissue following intravenous administration
to adult female CD-1 mice. Toxicol Lett 2012;211:114e9.
[18] Teeguarden JG, Calafat AM, Ye X, et al. Twenty-four hour human urine and
serum proles of bisphenol a during high-dietary exposure. Toxicol Sci
2011;123:48e57.
[19] Brent RL. Commentary: urine chemical content may be a false measure
of environmental exposure. Pediatrics 2013;132: e747e8. http://pediatrics.
aappublications.org/content/132/3/e747.full.html.
[20] Nagel SC, vom Saal FS, Thayer KA, et al. Relative binding afnity-serum
modied access (RBA-SMA) assay predicts the relative in vivo bioactivity of
the xenoestrogens bisphenol A and octylphenol. Environ Health Perspect
1997;105:70e6.
[21] vom Saal FS, Akingbemi BT, Belcher SM, Birnbaum LS, Crain DA, Eriksen M,
et al. Chapel Hill bisphenol A expert panel consensus statement: integration of
mechanisms, effects in animals and potential to impact human health at
current levels of exposure. Reprod Toxicol 2007;24:131e8.
[22] Center for the Evaluation of Risks to Human Reproduction. Draft meeting
summary. Expert panel evaluation of bisphenol A. CERHR (NIEHS); 2007.
[23] Vandenberg LN, Hauser R, Marcus M, Olea N, Welshons WV. Human exposure
to bisphenol A (BPA). Reprod Toxicol 2007;24:139e77.
[24] National Toxicology Program. NTP-CERHR Monograph on the potential hu-
man reproductive and developmental effects of bisphenol A. Research Tri-
angle Park, NC: NTP; 2008.
[25] Hunt PA, Koehler KE, Susiarjo M, Hodges CA, IIagan A, Voigt RC, et al.
Bisphenol a exposure causes meiotic aneuploidy in the female mouse. Curr
Biol 2003;13:546e53.
[26] Pacchierotti F, Ranaldi R, Eichenlaub-Ritter U, Attia S, Adler ID. Evaluation of
aneugenic effects of bisphenol A in somatic and germ cells of the mouse.
Mutat Res 2008;651:64e70.
[27] Hunt PA, Susiarjo M, Rubio C, Hassold TJ. The bisphenol A experience: a
primer for the analysis of environmental effects on mammalian reproduction.
Biol Reprod 2009;81:807e13.
[28] Teeguarden JG, Hanson-Drury S. A systematic review of bisphenol A low
dose studies in the context of human exposure: a case for establishing
standards for reporting low-dose effects of chemicals. Food Chem Toxicol
2013;62:935e48.
[29] Kortenkamp A, Faust M, Scholze M, Backhaus T. Low-level exposure to mul-
tiple chemicals: reason for human health concerns? Environ Health Perspect
2007;115(Suppl. 1):106e14.
[30] Calafat AM, Weuve J, Ye X, Jia LT, Hu H, Ringer S, et al. Exposure to bisphenol A
and other phenols in neonatal intensive care unit premature infants. Environ
Health Perspect 2009;117:639e44.
[31] Richter CA, Birnbaum LS, Farabollini F, Newbold RR, Rubin BS, Talsness CE,
et al. In vivo effects of bisphenol A in laboratory rodent studies. Reprod
Toxicol 2007;24:199e224.
[32] Sharpe RM. Is it time to end concerns over the estrogenic effects of bisphenol
A? Toxicol Sci 2010;114:1e4.
[33] Newbold RR, Jefferson WN, Padilla-Banks E. Long-term adverse effects of
neonatal exposure to bisphenol A on the murine female reproductive tract.
Reprod Toxicol 2007;24:253e8.
[34] Newbold RR, Jefferson WN, Padilla-Banks E. Prenatal exposure to bisphe-
nol A at environmentally relevant doses adversely affects the murine fe-
male reproductive tract later in life. Environ Health Perspect 2009;117:
879e85.
[35] Beronius A, Ruden C, Hakansson H, Hanberg A. Risk to all or none? A
comparative analysis of controversies in the health risk assessment of
bisphenol A. Reprod Toxicol 2010;29:132e46.
[36] Hengstler JG, Foth H, Gebel T, Kramer PJ, Lilienblum W, Schweinfurth H, et al.
Critical evaluation of key evidence on the human health hazards of exposure
to bisphenol A. Crit Rev Toxicol 2011;41:263e91.
[37] Viguie C, Collet SH, Gayrard V, Picard-Hagen N, Puel S, Roques BB, et al.
Maternal and fetal exposure to bisphenol a is associated with alterations of
thyroid function in pregnant ewes and their newborn lambs. Endocrinology
2013;154:521e8.
[38] Yaoi T, Itoh K, Nakamura K, et al. Genome-wide analysis of epigenomic al-
terations in fetal mouse forebrain after exposure to low doses of bisphenol A.
Biochem Biophys Res Commun 2008;376:563e7.
[39] Kouzarides T. Chromatin modications and their function. Cell 2007;128:
693e705.
[40] Skinner MK, Manikkam M, Guerrero-Bosagna C. Epigenetic transgenerational
actions of endocrine disruptors. Reprod Toxicol 2011;31:337e43.
[41] Dolinoy DC, Huang D, Jirtle RL. Maternal nutrient supplementation counter-
acts bisphenol A-induced DNA hypomethylation in early development. Proc
Natl Acad Sci USA 2007;104:13056e61.
[42] Salian S, Doshi T, Vanage G. Impairment in protein expression prole of
testicular steroid receptor coregulators in male rat offspring perinatally
exposed to bisphenol A. Life Sci 2009;85:11e8.
[43] Wolstenholme JT, Edwards M, Shetty SR, Gatewood JD, Taylor JA, Rissman EF,
et al. Gestational exposure to bisphenol A produces transgenerational changes
in behaviors and gene expression. Endocrinology 2012;153:3828e38.
[44] Environmental Health Research Foundation. FDA Update on Bisphenol A
(BPA) EHRF; 2011. pp. 1e3 (05/23/2011).
R.W. Tyl / Seminars in Fetal & Neonatal Medicine 19 (2014) 195e202 201
[45] Wolstenholme JT, Rissman EF, Connelly JJ. The role of bisphenol A in shaping
the brain, epigenome and behavior. Horm Behav 2011;59:296e305.
[46] Nakamura K, Itoh K, Yaoi T, Fujiwara Y, Sugimoto T, Fushiki S. Murine
neocortical histogenesis is perturbed by prenatal exposure to low doses of
bisphenol A. J Neurosci Res 2006;84:1197e205.
[47] Patisaul HB, Fortino AE, Polston EK. Neonatal genistein or bisphenol-A
exposure alters sexual differentiation of the AVPV. Neurotoxicol Teratol
2006;28:111e8.
[48] Kundakovic M, Champagne FA. Epigenetic perspective on the developmental
effects of bisphenol A. Brain Behav Immun 2011;25:1084e93.
[49] Ho SM, Tang WY, Belmonte de Frausto J, Prins GS. Developmental exposure to
estradiol and bisphenol A increases susceptibility to prostate carcinogenesis
and epigenetically regulates phosphodiesterase type 4 variant 4. Cancer Res
2006;66:5624e32.
[50] Cheng RYS, Tang W-y. Environment, epigenetics, and diseases, in toxicology
and epigenetics. New York: Wiley; 2012. pp. 5e24.
[51] Yeo M, Berglund K, Hanna M, Guo JU, Kittur J, Torres MD, et al. Bisphenol A
delays the perinatal chloride shift in cortical neurons by epigenetic effects on
the Kcc2 promoter. Proc Natl Acad Sci USA 2013;110:4315e20.
[52] Amir RE, Van den Veyver IB, Wan M, Tran CQ, Francke U, Zoghbi HY. Rett
syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-
binding protein 2. Nat Genet 1999;23:185e8.
[53] ROUND-UP. ROUND-UP: Consumer chemical affects development of isolated
neurons (PNAS*) e experts respond. Australian Science Media Centre; 2013.
http://www.smc.org.au/2013/02/round-up-consumer-chemical-affects-
development-of-isolated-neurons-pnas-experts-respond/.
[54] European Food Safety Authority. Toxicokinetics of bisphenol A: scientic
opinion of the panel on food additives, avourings, processing aids and ma-
terials in contact with food (AFC). EFSA J 2008;759:1e10. http://www.efsa.
europa.eu/EFSA/efsa_locale-1178620753812_1211902017492.htm.
[55] Ginsberg G, Rice DC. Does rapid metabolism ensure negligible risk from
bisphenol A? Environ Health Perspect 2009;117:1639e43.
[56] Gioiosa L, Fissore E, Ghirardelli G, Parmigiani S, Palanza P. Developmental
exposure to low-dose estrogenic endocrine disruptors alters sex differences in
exploration and emotional responses in mice. Horm Behav 2007;52:307e16.
[57] Yu C, Tai F, Song Z, Wu R, Zhang X, He F. Pubertal exposure to bisphenol A
disrupts behavior in adult C57BL/6J mice. Environ Toxicol Pharmacol 2011;31:
88e99.
[58] Tando S, Itoh K, Yaoi T, Ikeda J, Fujiwara Y, Fushiki S. Effects of pre- and
neonatal exposure to bisphenol A on murine brain development. Brain Dev
2007;29:352e6.
[59] Nakamura K, Itoh K, Dai H, Han L, Wang X, Kato S, et al. Prenatal and lacta-
tional exposure to low-doses of bisphenol A alters adult mice behavior. Brain
Dev 2012;34:57e63.
[60] Braun JM, Kalkbrenner AE, Calafat AM, Yolton K, Ye X, Dietrich KN, et al.
Impact of early-life bisphenol A exposure on behavior and executive function
in children. Pediatrics 2011;128:873e82.
[61] Selgrade MK, Blain RB, Fedak KM, Cawley MA. Potential risk of asthma asso-
ciated with in utero exposure to xenobiotics. Birth Defects Res C Embryo
Today 2013;99:1e13.
[62] Spanier AJ, Kahn RS, Kunselman AR, Hornung R, Xu Y, Calafat AM, et al. Pre-
natal exposure to bisphenol A and child wheeze from birth to 3 years of age.
Environ Health Perspect 2012;120:916e20.
[63] Chevrier J, Gunier RB, Bradman A, Holland NT, Calafat AM, Eskenazi B, et al.
Maternal urinary bisphenol A during pregnancy and maternal and neonatal
thyroid function in the CHAMACOS study. Environ Health Perspect 2013;121:
138e44.
[64] Harley KG, Aguilar Schall R, Chevrier J, Tyler K, Aguirre H, Bradman A, et al.
Prenatal and postnatal bisphenol A exposure and body mass index in child-
hood in the CHAMACOS cohort. Environ Health Perspect 2013;121:514e20.
520e1e6.
[65] Donohue KM, Miller RL, Perzanowski MS, Just AC, Hoepner LA, Arunajadai S,
et al. Prenatal and postnatal bisphenol A exposure and asthma development
among inner-city children. J Allergy Clin Immunol 2013;131:736e42.
[66] Angle BM, Do RP, Ponzi D, Stahlhut RW, Drury BE, Nagel SC, et al. Metabolic
disruption in male mice due to fetal exposure to low but not high doses of
bisphenol A (BPA): evidence for effects on body weight, food intake, adipo-
cytes, leptin, adiponectin, insulin and glucose regulation. Reprod Toxicol
2013;42:256e68.
R.W. Tyl / Seminars in Fetal & Neonatal Medicine 19 (2014) 195e202 202
Review
Carcinogenic risks of prenatal ionizing radiation
Robert L. Brent
*
Thomas Jefferson University, Alfred I. duPont Hospital for Children, Room 308, A/R Building, PO Box 269, Wilmington, DE 19899, USA
Keywords:
Cancer
Carcinogenic risks
Fetus
Ionizing radiation
s u m m a r y
The risk of cancer in offspring who have been exposed to diagnostic X-ray procedures while in utero has
been debated for 55 years. High doses at high dose rates to the embryo or fetus (e.g. >0.5 Gy) increase
the risk of cancer. This has been demonstrated in human epidemiology studies as well as in mammalian
animal studies. Most pregnant women exposed to diagnostic X-ray procedures or the diagnostic use of
radionuclides receive doses to the embryo or fetus <0.1 Gy. The risk of cancer in offspring exposed in
utero at a low dose such as <0.1 Gy is controversial and has not been determined.
2013 Elsevier Ltd. All rights reserved.
1. Introduction
In 1950e51, I was working in the radiation embryology section
of the University of Rochester Medical Centers Atomic Energy
Project. We had submitted an abstract to the Anatomy Society
meetings in Detroit [1]. The completed manuscript that was sub-
mitted to Cancer Research was titled, Cancer induced in rat em-
bryos by roentgen irradiation. The editors rejected the manuscript
and stated that if the embryos had developed cancer, there would
have been a much higher mortality. So we changed the title to
Neoplasia induced in rat embryos by roentgen irradiation and the
manuscript was accepted [2]. We examined the tumors as they rst
appeared and continued to grow (Fig. 1). Many of the tumors
became anaplastic and contained many undifferentiated cells with
a high mitotic index. At birth, most of the tumors were gone.
However, there were a few pyknotic cell remnants that were still
present. We followed 300 radiated survivors and controls for 4
years and these irradiated animals did not have a higher incidence
of cancer than the controls.
We put this project aside with the tentative conclusion that the
embryo was less vulnerable to the carcinogenic effects of low ex-
posures of ionizing radiation than the postnatal animal.
Liane Russell [3,4] and our laboratory [5,6] had already
described the all or none phenomenon, which indicated that the
pre-somite mammalian embryo was less vulnerable to the terato-
genic effects of ionizing radiation. The embryo was very vulnerable
to the lethal effects of radiation; however, the surviving embryos
did not have an increased risk of birth defects.
When Alice Stewart published her research results, a 60-year
controversial discussion was initiated. Stewart et al. [7e10] sug-
gested that the human embryo was more vulnerable to the leuke-
mogenic effects of radiation and in later publications concluded
that other childhood cancers also occur more frequently in persons
exposed in utero to diagnostic radiologic procedures (primarily
pelvimetry) (Fig. 2). These authors initially estimated that a 1e2 rad
in-utero radiation exposure increases the risk of leukemia devel-
oping in the offspring by a factor of 1.5 to 2.0 over the natural
incidence. This incidence is considerably greater than the increase
resulting from 2 rad delivered to an adult population. In fact, an
increase in the incidence of leukemia after an adult population
exposure of 2 rad would be difcult to document, even for very
large population groups [11,12]. Dr Stewart became a spokesperson
for anti-radiation groups. She appeared as a plaintiff expert in ra-
diation litigation and was even a plaintiff expert against her own
country in a case before the World Court inwhich Ireland was suing
the UK, claiming that a British nuclear facility (Sellaelds Fuel
Handling Plant) was contaminating the Irish sea and causing
increased cases of birth defects and cancer in the inhabitants on the
east coast of Ireland. After more than a decade of litigation the
World Court decided in favor of the UK [13]. Dr Stewart claimed
that the embryo was many times more vulnerable to the carcino-
genic effects of radiation than children and she was critical of sci-
entists who disagreed with her [8,14].
As a medical and graduate student and part-time instructor, I
did not have time to further pursue the question of the resistance of
the embryo to the carcinogenic effects of radiation. However, there
were many publications exposing animals to carcinogenic agents.
In particular, urethan (urethane; ethyl carbamate) was used by
Klein [15] and Vesselinovitch et al. [16] to produce neoplasia in
rodents. Only a few of the investigators utilizing urethan exposed
pregnant animals to this carcinogenic agent. Klein [15] reported
* Tel.: 1 302 651 6880; fax: 1 302 651 6888.
E-mail address: rbrent@nemours.org.
Contents lists available at ScienceDirect
Seminars in Fetal & Neonatal Medicine
j ournal homepage: www. el sevi er. com/ l ocat e/ si ny
1744-165X/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.siny.2013.11.009
Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213
that cesarean-delivered mice exposed in utero had signicantly
fewer lung tumors than animals treated postnatally. Signicantly
more tumors per lung were observed in mice injected with urethan
at 47 days of age than at birth, suggesting an increased suscepti-
bility with age. Vesselinovitch et al. [16] exposed pregnant mice on
multiple days in mid pregnancy (days 12e18). The incidence of liver
and lung tumors was signicantly higher in mice exposed to this
carcinogen at the end of gestation. Neonatally treated animals
developed all of the tumor types more readily than those exposed
to the carcinogen in utero and also developed leukemia which did
not occur in the in-utero-exposed population. The urethan animal
studies reinforced the animal studies from our laboratory, which
indicated that the fetus had lower carcinogenic risks from muta-
genic or carcinogenic agents when compared to the postnatal an-
imals vulnerability.
2. Human studies concerning the vulnerability of the embryo
to the carcinogenic risks of ionizing radiation
Lilienfeld [17] reviewed the epidemiologic considerations with
respect to leukemogenesis. His results, conrmed by others [18e
21] support the thesis that diagnostic radiation absorbed in utero
was associated with an increased risk of leukemia. Six of nine
studies reported in Lilienfelds paper indicate a 1.3e1.8-fold in-
crease in the risk of leukemia after diagnostic radiation exposure in
utero. Lilienfeld states: When one considers the variety of control
groups used and the sampling variability, the results are remark-
ably consistent in showing an excess frequency of leukemia among
children of radiation-exposed pregnant mothers [17]. Diamond
et al. [22] conrmed and extended the observation of a three-fold
increased incidence of leukemia in children exposed to diagnostic
radiation in utero. Interestingly, this effect did not occur in the
African-American population. When MacMahon [23] extended his
studies, the 1.5-fold excess leukemia incidence remained, but the
excess in other childhood cancers was no longer present (Table 1).
Fig. 1. (A, B) Fifteen-day-old rat fetuses exposed to 1.5 Gy at day 9 post conception. The arrows in 1b are pointing to tumors under the scalp that can be seen in histological sections
in 3a, 3b, and 3c. The tumor growths are derived from outgrowths of the radiated neural tube as seen in 2a, 2b and 2c. Some of the growths dedifferentiated into more aggressive-
appearing tissues (4). At the time of the birth of the fetuses almost all the growths had regressed except for a few remnants of pyknotic cells located between the brain and the scalp
(5a, 5b, and 5c). All photographs are reproduced with permission from Wilson et al. [2].
100
10
1
Fetus 1 10 20 30
Stewart et al.
810
ABCC
28,38,55
Animal data
1,2,61,6466,108
Age (years)
R
i
s
k
o
f
l
e
u
k
e
m
i
a
p
e
r
1
0
6
p
e
r
s
o
n
s
f
o
l
l
o
w
i
n
g
0
.
0
1
t
o
0
.
0
2
G
y
40 50 60 70
Fig. 2. Risk of cancer from in-utero radiation. When Stewart rst reported the risks of
cancer in the offspring of pregnancies in which the mother had been exposed to
diagnostic radiological studies in the 1950s, the risk of leukemia was stated as one to
two orders of magnitude greater than the risk of cancer following similar exposures in
childhood. Children were believed to be slightly more vulnerable than adults. Animal
studies were inconsistent, but many of the animal studies were negative and many of
the studies did not expose the pregnant animals to doses of <0.10 Gy. ABCC, Atomic
Bomb Casualty Commission.
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 204
Table 1
Risk of specic and total childhood cancers in offspring of women undergoing prenatal diagnostic X-ray procedures: caseecontrol studies.
a
Reference, country, birth year No. of cases/no.
of controls
Type of
control
Source of X-ray
procedure
information
Controls with
abdominal X-ray
procedures (%)
Any X-ray
procedure
Estimated
RR (95% CI)
b
Abdominal
procedure
Pelvimetry
Acute lymphoblastic leukemia
Bithell and Stewart [24],
c
UK, 1943e1967
2007/8513 Population Interview,
medical records,
questionnaire
11.5 1.5 (1.3e1.8)
Van Steensel-Moll et al. [25],
The Netherlands, 1959e1980
517/509 Population Questionnaire 3.7 2.2 (1.2e3.8)
Shu et al. [26], Shanghai, China,
1960e1986
172/618 Population Interview 7.1 1.6 (0.9e2.8) 2.0 (0.7e3.8)
Magnani et al. [27], Turin, Italy
(years not provided)
142/307 Hospital Interview 5.5 11 (not provided)
Naumberg et al. [28], Sweden,
1973e1989
449/450 Population Medical records 9.8 1.0 (0.6e1.7) 1.1 (0.8e1.6)
Shu et al. [29], USA and
Canada, 1972e1992
1842/1986 Population Interview 6 (before 1980)
2.3 (1981e1986)
1.8 (after 1986)
1.2 (0.8e1.7)
(all ages)
2.4 (1.2e5.0)
(ages 11e14
years)
Acute myeloid leukemia
Bithell and Stewart [24],
c
UK, 1943e1967
866/8513 Population Interview;
medical records;
questionnaire
11.5 1.5 (1.2e1.8)
Shu et al. [26], Shanghai, China,
1960e1986
92/618 Population Interview 7.1 1.4 (0.6e3.0) 0.6 (0.1e5.0)
Van Duijn et al. [30],
The Netherlands, 1969e1979
80/240 Population Questionnaire 3 2.4 (0.8e7.0)
All leukemias
Kaplan [31], California, USA,
1943e1967
150/150 Friends Interview 16 1.7 (1.1e2.7)
Polhemus and Koch [18],
Los Angeles, California, USA
(years not provided)
251/251 Hospital Questionnaire 23.1 1.2 (0.8e1.8)
Graham et al. [20], Baltimore,
Minneapolis, New York State,
USA (Tristate Study, 1969e1979)
313/854 Population Medical records 23.4 1.4 (0.9e2.3)
Salonen and Saxen [32], Finland,
1945e1968
373/373 Population Medical records 49.3 1.0 (0.5e1.9)
Hirayama [33], Japan, 1969e1977 4607/5968 Other cancers Not provided 10.6 1.6 (1.4e1.8)
Monson and MacMahon (1984)
[34],
d
Northeast USA, 1947e1960
704/14 276 Hospital Medical records 9.4 1.5 (1.2e2.0)
e
Shu et al. [35], Shanghai, China,
1986e1991
166/166 Population Interview 7.1 2.4 (0.5e10.6)
Infante-Rivard et al. [36],
InfanteeRivard and Deadman [37],
Canada, 1980e1993
701/701 Population Interview 0.8 (0.6e1.3)
Rajaraman et al. [38], UK,
1976e1996
1253/4857 Population Medical records 1.2 1.4 (0.9e2.0)
CNS tumors
Bithell and Stewart [24],
c
UK,
1943e1967
1332/8513 Population Interview;
medical records;
questionnaire
11.5 1.4 (1.2e1.7)
Salonen and Saxen [32], Finland,
1945e1968
245/245 Population Medical records 49.3 1.1 (0.3e4.2)
Preston-Martin et al. [39],
Los Angeles, California, USA,
1948e1977
209/209 Friends,
neighborhood
Interview 15.0 1.3 (not provided)
Monson and MacMachon [34],
d
Northeast USA, 1947e1960
298/14 276 Hospital Medical records 9.4 1.2 (0.8e1.7)
f
Bunin et al. [40], Greater Delaware
Valley, USA, 1980e1986
155 astro/32
166 PNET/321
Population Interview 1.1 (0.3e3.9)
0.8 (0.3e2.3)
Schuz et al. [41], Germany,
1988e1993
466/2458 Population Interview 0.8 (0.4e1.4)
Stalberg et al. [42], Sweden,
1975e1984
512/524 (total CNS)
105 PNET/524
191 astro/524
Population Medical records;
registry
9.1 1.1 (0.7e1.6)
1.7 (0.9e3.2)
0.8 (0.4e1.5)
Rajaraman et al. [38], UK,
1976e1996
482/4857 Population Medical records 1.2 1.1 (0.6e1.8)
Neuroblastoma
Bithell and Stewart [24],
c
UK,
1943e1967
720/8513 Population Interview;
medical records;
questionnaire
11.5 1.5 (1.2e1.8)
(continued on next page)
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 205
Several associations of these data should be pointed out. In the
studies of Stewart and colleagues [7,10], there was a higher inci-
dence of previous abortion in the mothers receiving pelvimetry,
and the children in the pelvimetry group had a higher incidence of
upper respiratory infections before the development of leukemia.
Others reported that infants with a strong family history of allergy
are also more susceptible to radiation-induced leukemia when
exposed to diagnostic radiation in utero. The problem with these
data is that patients with an allergic history and no exposure to
radiation had a higher frequency of leukemia than did other groups
that had received radiation in utero [48].
Of the 86 persons exposed in utero at Nagasaki, none developed
leukemia [8]. These persons received considerably higher doses of
radiation than did those patients in the previous studies.
Shiono et al. [49] examined the potential risk of diagnostic X-
rays in the 44 908 pregnant patients studied in the Collaborative
Perinatal Project of the National Institute of Neurological and
Communicative Disorders and Stroke. These ndings were sur-
prising in that they reported a statistically increased relative risk for
malignancies in the offspring of mothers who were exposed before
pregnancy (preconception) [relative risk (RR): 2.61; 90% condence
limits (CL): 1.26e5.85]. However, there was no statistically signi-
cant increase in the risk of malignant or benign tumors in the
offspring of mothers who were exposed to radiation during preg-
nancy (RR: 1.09; 90% CL: 0.47e2.40 for malignant neoplasms; and
RR: 0.94; 90% CL: 0.46e1.82 for benign neoplasms). Court-Brown
et al. [50] evaluated the incidence of leukemia in 39 166 offspring
of mothers who had been irradiated in utero, and Salonen [51]
studied the relationship between pregnancy radiation exposure
and childhood cancers. Neither study could establish a statistically
signicant increase in leukemia (Table 2).
Although it is true that the population of offspring exposed in
utero in Hiroshima and Nagasaki did not have an increased
incidence of leukemia and other childhood cancers during the
childhood years, they have of course developed cancer as adults
(Fig. 3).
A larger number of cancers must accumulate before we can
reliably establish a risk of cancer in adults who were exposed in
utero. Whether there is an increased risk has been partially
answered by Preston et al. [67] One conclusion is certain: the risk of
developing cancer as an adult from in-utero radiation is below the
risks of childhood cancer that have been suggested by several in-
vestigators (Fig. 4) [7e10,17e19,21].
Hoshino et al. [68] reported no increase in leukemia in a study of
17 000 children of parents who had received radiation before
conception. The question arises as to what extent the same biases
that contribute to the increased risk of leukemia in the cases of
radiation exposure before conception also affect the in-utero ra-
diation cases. Graham et al. [20] pointed out that children of
mothers with a history of abortion or stillbirths also had children
with a higher risk of leukemia.
Miller [69] and others [56,70,71] do not believe that the risk of
prenatal radiation is as great as Stewart suggests. Miller writes:
It is surprising that in Stewarts studies minimal doses of x-rays
are equally oncogenic whether exposure occurred before
conception or during pregnancy, whether the neoplasm studied
was leukemia or any other major cancer of childhood, and
whether the study was based on interviews, which may be
biased, or from hospital records. Taken in aggregate, the simi-
larity of results, in the absence of a doseeresponse effect or of
supporting data from animal experimentation, raises a question
about biologic plausibility of a causal relationship.
Furthermore, Miller [69] points out that siblings of leukemic
children have a risk of childhood leukemia of 1 in 720 in the rst 10
Table 1 (continued)
Reference, country, birth year No. of cases/no.
of controls
Type of
control
Source of X-ray
procedure
information
Controls with
abdominal X-ray
procedures (%)
Any X-ray
procedure
Estimated
RR (95% CI)
b
Abdominal
procedure
Pelvimetry
Bone tumors
Bithell and Stewart [24],
c
UK,
1943e1967
244/8513 Population Interview;
medical records;
questionnaire
11.5 1.1 (0.7e1.7)
Ewings sarcoma
Winn et al. (1992) [43], USA,
multicenter (years not provided)
204/204
191/191
Population
Siblings
Interview
Interview
27.5
17.3
0.8 (0.5e1.2)
1.5 (0.8e3.2)
Rhabdomyosarcoma
Gufferman et al. [44], USA,
multicenter, 1962e1988
319/319 Population Interview 6.8 1.4 (0.7e2.9)
Total childhood cancer
Bithell and Stewart [24],
c
UK,
1943e1967
8513/8513 Population Interview;
medical records;
questionnaire
11.5 1.5 (1.3e1.6)
MacMahon [45],
d
Northeast USA,
1947e1954
556/7230 Hospital Medical records 10.6 1.5 (1.2e1.8)
e
Monson and MacMahon [34],
d
Northeast USA, 1947e1960
1342/14 276 Hospital Medical records 9.4 1.3 (0.95e1.7)
e
Rajaraman et al. [38], UK,
1976e1996
2690/4857 Population Medical records 1.2 1.1 (0.9e1.4)
RR, relative risk; CI, condence interval; CNS, central nervous system; PNET, primitive neuroectodermal tumour.
a
Adapted from Little. [46].
b
RR and CI values are presented to one decimal place (rounded), except that one CI value is presented as 0.95 (rather than 1.0).
c
Initial Oxford Survey of Childhood Cancer CaseeControl Study (Stewart et al.
10
) was extended to include 12 additional birth years and an increase to 8513 cancer cases
from 1299 in the original study (Bithell and Stewart [24]). A third publication from the same investigation included cases born during 1948 to 1978 who died during 1953 to
1979 (Knox et al. [47]).
d
Initial study of MacMahon [45] was extended by Monson and MacMahon [34] to include ve additional hospitals, six additional birth years, seven additional years of
childhood cancer deaths, an increase to 1 429 400 children from 734 243 in the original study.
e
Results for children who died from cancer before their 10th birthday.
f
Results for children who died from cancer before their 20th birthday; condence interval computed from data in Monson and MacMahon. [34].
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 206
years of their life, which is greater than the 1:2000 risk of leukemia
after pelvimetry exposure and the 1:3000 probability of leukemia
in the general population of children followed for 10 years (Table 3).
Stewart and Kneale [7] reinforces the contention that radiation may
not be the only etiologic factor responsible for the induction of
malignancy because of unirradiated siblings of the irradiated pa-
tient population with a higher incidence of leukemia also had an
incidence higher than in control siblings and in control patients.
This observation certainly would indicate that genetic or other
environmental factors may be important in the etiology of
leukemia.
At present, some investigators believe that in-utero exposure to
small amounts of radiation increased the risk of leukemia and other
malignancies, whereas others seriously question the contention
that the embryo is markedly more sensitive to the leukemogenic
effects of irradiation when compared with the child or adult. Until
the mechanism is understood, there will be doubt concerning the
magnitude of the role of in-utero diagnostic radiology studies in
leukemia induction. The increased incidence of cancer in children
exposed to in-utero diagnostic radiation should be claried in view
of the fact that much higher doses of radiation to animal embryos
and to the children exposed in utero at Hiroshima and Nagasaki
have not resulted in a marked increase in the incidence of cancers
from higher doses of radiation, which one would expect if the
embryo were as sensitive to the carcinogenic effects of radiation as
Stewart and colleagues suggest (Table 2) (Figs. 3 and 4) [74e77].
One cannot overemphasize either the importance of the multi-
plicity of factors or the difculties involved in identifying and con-
trolling for such factors. Even laboratory experiments concerned
with tumor production are difcult to interpret. For example, Ross
and Bras [48] reported that the incidence of spontaneous tumors
varied with the diet and weight of the animals. Heavier animals on
high-protein diets had a higher incidence of tumors than did the
lighter rats onlow-proteindiets. Hence, there are many unanswered
questions pertaining to the relationship between leukemia and
malignancy and in-utero radiation exposure.
Because of the introduction of new diagnostic techniques, such
as the use of ultrasound, and because of the concerns about the
risks of radiation, fewer pregnant patients will be exposed in the
future. Therefore it is unlikely that adequate numbers of exposed
patients will be available to evaluate the carcinogenic risks of in-
utero diagnostic radiation. MacMahon [21] in his editorial in the
Table 2
Cancer mortality rates in cohorts of children whose mothers underwent diagnostic
X-ray procedures during pregnancy.
a
Reference, location No. of cancer deaths;
no. of children
exposed in utero
b
Total cancer
[RR (95% CI)]
Leukemia
[RR (95% CI)]
Murray et al. [52]
(Rochester, NY)
3 (L); w6740
c
0.9 (0.3e3.1)
Court Brown et al. [50]
(Edinburgh)
d
9 (L); 39 166 0.9 (0.4e1.6)
Lewis [53] (London) 1 (L); 11 443 0.4 (0.1e2.6)
Griem et al. [54]
(Chicago)
4 (I L, 3 O); 982 1.2 (0.4e4.0) 0.4 (0.1e2.6)
Oppenheim et al.
[55,56] (Chicago)
e
1 (L); 393 0.7 (0.1e5.0)
Diamond et al. [22]
(Baltimore)
13 (6 L, 7 O); 19 889 1.1 (0.5e2.1) 1.6 (0.6e4.6)
Shiono et al. [49]
(USA, multicenter)
7; w5000
f
1.1 (0.5e2.4)
Golding et al. [57]
(UK, national)
12; w3000
g
1.2 (0.6e2.5)
Combined small
cohorts (ICRP [58])
7 4.6 (0.9e25.1)
Dempster [59] [0; 148]
Milis et al. [60] [0; 190]
Lejeune et al. [61] [2; 491]
Magnin [62] [1; 5353]
Nokkentved [63] [0; 152]
Hagstrom et al. [64],
h
[4; 649]
Ray et al. [65]
(Ontario, Canada)
4; 5590 0.7 (0.3e1.8)
a
Adapted from Doll and Wakeford [66] and ICRP. [58].
b
The number of leukemias (L) and other cancers (O) are given when available.
c
A total of 140 438 children aged <20 years were included; w6460 of the 6740
exposed mothers had pelvimetry or other abdominal X-ray procedures during
pregnancy.
d
Doll pointed out his concerns about the adequacy of the identication of irra-
diated women that arose when he tried to extend the Court Brown et al. [50] study.
Doll indicated that some of the ndings may therefore be unreliable (Doll and
Wakeford [66]).
e
Subsequent follow-up study of Griem et al. [54].
f
A total of 55 908 live births were studied; w10% of the mothers had abdominal
X-ray procedures during pregnancy.
g
A total of 16 193 live births were studied; w18% of the mothers had abdominal,
non-abdominal or dental X-ray procedures during pregnancy.
h
Exposure in Hagstrom et al. [64] was to radioactive iron (
59
Fe) administered
during pregnancy. The RR and 95% CI for this small cohort was 6.1 (1.7e15.8).
Fig. 3. Solid cancer risk patterns for in-utero and childhood exposures e atom bomb survivors. The graphs indicate that there is a clear and signicant difference in the risk of tumor
development in the embryo-irradiated and childhood-irradiated populations. During the course of the 55 year post-radiation period tumors are accumulating at a much lower rate
in the irradiated in-utero population. ERR, excess relative risk; PY, person-years; CI, condence interval. Reproduced with permission from Preston et al. [67].
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 207
New England Journal of Medicine wrote that It seems likely that the
question of the association between fetal irradiation and childhood
cancer will fade into medical history unresolved and remain a
source of more confusion than enlightenment.
There is little disagreement with the concept that radiation may
represent a carcinogenic risk to the embryo and adult and that
there may be different risks per rad at different stages of devel-
opment. The concept that is difcult to explain from a basic science
viewpoint is, Why would embryonic cells be orders of magnitude
more sensitive to radiation-induced cancer than cells of children or
adults?
Recent publications have provided added perspectives, data,
and interpretations of the prior publications [78e86]. Boice and
Miller [78] point out that numerous epidemiologic studies have
been performed. Positive associations for an increased incidence of
cancer after in-utero diagnostic radiation exposures have been
derived almost exclusively from caseecontrol studies (Tables 1e3),
whereas almost all of the cohort studies have found no association
(Figs. 3 and 4). It is of great interest that the in-utero atomic bomb
population did not demonstrate an increase in childhood leukemia
despite the fact that many in the in-utero population were exposed
to high doses of acute irradiation.
Because many of the positive associations have been derived
from caseecontrol studies, the questions of confounding factors
have been raised to explain the ndings. Twin studies have been
used to eliminate some of the confounding factors [87e89].
Although the reports of Mole [89] and Harvey et al. [87] were
positive, the Rodvall study [88] was not statistically signicant.
The most recent estimates of the carcinogenic risk of in-utero
radiation have moved from two extremes. The rst viewpoint
popularized by Stewart [8,10] suggested a risk of one or two orders
greater than the carcinogenic risk of postnatal exposure to children
and adults. Based on animal data and the Radiation Effects Research
Foundation (RERF) data, it appeared that the embryo and fetus may
actually have a lower risk than children exposed postnatally. Boice
and Miller [78] have concluded that the more recent data have
reduced the discrepancy between these two extreme viewpoints.
Based on the reports of Muirhead and Kneale [82] and Mole [89]
they believe that The risk estimate associated with intrauterine
radiation is not substantially greater than that seen following
childhood irradiation.
Doll and Wakeford [90] expressed their opinion about the
carcinogenic effect of intrauterine radiation and concluded that
Irradiation of the fetus in utero increases the risk of childhood
cancer, and increases the risk fromexposures of the order of 10 mGy,
and that in these circumstances the excess risk is w6% per Gy.
The atomic bomb survivors of in-utero irradiation have been
followed into adulthood, and the incidence of cancer in this pop-
ulation has been studied [86]. Although some of the exposures to
this population of pregnant women was considerably higher than
the exposure from the population exposed to radiation from
pelvimetry, there was only a small excess of adult tumors among
the atomic bomb survivors exposed in utero (Tables 2 and 3).
Boice and Miller [78] conclude:
Learned debate continues as to the causal nature of low level
intrauterine radiation and subsequent cancer risk. The associa-
tion is not questioned, but the etiologic signicance is. Different
scientists interpreting the same data have different opinions as
to the causal nature of the association and the possible level of
risk.
100
10
1
Fetus 1 10 20
Age (years)
Stewart et al.
810
ABCC
28,38,55
Animal data
1,2,61,6466,108
R
i
s
k
o
f
l
e
u
k
e
m
i
a
p
e
r
1
0
6
p
e
r
s
o
n
s
f
o
l
l
o
w
i
n
g
0
.
0
1
t
o
0
.
0
2
G
y
incorrect
30 40 50 60 70
Fig. 4. Risk of leukemia in children following radiation exposure during pregnancy.
This graph illustrates the modication proposed for Fig. 2, which indicates that the risk
of leukemia from radiation of the embryo/fetus is not one to two orders greater than
for the child, but actually lower than the risk for the child.
Table 3
Risk of leukemia.
a
Group Risk Latency
Identical twin of a leukemic twin 1:5 Weeks to months
Radiation-induced polycythemia 1:6 10e15 years
Bloom syndrome 1:8 <10 years of age
Hiroshima survivors <1000 m
hypocenter
1:60 3e12 years
Down syndrome 1:95 Weeks to months
Radiation treatment of ankylosing
spondylitis
1:270 15 years
Siblings of a leukemic child 1:720 10 years
Combined background risk of
leukemia plus radiation risk
from Stewart
1:2000 10 years
Additional risk of in-utero
diagnostic radiation studies
(Stewart et al. [8])
1:6000 10 years
In-utero diagnostic radiation
(RERF) data and other cohort
studies
Risk the same for
exposure during
childhood but
actual risk is
uncertain (Miller
[72]; Brent [73])
Lifetime
US Caucasian aged <15 years 1:3000 10 years
RERF, Radiation Effects Research Foundation.
a
Modied from Miller et al. [69].
Box 1
Arguments supporting a causal association between prenatal
radiation and childhood leukemia and cancer
1. Consistency. Practically all studies are statistically
consistent, with relative risk of 1.40 for leukemia
[23,66,91].
2. Dose response. Risk of childhood cancer was found to
increase with number of X-ray lms [68].
3. Coherence. Apparent lower risk of childhood cancer in
birth cohorts born in years when dose per lm was lower
[66,92].
4. Recall bias is unlikely to be a major factor [45].
5. Confounding variables have been sought, but none has
been found [68,34].
6. Selection bias related to reason for radiographic exami-
nation is not supported by caseecontrol studies of twins
[87,89].
7. Risk estimates after intrauterine exposures are generally
comparable to risks after childhood exposures for leu-
kemia [82,91].
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 208
Boxes 1 and 2 summarize the opinion of John Boice and Robert
Miller regarding the controversy pertaining to the carcinogenic risk
of ionizing radiation exposure to the developing embryo following
the National Council on Radiation Protection and Measurements
(NCRP) annual meeting dealing with the developmental, repro-
ductive, carcinogenic and mutagenic risks of ionizing radiation
published in 1999 [78].
During the period from 1997 to 2013 there were other publi-
cations dealing with the carcinogenic risk of exposing the embryo
to ionizing radiation besides Boice and Miller [78], the RERF pro-
gram in Japan [67] and the publications of Wakeford
[66,90,100,101]. Wakeford continued his estimate of the risks of
cancer following in-utero radiation and persisted with the
conclusion that the embryo was more vulnerable to the carcino-
genic effects of ionizing radiation than the risks in exposed
children.
The most important recent publication dealing with the carci-
nogenic risks of in-utero radiation was that by Preston et al. [67]
ICRP 90 was published in 2003 and was titled Effects of prenatal
irradiation (embryo and fetus). In 2013 an update of NCRP Hand-
book 54 was published as NCRP Report 174, Preconception and
prenatal radiation exposure: health effects and protective guid-
ance. [102] Dr Martha Linet had nal responsibility for the section
on oncogenesis. Contributions and suggestions were submitted by
other committee members to Dr Linet. Dr Roy Shores critiques
from the RERF program in Japan were particularly helpful.
Before reviewing the Preston et al. [67] the data pertaining to
the caseecontrol and cohort studies are examined.
2.1. Caseecontrol and cohort studies
Alice Stewart and her colleagues were the rst to indicate that
diagnostic radiological studies of pregnant women could signi-
cantly increase the risk of cancer in the offspring [9,10]. Similar
caseecontrol studies were performed in eight other countries and
they are summarized in Table 1. Some investigators were skeptical
of Dr Stewarts conclusions because the results were partly based
on medical histories obtained fromthe mother rather than fromthe
medical records. However, later analyses utilized primarily medical
records and the increased RR > 1.0 persisted. Attempts to deter-
mine the actual fetal exposure of the pregnant women were not
successful, since the exposures were never measured on the
pregnant women whose offspring were part of the analysis.
There are 35 caseecontrol studies listed in Table 1 dealing with
childhood lymphoblastic leukemia, acute myeloid leukemia, all
leukemias, CNS tumors, neuroblastoma, bone tumors, Ewing sar-
coma, rhabdomyosarcoma and total childhood cancer. Twenty-ve
of the caseecontrol studies were not statistically signicant. Ten of
the studies with RR >1 were statistically signicant. The consensus
was that the caseecontrol studies supported a RR of 1.2e1.3 based
on a meta-analysis [101].
3. Epidemiological cohort studies (Table 2)
There were 17 cohort studies of the offspring of womenwho had
been exposed to radiation during their pregnancy. None of the
studies individually reported a statistically increased RR for cancer
in the offspring. Cohort investigations to assess childhood cancer
risks among those undergoing diagnostic X-ray procedures
involving in-utero exposure included radiation-exposed pop-
ulations ranging in size from <200 to nearly 40 000 children. The
largest cohort study in this group was the report of Court Brown
et al. [50] This article was published in 1960 just a few years after
Giles et al. [9] and Stewart et al. [10] had indicated that the embryo
may be much more vulnerable to the carcinogenic effects of radi-
ation. There were 39 166 exposed and more than 1.5 million un-
exposed. There were nine leukemia subjects in the radiated
population and 14 in the controls [RR: 0.9; 95% condence interval
(CI): 0.4e1.6]. None of the cohort studies was statistically signi-
cant. The problemof cohort studies is that very large populations of
exposed individuals are needed, which was not the case with most
of these cohort studies.
4. Environmental exposures: atomic bomb survivors (Preston
et al. [67])
Although there was no evidence of a dose-related increase in
cancer mortality at ages prior to 15 years of age among the w2500
persons who were in utero at the time of the bombings [103], as the
cohort has grown older, a statistically signicant excess relative risk
(ERR) of solid cancers became apparent (ERR: 2.1 Gye1; 90% CI:
0.2e6.0), based on 10 deaths among those with weighted uterine
doses >0.01 Gy [80]. In a follow-up of cancer mortality risks during
1950 to 1992 comparing risks among a subset of persons who were
in utero versus those who were 0 to <6 years of age at the time of
the bombings, there were only two deaths from leukemia (both
exposed to relatively low doses and none during childhood) in the
in-utero cohort versus 24 among children <6 years of age at
exposure (Fig. 3) [80].
Subsequently, Preston et al. [67] compared solid cancer inci-
dence risks among in-utero cohort members aged 12e55 years
during 1958 to 1999 (based on 94 cancers) with risks among sur-
vivors who were aged <6 years of age at the time of the bombings
(based on 649 cancers). The difference in ERRs and excess absolute
risks (EARs) between the two cohorts suggests that lifetime cancer
risks at age 50 years following in-utero exposure are lower than
risks for early childhood exposures. However, the investigators
state, Additional follow-up of this cohort is necessary before
denitive conclusions can be made about the nature of the risk for
those exposed in utero [67] (Figs. 3 and 4). The difference in ERRs
and EARs between the two cohorts suggests that lifetime cancer
risks at 50 years of age following in-utero exposure are lower than
risks for early childhood exposure.
However, the investigators state that additional follow-up of
this cohort is necessary before denitive conclusions can be made
about the nature of the risks for those exposed in utero.
67
The
Box 2
Grounds for uncertainty regarding the causal nature of the as-
sociation between prenatal radiation and childhood cancer
1. Atomic bomb in-utero study nds no excess of childhood
cancer deaths [93], whereas a lower limit of 5.2 extra
cancer deaths was predicted from the risk model based
on obstetric X-ray data [7]. The central estimate of excess
cancer deaths predicted was about 10.
2. All major cohort studies are negative [22,50,91].
3. Biological implausibility; the equality of relative risks
associated with obstetric X-rays for leukemia and solid
tumors is perplexing given the variability in tissue
radiosensitivity, dissimilar origins, and different inci-
dence patterns [94,95]. The extended MacMahon study
did not nd an increased risk for solid cancers [34].
4. Risk estimates appear greater for in-utero versus
newborn exposures, for solid cancers [91].
5. Twin cohorts have lower risk of childhood cancer than
singletons despite more frequent X-rays [91,96,97].
6. Supporting animal evidence is weak [95,98,99].
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 209
investigators also note that this study cannot provide information
on the effect of radiation on the incidence of childhood cancers
because comprehensive data on solid cancer incidence are un-
available for the period from 1945 to 1957. Mortality follow-up for
the in-utero cohort, however, was available from 1950 and indi-
cated no deaths from childhood leukemia [80]. Another limitation
is the small numbers of cancers in each dose category in the in-
utero cohort.
Nevertheless, this investigation is the only cohort study with
long-term, continuous, active follow-up of a population with in-
utero radiation exposure and high-quality estimated doses for
each subject.
5. Animal studies
Many animal studies were performed in which pregnant ani-
mals were irradiated with ionizing radiation and the risk of cancer
in the offspring was evaluated in the offspring. A major problem
with these studies is that very few of the protocols utilized expo-
sures in the diagnostic range of clinical X-ray studies (<0.10 Gy). So
very few of the studies were planned to answer the question of the
risk of diagnostic radiation to pregnant women.
Several studies have reported excess risks of various tumors in
mice after in-utero irradiation, mostly after whole-body doses
higher than 2 Gy. Offspring of BC3F1 mice who received whole-
body in-utero doses (17th day post coitus) ranging from 0.3 to
2.1 Gy (41e58 animals in each dose group) developed small in-
creases in liver tumor occurrence [104]. Offspring of B6C3F1 mice
exposed to a whole-body dose of 3.8 Gy in utero developed
increased risk of pituitary, ovarian, liver, and bone tumors; an in-
crease in lung tumors was statistically signicant after doses of 1.9,
3.8, and 5.7 Gy; and an elevation in malignant lymphoma, lym-
phocytic type, was statistically signicant after 5.7 Gy [105].
In-utero irradiation [0.3 and 1 Gy (X-rays, whole body)] during
day 10.5 postconception (PT HT F
1
) did not induce an increased
incidence of neoplasms in the offspring [106].
Studies assessing tumor risks in different strains of mice
demonstrate high susceptibility of the ovaries for radiation-related
tumor induction during the fetal period, with 0.25 Gy the lowest
dose associated with a statistically signicant increase
[105,107,108].
Other investigations found no excess cancer after in-utero
irradiation of mice with 3 or 2 Gy [99,109], although each of
these studies showed increased risks of cancer in the mice
following administration of similar doses postnatally. Although
investigators found no excess cancer in BC3F1 mice after
in-utero exposures to 0.3 Gy, increased risks were seen in mice
given the same dose postnatally [84].
Rugh et al. [110] in a very large study irradiated mice with 1 Gy
on each day post conception and observed the incidence of tumors
in the offspring as adults. There was no statistically signicant in-
crease in the incidence of tumors in adult animals from irradiation
in utero on any day. Brent and Bolden [6] exposed pregnant mice to
doses of 0.3, 0.6, and 0.9 Gy at 0.5, 7.5, 8.5, 12.5, and 16.5 days post
conception. They also did not observe an increase in the incidence
of tumors. However, the pre-sexually mature mouse was more
vulnerable than the adult mouse to the leukemogenic effect of
radiation.
Offspring of pregnant beagles treated with mean doses of 0.16 or
0.81 Gy at 8th, 28th, or 55th days post coitus (120 dogs in each dose
and treatment day group) experienced increases in mortality from
total cancers that were not statistically signicant, and statistically
signicant elevated mortality risks from lymphoma. Detailed
assessment revealed that the increased risk of fatal neoplasms was
most pronounced in beagles irradiated in the neonatal period [111].
These data suggest that irradiation in both the fetal and neonatal
periods are associated with an increase of early onset and lifetime
cancer risk. However, the lower-dose group (0.16 Gy) did not have
an increased incidence of tumors.
Warkany et al. [112] studied the interaction of ethylnitrosourea
and X-irradiation in rats. The original goal of the investigators was
to determine the effect of X-irradiation administered on the 16th
day post conception on the incidence of tumors following the
administration of ethylnitrosourea on the 20th day post conception.
Sixteen months after delivery 62.2% of the rats that had received
only the ethylnitrosourea during the fetal period had neurogenic
tumors. After fetal irradiation on the 16th day post conception,
followed by ethylnitrosourea 4 days later, 16.7% of the rats devel-
oped neurogenic tumors. The mechanisms of these unexpected
ndings, whereby irradiation before receiving an oncogenic drug
reduced the incidence of cancer, have not been determined.
Nakano et al. [113] irradiated mice at various stages of preg-
nancy with 1 or 2 Gy. Translocation frequencies in the peripheral
blood T-cells, spleen cells, and bone-marrowcells were determined
when the offspring were aged 20 weeks. The translocation fre-
quency was very low in the mice that were irradiated in utero
(0.8%). The mice irradiated during days to weeks after birth had
translocation frequencies of 5%. The authors suggested that the
abnormal cells in the fetus were replaced by normal fetal stem cells
during the postnatal growth of the animal. If this phenomenon
occurs in humans, it could explain why the fetus may be less
vulnerable to the oncogenic effect of radiation than the child.
Earlier research supporting the ndings of Nakano et al. [113] found
that X-irradiation of the rat embryo during early organogenesis
resulted in the production of hundreds of small growths that
resembled well-differentiated ependymomas or retinoblastomas
(Fig. 1) [1,2]. As the embryo developed, some of the tumors dedif-
ferentiated into more primitive growths. However, at term almost
all the cytogenetically abnormal cells had regressed esimilar to the
result reported by Nakano et al [113].
6. Counseling patients about the in-utero carcinogenic risks
of ionizing radiation
Although it is our opinion that a dose of <0.10 Gy to the
implanted embryo does not result in a signicant increase risk for
congenital malformations, intrauterine growth restriction, or fetal
death (deterministic effects), low-risk tumorigenic or genetic haz-
ards cannot be ruled out. Even if one believed that the tumorigenic
(leukemogenic) effects of low-level radiation were real, let us
examine how difcult it would be to use this information in
counseling a patient who has received a dose of perhaps 2 rad
(0.02 Gy) during her pregnancy. According to Stewart et al. [7,8,10],
the risk of leukemia after this exposure in utero is 1:2000 versus
1:3000 in unexposed controls over a 10-year period (Table 3). If one
were inclined to recommend therapeutic abortion for this preg-
nancy because the probability of developing leukemia is 50%
greater than controls, one would perform abortions in almost 2000
exposed non-leukemic subjects for every leukemic subject saved.
It is one thing to avoid radiation because of a potential or conjec-
tured hazard, but it is another matter to recommend therapeutic
abortion on this basis. If a physician were inclined to accept this
increased probability (1:2000) as a risk great enough to recom-
mend therapeutic abortion, he or she would be placed in a serious
dilemma because there are other epidemiologic situations in which
the risk of leukemia is greater. In fact, the hypothetical incremental
risk for 2 rad of in-utero radiation is 1:6000 over a 10-year period. It
is the combination of the control risk plus the incremental radiation
risk that results in a 1:2000 risk for these patients. From Table 1 it
should be clear that the risk of leukemia is greater in unirradiated
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 210
siblings of leukemics (1:720) than in patients subjected to diag-
nostic radiation (1:2000) if one uses Stewarts risk estimate.
Certainly, the position that all future pregnancies of parents
with one leukemic child should be aborted would be untenable.
One can carry this argument to its ridiculous extreme by advocating
that all pregnancies should be aborted because of the risk of serious
malformations is w30 per 1000 deliveries (Table 3) and this does
not include the probability of postnatal diseases occurring in these
offspring. Some may interpret this as a facetious discussion, but the
clinician and the patient must recognize that spontaneous risks of
pregnancy are two orders of magnitude greater than the theoretical
risks of diagnostic radiation (Box 1).
7. Conclusions and counseling advice
The radiation risks determined from the A-bomb exposure to
the populations in Hiroshima and Nagasaki have been referred to as
the "gold standard" for determining radiation risks because the
study was a large cohort study with a major effort to determine the
actual exposure of each survivor. Yet we know that if individuals or
populations are exposed to x-rays or gamma rays of different
photon energy or length of exposure, the risks may deviate from
the risks determined for the A-bomb survivors. Furthermore, there
was a neutron component in the radiation from the A-bomb det-
onations. So one can assume that exposures to the embryo from an
IVP or uoroscopy in a pregnant mother may represent a different
risk to the embryo per mGy than the A-bomb data. So that is why it
is problematic to provide denitive carcinogenic risks for diag-
nostic radiologic studies utilizing x-rays or radionuclides. This
article has discussed the subject of the carcinogenic risk of ionizing
radiation to the embryo.
1. There is no doubt that if the exposure is high enough and of
short duration, the carcinogenic risks are increased. Protracted
continuous radiation of 0.02 Gy/day does not represent an
increased carcinogenic risk. Acute radiation of 1.0 Gy at mid-
gestation does represent an increased carcinogenic risk.
2. The embryo is less vulnerable to the carcinogenic effects of ra-
diation the earlier in gestation it is exposed.
3. There are at least three viewpoints on the carcinogenic risks of
<0.10 Gy embryonic radiation.
(a) There is the scholarly, conservative view of Martha Linet
who writes that the risk is very small and would not justify
canceling a radiological study in a pregnant woman if the
study is medically indicated. She also suggests that we wait
to determine whether the risk increases based on future
data from the Preston et al. study, which stated that addi-
tional follow-up of this cohort is necessary before denitive
conclusions can be made about the nature of the risks for
those exposed in utero. [67]
(b) Richard Wakeford has been interested in this subject for
decades. We rst met many years ago when we were de-
fense experts in litigation between the UK and Ireland
regarding the allegation that the Sellaeld Nuclear Facility
was discharging nuclear waste that was responsible for an
increase in cancer and birth defects in the inhabitants on the
East Coast of Ireland. The World Court deliberations ended
after 10 years with a defense verdict. Wakefords views have
changed with time. One of his most recent publications in-
dicates that he believes that 20% of childhood leukemia in
the UK is due to background radiation [114]. He still is the
proponent of the idea that the embryo is more vulnerable to
the carcinogenic effects of radiation than the child.
(c) I am not one who is reluctant to make predictions. I agree
with Martha Linet regarding the risks of embryonic ionizing
radiation. However, I would predict that in the next 20 years
we will learn that the risk of cancer from embryonic radia-
tion will be further reduced from the ndings of the Preston
et al. 2008 study. At my present age I will not be alive to
know the results. I believe that the omnipotential (stem)
cells protective effect that was present in the embryo at the
time of the radiation will continue to be manifested.
7.1. Counseling an individual patient
If a pregnant woman has had a diagnostic radiological proce-
dure that exposed her embryo or who has been scheduled for an X-
ray that will expose her embryo and is concerned about the
increased risk of cancer fromthe exposure, howshould a counselor
respond?
The majority of diagnostic radiological studies expose the
embryo to <0.10 Gy (<10 rad) which is a very low exposure. Based
on all the studies we have available, the risk of cancer to the embryo
is very low and possibly so low that we may never be able to
measure the risk. Therefore, diagnostic radiological studies that are
considered to be important for optimal patient care should be
performed. It is important to be aware of the background risk of
cancer for all individuals, which is 23% for potentially lethal can-
cers. Fortunately, each year the percentage of cancers that are cured
is increasing. The background risk of cancer is hundreds or more
times the theoretical risks of diagnostic radiological exposures to
the embryo.
Conict of interest statement
None declared.
Funding sources
None.
References
[1] Brent RL, Jordan HC. Possible late tumor development in rat embryos irra-
diated with 100 r of x-radiation on the 9th day of gestation. Anat Rec
1951;109:13 (abstract).
[2] Wilson JG, Brent RL, Jordan HC. Neoplasia induced in rat embryos by
roentgen irradiation. Cancer Res 1952;12:222e8 (also appeared in USAECD
UR-183, 1951).
[3] Russell LB. X-ray induced developmental abnormalities in the mouse and
their use in analysis of embryological patterns. I. External and gross visceral
changes. J Exp Zool 1950;114:345e602.
[4] Russell LB. X-ray-induced developmental abnormalities in the mouse and
their use in analysis of embryological patterns. II. Abnormalities of the
vertebral column and thorax. J Exp Zool 1956;131:329e95.
[5] Wilson JG, Brent RL, Jordan HC. Differentiation as a determinant of the re-
action of rat embryos to x-irradiation. Proc Soc Exp Biol Med 1953;82:67e70
(also appeared in USAECD UR-243, 1953).
[6] Brent RL, Bolden BT. Indirect effect of x-irradiation on embryonic develop-
ment. V. Utilization of high doses of maternal irradiation on the rst day of
gestation. Rad Res 1968;36:563e70.
[7] Stewart A, Kneale GW. Radiation dose effects in relation to obstetric x-rays
and childhood cancers. Lancet 1970;1:1185e8.
[8] Stewart A, Webb D, Giles D. Malignant disease in childhood and diagnostic
irradiation in utero. Lancet 1956;2:447.
[9] Giles D, Hewitt D, Stewart A, Webb J. Malignant disease in childhood and
diagnostic irradiation in utero. Lancet 1956;271(6940):447.
[10] Stewart A, Webb D, Hewitt D. A survey of childhood malignancies. BMJ
1958;1:1495e508.
[11] National Academy of Sciences/National Research Council. The effects of
populations of exposure to low levels of ionizing radiation. Report of the
Advisory Committee on the Biological Effects of Ionizing Radiations. Wash-
ington, DC: NAS/NRC; 1972.
[12] Lewis EB. Leukemia and ionizing radiation. Science 1957;125:865e972.
[13] Short & others v. BNFL and others. Legal case, World Court decision, 1999.
[14] Kneale GW, Stewart AM. Age variation in the cancer risks from foetal irra-
diation. Br J Cancer 1977;36:501e10.
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 211
[15] Klein M. Induction of lung adenomas following exposure of pregnant,
newborn, and immature male mice to urethane. Cancer Res 1954;14:438e40.
[16] Vesselinovitch SD, Mihailovich N, Rao KVN, Itze L. Perinatal carcinogenesis
by urethane. Cancer Res 1971;31:2143e7.
[17] Lilienfeld AM. Epidemiological studies of the leukemogenic effects of radia-
tion. Yale J Biol Med 1966;39:143e64.
[18] Polhemus D, Koch R. Leukemia and medical irradiation. Pediatrics 1959;23:
453e61.
[19] Ford D, Patterson T. Fetal exposure to diagnostic x-rays and leukemia and
other malignant diseases in childhood. J Natl Cancer Inst 1959;22:1093e104.
[20] Graham S, Levin MI, Lilienfeld AM. Preconception, intrauterine and postnatal
irradiation as related to leukemia. J Natl Cancer Inst Monogr 1966;19:347e71.
[21] MacMahon B. Prenatal x-ray exposure and twins. N Engl J Med 1985;312:
576e7.
[22] Diamond EL, Schmerler H, Lilienfeld AM. The relationship of intrauterine
radiation to subsequent mortality and development of leukemia in children:
a prospective study. Am J Epidemiol 1973;97:283e313.
[23] MacMahon B, Hutchinson GB. Prenatal x-ray and childhood: a review. Acta
Union Int Contra Cancrum 1964;20:1172e4.
[24] Bithell JF, Stewart AM. Prenatal irradiation and childhood malignancy: a
review of British data from the Oxford Survey. Br J Cancer 1975;31:271e87.
[25] Van Steensel-Moll HA, Valkenburg HA, Vandenbroucke JP, Van Zanen GE. Are
maternal fertility problems related to childhood leukaemia? Int J Epidemiol
1985;14:555e9.
[26] Shu XO, Gao YT, Brinton LA, Linet MS, Tu JT, Zheng W, et al. A population-
based caseecontrol study of childhood leukemia in Shanghai. Cancer
1988;62:635e44.
[27] Magnani C, Pastore G, Luzzatto L, Terracini B. Parental occupation and other
environmental factors in the etiology of leukemias and non-Hodgkins
lymphomas in childhood: a caseecontrol study. Tumori 1990;76:413e9.
[28] Naumberg E, Bellocco R, Cnattingius S, Hall P, Boice Jr JD, Ekbom A. Intra-
uterine exposure to diagnostic x-rays and risk of childhood leukemia sub-
types. Radiat Res 2001;156:718e23.
[29] Shu XO, Potter JD, Linet MS, Severson RK, Han D, Kersey JH, et al. Diagnostic
x-rays and ultrasound exposure and risk of childhood acute lymphoblastic
leukemia by immunophenotype. Cancer Epidemiol Biomarkers Prev
2002;11:177e85.
[30] Van Duijn CM, Van Steensel-Moll HA, Coebergh JW, Van Zanen GE. Risk
factors for childhood acute non-lymphocytic leukemia: an association with
maternal alcohol consumption during pregnancy? Cancer Epidemiol Bio-
markers Prev 1994;3:457e60.
[31] Kaplan HS. An evaluation of the somatic and genetic hazards of the medical
uses of radiation. Am J Roentgenol Radium Ther Nuc Med 1958;80:696e706.
[32] Salonen T, Saxen L. Risk indicators in childhood malignancies. Int J Cancer
1975;15:941e6.
[33] Hirayama T. Descriptive and analytical epidemiology of childhood malig-
nancy in Japan. In: Kobayashi N, editor. Recent advances in management of
children with cancer. Tokyo: Childrens Cancer Association of Japan; 1979.
pp. 27e43.
[34] Monson RR, MacMahon B. Prenatal x-ray exposure and cancer in children. In:
Boice Jr JD, Fraumeni JF, editors. Radiation carcinogenesis: epidemiology and
biological signicance. New York: Raven Press; 1984. pp. 97e105.
[35] Shu X-O, Reaman GH, Lampkin B, Sather HN, Pendergrass TW, Robison LL.
Association of paternal diagnostic x-ray exposure with risk of infant leuke-
mia. Investigators of the Childrens Cancer Group. Cancer Epidemiol Bio-
markers Prev 1994;3:645e53.
[36] Infante-Rivard C, Mathonnet G, Sinnett D. Risk of childhood leukemia asso-
ciated with diagnostic irradiation and polymorphisms in DNA repair genes.
Environ Health Perspect 2000;108:495e8.
[37] Infante-Rivard C, Deadman JE. Maternal occupational exposure to extremely
low frequency magnetic elds during pregnancy and childhood leukemia.
Epidemiology 2003;14:437e41.
[38] Rajaraman P, Simpson J, Neta G, Berrington de Gonzalez A, Ansell P, Linet MS,
et al. Early life exposure to diagnostic radiation and ultrasound scans and
risk of childhood cancer: caseecontrol study. Br Med J 2011;342:d472.
[39] Preston-Martin S, Yu MC, Benton B, Henderson BE. N-Nitroso compounds
and childhood brain tumors: a caseecontrol study. Cancer Res 1982;42:
5240e5.
[40] Bunin GR, Buckley JD, Boesel CP, Rorke LB, Meadows AT. Risk factors for
astrocytic glioma and primitive neuroectoderman tumor of the brain in
young children: a report from the Childrens Cancer Group. Cancer Epidemiol
Biomarkers Prev 1994;3:197e204.
[41] Schuz J, Kaletsch U, Kaatsch P, Meinert R, Michaelis J. Risk factors for
pediatric tumors of the central nervous system: results from a German
population-based caseecontrol study. Med Pediatr Oncol 2001;36:274e
82.
[42] Stalberg K, Haglund B, Axelsson O, Cnattingius S, Pfeifer S, Kieler H. Prenatal
x-ray exposure and childhood bran tumours: a population-based case-con-
trol study on tumor subtypes. Br J Cancer 2007;97:1583e7.
[43] Winn DM, Li FP, Robison LL, Mulvihill JJ, Daigle AE, Fraumeni Jr JF. A casee
control study of the etiology of Ewings sarcoma. Cancer Epidemiol Bio-
markers Prev 1992;1:525e32.
[44] Gufferman S, Ruymann F, Ognjanovic S, Erhardt EB, Maurer HM. Prenatal x-
ray exposure and rhabdomyosarcoma in children: a report from the Chil-
drens Oncology Group. Cancer Epidemiol Biomarkers Prev 2009;18:1271e6.
[45] MacMahon B. Prenatal x-ray exposure and childhood cancer. J Natl Cancer
Inst 1962;28:1173e91.
[46] Little MP. A comparison of the risk of stillbirth associated with paternal pre-
conception irradiation in the Sellaeld workforce with that of stillbirth and
untoward pregnancy outcome among Japanese atomic bomb survivors.
J Radiol Prot 1999;14:361e73.
[47] Knox EG, Stewart AM, Kneale GW, Gilman EA. Prenatal irradiation and
childhood cancer. J Soc Radiol Prot 1987;7:177e88.
[48] Ross MH, Bras G. Tumor incidence patterns and nutrition in the rat. J Nutr
1965;87:245e60.
[49] Shiono PH, Chung CS, Myrianthopoulos NC. Preconception radiation, intra-
uterine diagnostic radiation, and childhood neoplasia. J Natl Cancer Inst
1980;65:681e6.
[50] Court-Brown WM, Doll R, Hill RB. Incidence of leukemia after exposure to
diagnostic radiation in utero. BMJ 1960;2:1539e45.
[51] Salonen T. Prenatal and perinatal factors in childhood cancer. Ann Clin Res
1976;8:27e42.
[52] Murray R, Heckel P, Hempelmann LH. Leukemia in children exposed to
ionizing radiation. N Engl J Med 1959;261:585e9.
[53] Lewis TLT. Leukaemia in childhood after antenatal exposure to x-rays. Br
Med J 1960;2:1551.
[54] Griem ML, Meier P, Dobben G. Analysis of the morbidity and mortality of
children irradiated in fetal life. Radiology 1967;88:347e9.
[55] Oppenheim BE, Griem ML, Meier P. Effects of low-dose prenatal irradiation in
humans: analysis of Chicago lying-in data and comparisons with other
studies. Radiat Res 1974;57:508e44.
[56] Oppenheim BE, Griem ML, Meier P. The effects of diagnostic x-ray exposure
on the human fetus. An examination of the evidence. Radiology 1975;114:
529e34.
[57] Golding J, Paterson M, Kinlen LJ. Factors associated with childhood cancer in
a national cohort study. Br J Cancer 1990;62:304e8.
[58] ICRP International Commission on Radiological Protection. Pregnancy and
medical radiation, ICRP Publication 84, Ann ICRP, vol. 30(1). New York:
Elsevier; 2003.
[59] Dempster WH. Radiation hazard in antenatal radiography. Lancet 1958;272:
159.
[60] Milis SD, Bruwer AJ, Banner EA, Davis GD, Gage RP. Effects of irradiation of
the fetus: ten-year follow-up of pelvimetry during pregnancy. Minn Med
1958;41:339e41.
[61] Lejeune J, Turpin R, Rethore MO, Mayer M. Results of a rst investigation of
the somatic effects of fetoembryonal irradiation in utero (special case of
heterochromia iridis). Rev Fr Etudes Clin Biol 1960;5:982e9 (in French).
[62] Magnin P. The fate of infants irradiated in utero. Analysis of a survey
involving 5,353 cases. Presse Med 1962;70:1199e202 (in French).
[63] Nokkentved K. Effects of radiation upon the human fetus. Follow-up study of
152 children exposed to irradiation during the rst 4 months of foetal life
due to X-ray examination of the maternal abdomen. Copenhagen: Ejnar
Munksgaards Forlag; 1968.
[64] Hagstrom RM, Glasser SR, Brill AB, Heyssel RM. Long term effects of radio-
active iron administered during human pregnancy. Am J Epidemiol 1969;90:
1e10.
[65] Ray JG, Schull MJ, Urquia ML, You JJ, Guttmann A, Vermeulen MJ. Major
radiodiagnostic imaging in pregnancy and the risk of childhood malignancy:
a population-based cohort study in Ontario. PloS Med 2010;7:1e8.
[66] Doll R, Wakeford R. Risk of childhood cancer from fetal irradiation. Br J Radiol
1997;70:130e9.
[67] Preston DL, Cullings H, Suyama A, Funamoto S, Nishi N, Soda M, et al. Solid
cancer incidence in atomic bomb survivors exposed in utero or as young
children. J Natl Cancer Inst 2008;100:428e36.
[68] Hoshino T, Itoga T, Kato H. Leukemia in the offspring of parents exposed to
the atomic bomb at Hiroshima and Nagasaki. Presented at the Japanese
Association of Hematology, Japan, 25e30 March 1965.
[69] Miller RW. Epidemiological conclusions from radiation toxicity studies. In:
Fry RJM, Grahn D, Greim ML, editors. Late effects of radiation. London: Taylor
& Francis; 1970. pp. 245e56.
[70] Brent RL, Gorson RO. Radiation exposure in pregnancy. In: Moseley RD,
Baker DH, Gorson RO, editors. Current problems in radiology. Chicago: Year
Book Medical Publishers; 1972. pp. 1e48.
[71] Jablon S. Comments. Health Phys 1973;24:257e8.
[72] Miller RW. Discussion. Severe mental retardation and cancer among atomic
bomb survivors exposed in utero. Teratology 1999;59:234e5.
[73] Brent RL. Utilization of developmental basic science principles in the eval-
uation of reproductive risks from pre- and postconception environmental
radiation exposures. Teratology 1999;59:182e204.
[74] Brent RL. Radiation teratogenesis. Teratology 1980;21:218e98.
[75] Wood J, Johnson KG, Omori Y. Mental retardation in children exposed in
utero to the atomic bombs in Hiroshima and Nagasaki. Am J Public Health
1967;57:1381e90.
[76] Wood J, Keehn R, Kawamoto S. The growth and development of children
exposed in utero to the atomic bombs in Hiroshima and Nagasaki. Am J
Public Health 1967;57:1374e80.
[77] Maroteaux P, Spranger J, Opitz JM. Le syndrome campomelique. Presse Med
1971;22:1157e62.
[78] Boice Jr JD, Miller RW. Childhood and adult cancer following intrauterine
exposure to ionizing radiation. Teratology 1999;59:227e33.
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 212
[79] Boice Jr JD, Inskip PD. Radiation-induced leukemia. In: Henderson ES,
Lister TA, Greaves MF, editors. Leukemia. 6th ed. Philadelphia: Saunders;
1996. pp. 195e209.
[80] Delongchamp RR, Mabuchi K, Yoshimoto Y, Preston DL. Cancer mortality
among atomic bomb survivors exposed in utero or as young children,
October 1950eMay 1992. J Radiat Res (Tokyo) 1997;147:385e95.
[81] Miller RW, Boice Jr JD. Cancer after intrauterine exposure to the atomic
bomb. J Radiat Res (Tokyo) 1997;147:396e7.
[82] Muirhead CR, Kneale GW. Prenatal irradiation and childhood cancer. J Radiol
Prot 1989;9:209e12.
[83] National Radiological Protection Board. Diagnostic medical exposures:
advice on exposure to ionizing radiation during pregnancy. Joint Guidance
from NRPB/CoR/RCR. Chilton, UK: NRPB; 1998.
[84] Pierce DA, Shimizu Y, Preston DL, Vaeth M, Mabuchi K. Studies of the mor-
tality of atomic bomb survivors. Report 12, Part I. Cancer: 1950e1990.
J Radiat Res (Tokyo) 1996;146:1e27.
[85] Thompson DE, Mabuchi K, Ron E, Soda M, Tokunaga M, Ochikubo S, et al.
Cancer incidence in atomic bomb survivors. Part II. Solid tumors 1958e1987.
J Radiat Res (Tokyo) 1994;137:S17e67.
[86] Yoshimoto Y, Delongchamp R, Mabuchi K. In utero exposed atomic bomb
survivors: cancer risk update. Lancet 1994;344:345e6.
[87] Harvey EB, Boice Jr JD, Honeyman M, Flannery JT. Prenatal x-ray exposure
and childhood cancer in twins. N Engl J Med 1985;312:541e5.
[88] Rodvall Y, Pershagen G, Hrubec Z, Ahlbom A, Pedersen NL, Boice Jr JD.
Prenatal x-ray exposure and childhood cancer in Swedish twins. Int J Cancer
1990;46:362e5.
[89] Mole RH. Antenatal irradiation and childhood cancer: causation or coinci-
dence? Br J Cancer 1974;30:199e208.
[90] Doll R, Wakeford R. Carcinogenesis after exposure to ionizing radiation.
Invest Rad 2000;35:527e33.
[91] UNSCEAR United Nations Scientic Committee on the Effects of Atomic Ra-
diation. Sources and effects of ionizing radiation, UNSCEAR 1994. Report to
the General Assembly with Scientic Annexes, No. E.94IX.11. New York:
United Nations Publications; 1994.
[92] Bithell JF, Stiller CA. A new calculation of the carcinogenic risk of obstetric x-
raying. Stat Med 1988;7:857e64.
[93] Jablon S, Kato H. Mortality among a-bomb survivors, 1950e1970. Technical
report 10-71. Hiroshima: Radiation Effects Research Foundation; 1970.
[94] Miller RW. Delayed radiation effects in atomic-bomb survivors. major ob-
servations by the Atomic Bomb Casualty Commission are evaluated. Science
1969;166:569e74.
[95] NAS/NRC (National Academy of Sciences/National Research Council). Health
effects of exposure to low levels of ionizing radiation. Committee on the
Biological Effects of Ionizing Radiations. BEIR V. Washington, DC: National
Academy Press; 1990.
[96] Inskip PD, Harvey EB, Boice Jr JD, Stone BJ, Matanoski G, Flannery JT, et al.
Incidence of childhood cancer in twins. Cancer Causes Control 1991;2:
315e24.
[97] Rodvall Y, Pershagen G, Hrubec Z, Alhbom A, Pedersen NL, Boice Jr JD.
Childhood cancer among Swedish twins. Cancer Causes Control 1992;3:
527e32.
[98] UNSCEAR (United Nations Scientic Committee on the Effects of Atomic
Radiation). Ionizing radiation: levels and effects. A Report of the United
Nations Scientic Committee on the Effects of Atomic Radiation to the
General Assembly, with Annexes. No. E.72.IZ.17. New York: United Nations
Publications; 1986.
[99] Upton AC, Odell Jr TT, Sniffen EP. Inuences of age at time of irradiation on
induction of leukemia and ovarian tumors in RF mice. Proc Soc Exp Biol Med
1960;104:769e72.
[100] Doll R, Wakeford R. Survivors exposed in utero or as young children. J Natl
Cancer Inst 2008;100:1482e3.
[101] Wakeford R, Little MP. Risk coefcients for childhood cancer after intra-
uterine irradiation: a review. Int J Radiat Biol 2003;79:293e309.
[102] Brent RL, Bushberg JT, Linet M, et al., editors. NCRP Report No. 174. Pre-
conception and prenatal radiation exposure: health effects and protective
guidance 24 May 2013. p. 351. Recommendations of the National Council on
Radiation Protection and Measurements, Bethesda, MD, USA.
[103] Preston DL, Kusumi S, Tomonaga M, Izami S, Ron E, Kuramoto A, et al. Cancer
incidence in atomic bomb survivors. Part III. Leukemia, lymphoma and
multiple myeloma, 1950e1987. Radiat Res 1994;137(2 Suppl.):S68e97.
[104] DiMajo V, Coppola M, Rebessi S, Covelli V. Age-related susceptibility of
mouse liver to induction of tumors by neutrons. Radiat Res 1990;124:
227e34.
[105] Sasaki S. Inuence of the age of mice at exposure to radiation on life-
shortening and carcinogenesis. J Radiat Res (Tokyo) 1991;32(Suppl.):
73e85.
[106] Nomura T, Nakajima H, Hatanaka T, Kinuta M, Hongyo T. Embryonic muta-
tion as a possible cause of in utero carcinogenesis in mice revealed by
postnatal treatment with 12-O-tetradecanoylphorbol-13-acetate. Cancer Res
1990;50:2135e8.
[107] Schmahl W, Kriegel H, Senft E. Can prenatal x-irradiation in mice act as an
initiator stimulus in a modied 2-stage Berenblum/Mottram experiment
with postnatal promotion with phorbol ester TPA? J Cancer Res Clin Oncol
1980;97:109e17.
[108] Uma-Devi P, Hossain M. Induction of solid tumours in the Swiss albino
mouse by low-dose foetal irradiation. Int J Radiat Biol 2000;76:366e72.
[109] Ellender M, Harrison JD, Kozlowski R, Suzluinska M, Boufer SD, Cox R. In
utero and neonatal sensitivity of ApeMin/ mice to radiation-induced in-
testinal neoplasia. Int J Radiat Biol 2006;82:141e51.
[110] Rugh R, Duhamel L, Skaredoff L. Relation of embryonic and fetal x-irradiation
to life time average weights and tumor incidence in mice. Proc Soc Exp Biol
Med 1966;121:714e8.
[111] Benjamin SA, Lee AC, Angleton GM, Saunders WJ, Keefe TJ,
Mallinckrodt CH. Mortality in beagles irradiated during prenatal and
postnatal development. I. Contribution of non-neoplasia diseases. Radiat
Res 1998;150:330e48.
[112] Warkany J, Mandybur TI, Kalter H. Oncogenic response of rats with x-ray-
induced microcephaly to transplacental ethylnitrosourea. J Natl Cancer Inst
1976;56:59e64.
[113] Nakano M, Kodama Y, Ohtaki K, Nakashima E, Niwa O, Toyoshima M, et al.
Chromosome aberrations do not persist in the lymphocytes or bone marrow
cells of mice irradiated in utero or soon after birth. Radiat Res 2007;167:
693e702.
[114] Wakeford R, Kendall GM, Little MP. The proportion of childhood leukaemia
incidence in Great Britain that may be caused by natural background
ionizing radiation. Leukemia 2009;24:770e6.
R.L. Brent / Seminars in Fetal & Neonatal Medicine 19 (2014) 203e213 213
Review
New genetic testing in prenatal diagnosis
Natalia Babkina, John M. Graham Jr.
*
Medical Genetics Institute, CedarseSinai Medical Center and Division of Medical Genetics, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA
Keywords:
Chromosomal microarray
Exome sequencing
Next-generation sequencing
Non-invasive prenatal screening
Preimplantation genetic testing
s u m m a r y
Determining a genetic diagnosis prenatally permits patients to make informed reproductive decisions
and to be counseled about possible fetal outcomes. Therefore, it is important for the provider to be aware
of the spectrum of genetic conditions and to use appropriate testing modality to obtain specic diag-
nosis. This article reviews genetic techniques available for prenatal diagnosis such as preimplantation
genetic testing, chromosomal microarray, non-invasive prenatal screening, and next-generation
sequencing. Chromosomal microarray has emerged as the rst diagnostic test for evaluation of multi-
ple congenital anomalies and developmental delay as most of the next-generation sequencing methods
do not detect copy-number variants (CNVs). Exome sequencing and whole genome sequencing are time-
consuming, so if this needs to be done to obtain an accurate genetic diagnosis, allow sufcient time.
2013 Elsevier Ltd. All rights reserved.
1. Introduction
Prenatal diagnosis plays an essential role in contemporary
obstetrical care, and with proper planning, it can be available for
pregnancies at risk for chromosomal and single-gene disorders.
Standard chromosome testing has been available since the 1960s
[1] and has been commonly used in the prenatal setting, when a
fetus has abnormal ndings on ultrasound. Prenatal genetic testing
has now become much more sophisticated with an improved level
of resolution. Standard chromosome testing is being superseded by
the use of chromosomal microarrays. Most forms of prenatal
diagnosis require invasive procedures for fetal-sample collection
and therefore, although considered safe, these procedures involve a
risk of fetal loss. The presence of cell-free fetal DNA in maternal
circulation has been used as a basis for the development of
non-invasive prenatal testing [2]. Application of the latest tech-
nologies, such as next-generation sequencing, which features both
high sensitivity and accuracy, is allowing for more accurate pre-
conception counseling when there is a previously affected relative,
thereby broadening the scope of prenatal diagnosis to include
many diseases that are both paternally and maternally inherited.
Determining genetic diagnosis prenatally permits patients to
make informed reproductive decisions and to be counseled about
possible fetal outcomes, management options and recurrence
risks. Therefore, it is important for the physician to be aware of the
full spectrum of genetic conditions, and to use appropriate testing
and referrals to genetic healthcare providers in order to obtain a
specic diagnosis.
2. Preconception counseling
The prevalence of paternal and maternal conditions that are
relevant to pregnancy outcomes varies according to many factors,
such as parental age, ethnicity, medical history, and family history.
Family history plays a critical role in assessing the risk of
inherited medical conditions and single-gene disorders [3]. In
general practice, the family history can be obtained using a
questionnaire or a three-generation pedigree. A family history
screening allows stratication of the risk level. Also, the use of a
family history screening has been shown to increase the likelihood
of detecting a patient at high risk of developing an inherited
medical condition by 20%, compared with medical record review
alone [4]. Family history of developmental delay, congenital mal-
formations, or other constellation of clinical ndings suggestive of
a genetic condition requires thorough evaluation. If a disorder in
the individuals family has been identied as having a genetic
cause, it may be possible to test parents to determine the risk for
having an affected child. For example, a family history of known
genetic conditions, such as TayeSachs or cystic brosis, should
prompt testing for at least the person with an affected relative. It is
paramount to identify the mutation in an affected individual, so
prenatal diagnosis or preimplantation genetic diagnosis in cases of
in-vitro fertilization (IVF) can be performed in order to test spe-
cically for an identied familial mutation and decrease the risk of
affected pregnancy. A history of recurrent pregnancy loss should
* Corresponding author. Address: 8700 Beverly Blvd, PACT Suite 400, Los Angeles,
CA 90048, USA. Tel.: 1 310 423 9909; fax: 1 310 423 2080.
E-mail address: john.graham@cshs.org (J.M. Graham).
Contents lists available at ScienceDirect
Seminars in Fetal & Neonatal Medicine
j ournal homepage: www. el sevi er. com/ l ocat e/ si ny
1744-165X/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.siny.2013.10.005
Seminars in Fetal & Neonatal Medicine 19 (2014) 214e219
prompt testing of both parents for chromosomal translocations. A
parent with a balanced translocation identied by standard
chromosome analysis can produce normal, balanced or unbal-
anced gametes that result in normal, balanced or unbalanced
fetus. Recent data show that apparently balanced chromosomal
rearrangements are associated with an abnormal phenotype in
6.7% of cases that may be due to cryptic genomic imbalances or to
the disruption of genes at the breakpoints [5].
Whereas some genetic conditions that affect pregnancy out-
comes are easily identied early in life, others are not and may
require additional diagnostic testing. Up until about 10 years ago,
w65e70% of patients with congenital malformations remained
undiagnosed, including many patients with multifactorial defects
(e.g. non-syndromic cleft lip with or without cleft palate or neural
tube defects) or polygenic defects (e.g. Hirschsprung disease or
hypospadias), 15e25% were thought to be genetic or genomic, and
10% were thought to be environmentally determined [6].
Previously, the diagnostic yield for children with intellectual
disability varied between 50% and 80%, with 17e47% having a ge-
netic/genomic cause and 20e50% remaining undiagnosed [7]. With
the advent of high-resolution chromosomal microarrays, an addi-
tional 10e15% of such patients receive a genomic diagnosis (which
is often a sporadic occurrence in an otherwise normal family) [8,9].
As the use of exome sequencing has increased over the last 2 years,
most laboratories are reporting an additional 25e40% success rate
in obtaining a genetic diagnosis (some of which are also sporadic).
The practical implication of all these recent developments is that
with modern genomic/genetic diagnostic techniques, up to 70e80%
of patients can receive a genetic/genomic diagnosis, with an addi-
tional 10% attributed to environmental inuences and maternal
conditions. Multifactoral defects such as isolated congenital heart
defects or orofacial clefts have had a large number of genetic and
environmental inuences identied, but genetic diagnosis for
multifactoral defects is not currently available for clinical use.
Fortunately, recurrence risks are relatively low, in the 3e5% range,
and this recurrence risk can be reduced by up to 50% through use of
folic acid prior to conception and during early pregnancy (covered
elsewhere in this volume). Exome sequencing can be quite useful
for polygeneic or conditions manifesting genetic heterogeneity.
Ancestry inuences the probability of being a carrier of many
disorders that affect pregnancy. Typically, there is no family history
of the condition as the carriers are asymptomatic. For example,
using the current recommended mutation panel, negative carrier
testing for cystic brosis in a Caucasian couple would result in a
different chance of having an affected child than negative testing in
an African-American couple since the carrier rate and the type of
mutations are different in these two populations.
Parental age is an important risk factor for adverse pregnancy
outcomes. The risk of chromosomal non-disjunction increases with
increasing maternal age that translates into higher risk of trisomy
21, 13 and 18. Recent data emphasize the importance of paternal
age as a risk factor for newdominant mutations. The male germline
accumulates point mutations due to replication errors and reduced
activity of repair enzymes, strands mispairing of short tandem re-
peats and longer exposure to environmental mutagens [10]. In
addition, in human sperm DNA is more methylated than oocyte
DNA, which may account for the greater number of paternally
derived point mutations occurring within a CpG dinucleotide [11].
Because of the large number of cell divisions during spermato-
genesis, the mutation rate for base substitutions is much higher in
men than in women, and increases with paternal age. The risk for
de-novo autosomal dominant mutations calculated by Friedman
was 0.3e0.5% among the offspring of fathers aged >40 years [12].
Recently, with the use of whole-genome sequencing the increase in
the rate of de-novo mutations has been estimated to be two
mutations per year [13]. The conditions most strongly associated
with advanced paternal age are those caused by point mutations in
the FGFR2, FGFR3 and RET genes and include Pfeiffer syndrome,
Crouzon syndrome, Apert syndrome, achondroplasia as well as
multiple endocrine neoplasia type 2A (MEN2A) and multiple
endocrine neoplasia type 2B (MEN2B) [14]. Certain dominant
conditions such as neurobromatosis type 1 (NF-1) caused by point
mutations or small deletions show a lesser association with
paternal age. Some genetic conditions such as Noonan syndrome,
Apert syndrome and MEN2B demonstrate germline mosaicism as a
result of selection in male germline stem cells that explains an
association of these conditions with advanced paternal age [15].
There is also a growing body of evidence that advanced paternal
age is associated with an increased risk for complex disorders such
as certain congenital anomalies, schizophrenia, and autism spec-
trum disorders [16,17]. For autosomes and sex chromosomes, there
is no strong evidence that aneuploidy is signicantly increased in
newborns as paternal age increases but two possible exceptions are
trisomy 21 and Klinefelter syndrome as recent data suggest a
paternal effect, either acting alone or in combination with a
maternal age effect [18].
Therefore, genetic risk stratication based on family history,
ethnicity and parental age is paramount for preconception
counseling. Increased awareness of the importance of using family
history as a screening tool, and of the value of preventive measures
and increased surveillance, can improve the outcomes. The avail-
ability of advanced technology allows for the identication of ge-
netic etiologies in many conditions for which specic diagnostic
testing is currently available.
3. Preimplantation genetic testing
Preimplantation genetic testing comprises all types of genetic
testing performed on the embryos obtained from an IVF cycle and
was rst described by Handyside et al. [19] when the sex of the
embryo was determined in two cases with a history X-linked ge-
netic disorders. Preimplantation genetic testing is divided into two
categories: preimplantation genetic screening (PGS) and preim-
plantation genetic diagnosis (PGD).
PGS is performed on the embryos obtained from the parents
with presumed normal karyotypes. Chromosomal aneuploidy is a
major factor in implantation failure and spontaneous abortions. It
was demonstrated that half of the embryos produced in vitro had
chromosomal abnormalities that signicantly decreased the im-
plantation rate [20]. PGS is used as an embryo screening method for
aneuploidy, as the morphological analysis is not reliable for aneu-
ploidy prediction. The most common method used for PGS is
analysis of the embryonic cells on day 3 after fertilization with
uorescent in-situ hybridization (FISH). Fast turnaround times and
high accuracy are the advantages of PGS with FISH but the limited
number of chromosomes that can be evaluated is a signicant
limitation. Prospective trials of PGS by FISH have not demonstrated
any improvement in pregnancy rates [21]. Comprehensive aneu-
ploidy screening using whole-genome array showed that aneu-
ploidy may occur in any of the 23 pairs of chromosomes. Recent
retrospective studies showed that PGS using genome-wide
approach and testing for all 23 pairs of chromosomes could
improve the pregnancy outcome in certain groups or patients.
Specically, in women aged >35 years with a history of recurrent
pregnancy loss, PGS is associated with reduced rst-trimester
spontaneous abortion rate [22]. However, prospective random-
ized trials have not demonstrated a denitive benet of PGS and
therefore, currently, PGS is not recommended for routine use.
PGD involves testing the embryos for a specic genetic disorder.
PGD requires prior identication of the genetic cause of the
N. Babkina, J.M. Graham Jr. / Seminars in Fetal & Neonatal Medicine 19 (2014) 214e219 215
condition in the family. Multiple conditions can be tested for,
including cystic brosis [23], Huntington disease [24], achondro-
plasia [25] and many others. PGDwas initially developed to identify
the embryos that carried genes for serious, childhood-onset dis-
eases. Recently, the American Society for Reproductive Medicine
reported that PGD for adult-onset conditions is ethically justiable
when the condition is serious and there is no known intervention
for the condition, or the available interventions are either inade-
quately effective or signicantly burdensome, thus justifying
testing for Huntington disease, Alzheimer disease, and cancer
susceptibility genes (e.g. BRCA1 and -2) [26].
Depending on the nature of the condition, different methods
can be used, such as FISH and chromosomal microarrays for CNVs
and DNA sequencing for single-gene disorders with known muta-
tion in the parent. Most commonly, the biopsy of the embryo is
performed on day 3 or day 5 after fertilization. Blastomere stage
biopsy on day 3 is associated with higher risk of damaging the
embryo and misdiagnosis secondary to possible mosaicism [27],
whereas trophectoderm biopsy performed at the blasctocyst stage
on day 5 is associated with better results and more accurate diag-
nosis [22,27]. Although the risk of mosaicism at blastocyst stage is
lower than at the blastomere stage, there still might be a discrep-
ancy between results in the trophectoderm that further develops
into placental tissue and the inner cell mass that becomes a fetus
(e.g. conned placental mosaicism). Therefore, even with the
availability of advanced techniques, the risk of embryo misdiag-
nosis exists, and it is currently recommended to offer chorionic
villus sampling (CVS) or amniocentesis to rule out genetic diagnosis
in the fetus, especially when there is a proven family history of a
genetic condition.
4. Standard karyotype and chromosomal microarray
Routine microscopic analysis of chromosomes (standard kar-
yotype) has been used for more than 50 years to detect large
chromosomal abnormalities such as translocations, deletions or
duplications, but it cannot identify genomic imbalances
<5 Mb. Submicroscopic deletions and duplications are identied in
15e20% of cases of autism and intellectual disability, thus chro-
mosomal microarray became a rst-line diagnostic test for the
evaluation of congenital anomalies and intellectual disability
postnatally [9]. Genomic microarrays provide a genome-wide
screen for genomic imbalances at a high resolution, allowing for
the detection of all known recurrent microdeletion and micro-
duplication syndromes when using resolution of 400 kb for older
arrays versus 7.7 kb for single nucleotide polymorphism (SNP) ar-
rays. Current microarray technology also allows genome-wide
detection of microscopic and submicroscopic copy number
changes <100 kb. The use of DNA-based microarrays eliminates the
need to use cell culture, decreases overall turnout time, and in
general is less labor intensive [28,29].
Microarray technology is based on hybridization of labeled pa-
tient DNA with specic probes with known genomic coordinates.
The measurement of signal intensity ratio of patient DNA to refer-
ence DNA from individual probes allows identication of gains or
losses of chromosomal material. Proof of principle for the use of
microarray in prenatal diagnosis was rst described by Rickman
et al. [28] DNA extracted from uncultured amniotic uid cells was
used for comparative genomic hybridization (CGH). Fetal DNA from
CVS specimens has also been shown to be suitable for performing
CGH. The feasibility of microarray analysis for prenatal detection of
CNVs was demonstrated using different oligonucleotide array CGH
platforms, as well as SNP arrays [30,31]. SNP arrays have additional
oligonucleotide probes in order to provide more even coverage of
the genome and improved detection of CNVs. The additional probes
for SNP array also allow genotyping based on allele frequency and
detection of copy-neutral runs of homozygosity, chromosomal
mosaicism, and uniparental disomy. Although SNP arrays can
detect uniparental isodisomy, parental parenatal samples are
required for the detection of uniparental heterodisomy [31]. By
design the microarrays can be further subdivided into targeted and
whole-genome arrays. Targeted arrays have the most dense probe
coverage in regions of the genome associated with known micro-
deletion/duplications syndromes or genes known to have clinical
implications for human disease. Whole-genome arrays have evenly
distributed probe coverage and a higher overall resolution.
A number of prospective and retrospective studies testing the
efcacy of microarray technique in detection of clinically signicant
chromosomal gains or losses during prenatal diagnosis have been
performed. The additional diagnostic yield for clinically signicant
CNVs has ranged from 1.3% to 5% [29,30,32]. Most of the studies are
small and include women at risk of having a fetus with chromo-
somal abnormalities based on evidence of structural defects
detected on ultrasound. The largest prospective study includes
4406 women who underwent prenatal diagnosis secondary to
advanced maternal age, abnormal screening results or abnormal
ndings on prenatal ultrasound. In samples with a normal karyo-
type, microarray analysis revealed clinically relevant deletions or
duplications in 6% with a structural anomaly and 1.7% of those
whose indications were advanced maternal age or positive
screening results [33].
However, microarrays provide no information about the struc-
ture of the genome, so structural rearrangements that are copy
number neutral, such as balanced translocations or inversions,
cannot be detected by array and require the use of other techniques.
Although inability to detect balanced rearrangements with array is
considered a limitation, truly balanced translocations without gain
or loss of material are less likely to be associated with a clinically
signicant phenotype in a neonate. However, apparently balanced
rearrangements identied by chromosomal analysis are associated
with a 6.7% risk of abnormal phenotype in a neonate that are
usually caused by a genomic gain or loss at breakpoints, that are not
detected with standard cytogenetic analysis and may be due to
cryptic genomic imbalances or to the disruption of genes at the
breakpoint. Array is useful not only in identifying the genomic
imbalances at the breakpoints, but also in detecting unexpectedly
complex rearrangements in other chromosomes. Schluth-Bolard
et al. used next-generation sequencing to locate breakpoints at
the molecular level in four patients with multiple congenital
anomalies and/or intellectual deciency, who were carrying
apparently balanced chromosomal rearrangements and genomic
imbalance excluded by microarray [5]. Thus, array CGH combined
with next-generation sequencing is a powerful tool allowing rapid
breakpoint cloning of apparently balanced chromosomal rear-
rangements at the molecular level. Systematic studies of apparently
balanced chromosomal rearrangements with abnormal phenotype
by array CGH have showed that the phenotype might occur due to
genomic imbalances near or far from breakpoints. A recent meta-
analysis estimated that 37% of two-breakpoint rearrangements
are unbalanced [34]. Polyploidies such as triploidy or tetrapolidy
cannot be detected by array CGH secondary to correction of DNA
concentration; however, additional allelic information provided in
SNP arrays improves polyploidy detection. High-level mosaicism
can be diagnosed with arrays, but detection of low-level mosaicism
is a potential limitation of microarrays. The capacity of array to
detect low-level mosaicism depends on standard deviation and
dynamic range. Several reports from the studies focused on post-
natal evaluation suggested that array was comparable to conven-
tional cytogenetic analysis in the detection rate of chromosomal
mosaicism.
N. Babkina, J.M. Graham Jr. / Seminars in Fetal & Neonatal Medicine 19 (2014) 214e219 216
Concern for the potential risk for detection of the variants of
unknown clinical signicance that might result in parental anxiety
is one of the limiting factors for the use of array technology pre-
natally. The reported incidence of variants of unknown signicance
varies from 0.5% to 2% depending on the array design and the
availability of parental samples for comparison. In the studies that
enrolled the fetuses with major malformations detected via pre-
natal ultrasound, the rate of the variants of unknown signicance
was reported as high as 12.2% [35]. A large prospective study that
included women with advanced maternal age, abnormal screening
results and/or abnormal ndings on ultrasound reported an inci-
dence of variants of unknown clinical signicance of 3.4% among
karyotypically normal cases [33]. However, with additional expe-
rience acquired over more recent years, only 1.5% cases remained in
the category of variants of unknown clinical signicance.
5. Non-invasive prenatal screening
Non-invasive prenatal screening using fetal cell-free DNA from
maternal blood became available with advancement of genomic
technology and development of next-generation sequencing. The
phenomenon of fetus-derived cell-free DNA circulating in a preg-
nant womans blood was rst described in 1997 [2]. At present, it is
widely accepted that apoptosis of the trophoblastic cells is the
primary source of cell-free fetal DNA, which explains the high
turnover of circulating DNA and its rapid clearance after delivery. In
maternal serum up to 10% of DNA is of fetal origin [36]. Fetal cell-
free DNA is present in maternal circulation in a quantity sufcient
for testing from10 weeks of gestation. Next-generation sequencing
technology allows massive parallel sequencing of millions of
amplied genetic fragments, and by using sophisticated bioinfor-
matics analysis, it is possible to isolate fetal sequences from
maternal sequence and to determine the dosage difference be-
tween fetal and reference sequences.
Several studies demonstrated high sensitivity, high specicity,
and low false-positive rates for detection of trisomy 13, 18, and 21
[37]. The detection rate of sex chromosome abnormalities is also
high [38]. Overall, the detection rates are substantially higher than
with maternal serum analytes, and are accompanied by a much
lower (<1% false-positive rate) [39]. Recently, it has been shown
that fetal RhD genotyping can be very accurately determined in all
three trimesters using circulating cell-free fetal DNA in the
maternal circulation [40]. Diagnosis of single-gene disorders is
being validated for translation into clinical practice. In pregnancies
at risk for an autosomal dominant disorder of paternal origin, the
study of the paternal mutation or alleles in the maternal plasma
could detect the fetal condition with regard to the pathology.
Huntington disease [41], achondroplasia [42], and myotonic dys-
trophy [43] are autosomal dominant conditions that have been
diagnosed through non-invasive prenatal testing. In couples with
different mutations for a recessive disorder, the absence of the
paternal mutation or detection of the normal paternal allele in
maternal plasma can rule out the possibility of the fetus inheriting
the disease. Detection of the paternal mutation increases the risk to
the fetus from 1 in 4 to 1 in 2. Recent advances in the development
of new technologies have detected both paternal and maternal
mutations, and a genome-wide genetic map of the fetus has been
constructed for the paternal and maternal haplotypes. Haplotype
analysis in conjunction with targeted next-generation sequencing
allowed non-invasive prenatal diagnosis for alpha- and beta-
thalassemia by detecting paternally inherited alleles in the
maternal plasma [44]. Other autosomal recessive disorders suc-
cessfully diagnosed with non-invasive prenatal testing have
included cystic brosis [45], propionic acidemia [46], congenital
adrenal hyperplasia [47], and Leber congenital amaurosis [48].
Although the studies are promising, there are certain limitations
for non-invasive prenatal screening. Sequence information identi-
ed through non-invasive prenatal screening is derived from the
placenta, which raises the possibility of a false-positive result due to
conned placental mosaicism, and similar to CVS, placental DNA
may not reect the true fetal DNA. The majority of studies on
non-invasive prenatal screening include women aged >35 years
who are at higher risk for aneuploidy. The data concerning non-
invasive prenatal screening in low-risk populations are limited,
but a decreased incidence of aneuploidy may be associated with
lower positive predictive value. Also, prior knowledge about the
prevalence of aneuploidies in the samples may have affected an
analysts decisions about how to classify ambiguous test results.
Current technology does not allow detection of chromosomal rear-
rangements and has not been validated for the majority of single-
gene disorders and CNVs. In cases of abnormal ndings detected
via ultrasound, invasive testing may still be necessary. Non-invasive
prenatal testing does not substitute for routine rst- and second-
trimester screening, as it does not detect neural tube defects and
has limitations in evaluation of structural abnormalities detected via
ultrasound and in multiple gestations. As the clinical utility of cell-
free DNA testing in the general population is unproven,
professional organizations such as the American College of Obste-
tricians and Gynecologists, the Society for MaternaleFetal Medicine
and the National Society of Genetic Counselors currently recom-
mend cell-free DNA testing only for high-risk pregnancies and they
suggest that positive results be conrmed through invasive testing.
6. Whole-exome and whole-genome sequencing
Next-generation sequencing has been used to identify causative
genes in many Mendelian disorders. About 85% of known disease-
causing mutations occur within the 1% of the genome encoding for
proteins. Therefore, whole exome sequencing represents a power-
ful method to cost-effectively capture the majority of protein-
coding regions to identify single nucleotide variants (SNVs), and
small insertion/deletions. It is now possible to non-invasively
identify the fetal genome with the shotgun sequencing of
maternal plasma DNA. Also, molecular counting of parental hap-
lotypes and fetal exome capture has allowed exome screening of
clinically relevant and deleterious alleles that were paternally
inherited or were de novo [49]. There are data concerning the
clinical diagnosis established by whole-genome sequencing of a
prenatal sample e for example, a 13-day sequence and analysis
pipeline detection of a disruption of CHD7 in a fetus with a sus-
pected diagnosis of CHARGE syndrome who had multiple anoma-
lies detected on ultrasound [50].
However, although comprehensive non-invasive whole-exome
or whole-genome sequencing has signicant potential, it also
brings multiple concerns such as technical complexity, prolonged
turnaround time, difculty with bioinformatic interpretation, and
ethical issues that limit its clinical use. Whole-exome or whole-
genome sequencing is still considered impractical for routine pre-
natal care secondary to cost and turnaround time. Eventually,
decreasing costs may allow personal genome sequencing
to become available for routine analysis, but annotation and
interpretation of variant information are essential to provide in-
formation that can be used to better manage an individuals con-
dition. The expanded use of non-invasive fetal exome or genome
sequencing would present several advantages and challenges.
Broader use might lead to the improved detection of Mendelian
disorders in families who would not otherwise have been offered
prenatal testing, as well as in families who might have refused
invasive testing because of risk of fetal loss. However, the concern is
that it will identify variants that are beyond the scope of
N. Babkina, J.M. Graham Jr. / Seminars in Fetal & Neonatal Medicine 19 (2014) 214e219 217
conventional prenatal screening and diagnosis, specically variants
that indicate increased risk for developing adult-onset conditions.
The relevance of such information is controversial for a current
pregnancy, but it may have signicant implications for the health of
the parent. Therefore, the broader implementation of non-invasive
fetal genome sequencing will lead to wider application of the re-
sults and increased availability.
7. Conclusions
Determining accurate genetic diagnosis prenatally permits
families to make informed reproductive decisions. Therefore, it is
important for the provider to be aware of the spectrum of potential
genetic conditions, and to use appropriate testing modalities to
obtain specic diagnosis.
References
[1] Steele MW, Breg WR. Chromosome analysis of human amniotic uid cells.
Lancet 1966;1:383e5.
[2] Lo YMD, Corbetta N, Chamberlain PF, Rai V, Sargent IL, Redman CW, et al.
Presence of fetal DNA in maternal plasma and serum. Lancet 1997;350:485e7.
[3] Yoon P, Scheuner M. The family history public health initiative. In: Centers for
Disease Control and Prevention. Genomics and population health: United
States 2003. Atlanta: CDC; 2004. 39e45.
[4] Frezzo TM, Rubinstein WS, Dunham D, Ormond KE. The genetic family history
as a risk assessment tool in internal medicine. Genet Med 2003;5:84e91.
[5] Schluth-Bolard C, Labalme A, Cordier MP. Breakpoint mapping by next gen-
eration sequencing reveals causative gene disruption in patients carrying
apparently balanced chromosome rearrangements with intellectual de-
ciency and/or congenital malformations. J Med Genet 2013;50:144e50.
[6] Brent RL. Environmental causes of human malformations: the pediatricians
role in dealing with these complex clinical problems caused by a multiplicity
of environmental and genetic factors. Pediatrics 2004;113:957e68.
[7] Moeschler JB, Shevell M. American Academy of Pediatrics Committee on Ge-
netics: clinical genetics evaluation of the child with mental retardation or
developmental delay. Pediatrics 2006;117:2304e16.
[8] Schoumans J, Ruivenkamp C, Holmberg E, Kyllerman M, Anderlid BM,
Nordenskjold M. Detection of chromosomal imbalances in children with
idiopathic mental retardation by array based comparative genomic hybridi-
zation (array-CGH). J Med Genet 2005;42:699e705.
[9] Poot M, Hochstenbach R. A three-step workow procedure for the interpre-
tation of array-based comparative genome hybridization results in patients
with idiopathic mental retardation and congenital anomalies. Genet Med
2010;12:478e85.
[10] Crow JF. The origins, patterns and implications of human spontaneous mu-
tation. Nat Rev Genet 2000;1:40e7.
[11] Glaser RL, Jabs EW. Dear old dad. Sci Aging Knowledge Environ 2004;(3). re1.
[12] Friedman JM. Genetic disease in the offspring of older fathers. Obstet Gynecol
1981;57:745e9.
[13] Kong A, Frigge ML, Masson G, Besenbacher S, Sulem P, Magnusson G, et al.
Rate of de novo mutations and the importance of fathers age to disease risk.
Nature 2012;488(7412):471e5.
[14] Jung A, Schuppe HC, Schill WB. Are children of older fathers at risk for genetic
disorders? Andrologia 2003;35:191e9.
[15] Goriely A, McVean GA, Rjmyr M, Ingemarsson B, Wilkie AO. Evidence for
selective advantage of pathogenic FGFR2 mutations in the male germ line.
Science 2003;301(5633):643e6.
[16] Byrne M, Agerbo E, Ewald H, Eaton WW, Mortensen PB. Parental age and
risk of schizophrenia: a caseecontrol study. Arch Gen Psychiatry 2003;60:
673e8.
[17] Cantor RM, Yoon JL, Furr J, Lajonchere CM. Paternal age and autism are
associated in a family-based sample. Mol Psychiatry 2007;12:419e21.
[18] Shi Q, Martin RH. Aneuploidy in human sperm: a review of the frequency and
distribution of aneuploidy, effects of donor age and lifestyle factors. Cytogenet
Cell Genet 2000;90:219e26.
[19] Handyside A, Kontogianni EH, Hardy K, Winston RM. Pregnancies from bio-
psied human preimplantation embryos sexed by Y-specic DNA amplication.
Nature 1990;344:768e70.
[20] Wilton L. Preimplantation genetic diagnosis for aneuploidy screening in early
human embryos: a review. Prenat Diagn 2002;22:512e8.
[21] Mastenbroek S, Twisk M, van der Veen F, Repping S. Preimplantation genetic
screening: a systematic review and meta-analysis of RCTs. Hum Reprod Up-
date 2011;17:454e66.
[22] Forman EJ, Tao X, Ferry KM, Taylor D, Treff NR, Scott Jr RT. Single embryo
transfer with comprehensive chromosome screening results in improved
ongoing pregnancy rates and decreased miscarriage rates. Hum Reprod
2012;27:1217e22.
[23] Verlinsky Y, Rechitsky S, Evsikov S, White M, Cieslak J, Lifchez A, et al. Pre-
conception and preimplantation diagnosis for cystic brosis. Prenat Diagn
1992;12:103e10.
[24] Van Rij MC, De Rademaeker M, Moutou C, Dreesen JC, De Rycke M, Liebaers I,
et al. Preimplantation genetic diagnosis (PGD) for Huntingtons disease: the
experience of three European centres. Eur J Hum Genet 2012;20:368e75.
[25] Rechitsky S, Verlinsky O, Amet T, Rechitsky M, Kouliev T, Strom C. Reliability
of preimplantation diagnosis for single gene disorders. Mol Cell Endocrinol
2001;183:S65e8.
[26] Ethics Committee of the American Society for Reproductive Medicine. Use of
preimplantation genetic diagnosis for serious adult onset conditions: a com-
mittee opinion. Fertil Steril 2013;100:54e7.
[27] Harton GL, Magli MC, Lundin K, Montag M, Lemmen J, Harper JC. European
Society of Human Reproduction and Embryology (ESHRE) PGD Consortium/
Embryology Special Interest Groupdbest practice guidelines for polar body
and embryo biopsy for preimplantation genetic diagnosis/screening (PGD/
PGS). Hum Reprod 2011;26:41e6.
[28] Rickman L, Fiegler H, Shaw-Smith C, Nash R, Cirigliano V, Voglino G, et al.
Prenatal detection of unbalanced chromosomal rearrangements by array CGH.
J Med Genet 2006;43:353e61.
[29] Van den Veyver IB, Patel A, Shaw CA, Pursley AN, Kang SH, Simovich MJ, et al.
Clinical use of array comparative genomic hybridization (aCGH) for prenatal
diagnosis in 300 cases. Prenat Diagn 2009;29:29e39.
[30] Shaffer LG, Coppinger J, Alliman S, Torchia BA, Theisen A, Ballif BC, et al.
Comparison of microarray-based detection rates for cytogenetic abnormalities
in prenatal and neonatal specimens. Prenat Diagn 2008;28:789e95.
[31] Faas BH, van der Burgt I, Kooper AJ, Pfundt R, Hehir-Kwa JY, Smits AP, et al.
Identication of clinically signicant, submicroscopic chromosome alterations
and UPD in fetuses with ultrasound anomalies using genome-wide 250 k SNP
array analysis. J Med Genet 2010;47:586e94.
[32] Park SJ, Jung EH, Ryu RS, Kang HW, Ko JM, Kim HJ, et al. Clinical imple-
mentation of whole genome array CGH as a rst-tier test in 5080 pre and
postnatal cases. Mol Cytogenet 2011;4:12.
[33] Wapner RJ, Martin CL, Levy B, Ballif BC, Eng CM, Zachary JM, et al. Chromo-
somal microarray versus karyotyping for prenatal diagnosis. N Engl J Med
2012;367:2175e84.
[34] Feenstra I, Hanemaaijer N, Sikkema-Raddatz B, Yntema H, Dijkhuizen T,
Lugtenberg D, et al. Balanced into array: genome-wide array analysis in 54
patients with an apparently balanced de novo chromosome rearrangement
and a meta-analysis. Eur J Hum Genet 2011;19:1152e60.
[35] DAmours G, Kibar Z, Mathonnet G, Fetni R, Tihy F, Desilets V, et al. Whole-
genome array CGH identies pathogenic copy number variations in fetuses
with major malformations and a normal karyotype. Clin Genet 2012;81:
128e41.
[36] Lun FM, Chiu RW, Allen Chan KC, Yeung Leung T, Kin Lau T, Lo YMD. Micro-
uidics digital PCR reveals a higher than expected fraction of fetal DNA in
maternal plasma. Clin Chem 2008;54:1664e72.
[37] Palomaki GE, Kloza EM, Lambert-Messerlian GM, Haddow JE, Neveux LM,
Ehrich M, et al. DNA sequencing of maternal plasma to detect Down
syndrome: an international clinical validation study. Genet Med 2011;13:
913e20.
[38] Finning KM, Chitty LS. Non-invasive fetal sex determination: impact on clin-
ical practice. Semin Fetal Neonatal Med 2011;13:69e75.
[39] Simpson JL. Cell-free fetal DNA and maternal serum analytes for monitoring
embryonic and fetal status. Fertil Steril 2013;99:1124e34.
[40] Moise Jr KJ, Boring NH, OShaughnessy R, Simpson LL, Wolfe HM, Baxter JK,
et al. Circulating cell-free fetal DNA for the detection of RHD status and sex
using reex fetal identiers. Prenat Diagn 2013;33:95e101.
[41] Gonzalez-Gonzalez MC, Garcia-Hoyos M, Trujillo-Tiebas MJ, Bustamante
Aragones A, Rodriguez de Alba M, Diego Alvarez D, et al. Improvement in
strategies for the non-invasive prenatal diagnosis of Huntington disease.
J Assist Reprod Genet 2008;25:477e81.
Practice points
The birth of a child with intellectual disabilities and/or
congenital anomalies is an upsetting event for any family,
so if a child is born with these problems, think ahead to
the next pregnancy and try to ascertain the cause before
the next pregnancy.
It is best to perform SNP chromosomal microarray before
attempting to sequence any gene(s) because many gene
sequencing techniques do not detect CNVs.
Exome sequencing and whole genome sequencing are
time consuming (at least 6 months), so if this needs to be
done to obtain a diagnosis that will provide accurate
prenatal diagnosis, allow sufcient time.
N. Babkina, J.M. Graham Jr. / Seminars in Fetal & Neonatal Medicine 19 (2014) 214e219 218
[42] Lim JH, Kim MJ, Kim SY, Kim HO, Song MJ, Kim MH, et al. Non-invasive pre-
natal detection of achondroplasia using circulating fetal DNA in maternal
plasma. J Assist Reprod Genet 2010;28:167e72.
[43] Amicucci P, Gennarelli M, Novelli G, Dallapiccola B. Prenatal diagnosis of
myotonic dystrophy using fetal DNA obtained from maternal plasma. Clin
Chem 2009;46:301e2.
[44] Papasavva T, van Ijcken WF, Kockx CE, Van den Hout MC, Kountouris P,
Kythreotis L, et al. Next generation sequencing of SNPs for non-invasive
prenatal diagnosis: challenges and feasibility as illustrated by an application
to b-thalassaemia. Eur J Hum Genet 2013;21:1403e10.
[45] Bustamante-Aragones A, Gallego-Merlo J, Trujillo-Tiebas MJ, de Alba MR,
Ugarte M, Ramos C. New strategy for the prenatal detection/exclusion of
paternal cystic brosis mutations in maternal plasma. J Cyst Fibros 2008;7:
505e10.
[46] Bustamante-Aragones A, Prez-Cerd C, Prez B, de Alba MR, Ugarte M,
Ramos C. Prenatal diagnosis in maternal plasma of a fetal mutation causing
propionic acidemia. Mol Genet Metab 2008;95:101e3.
[47] Chiu RW, Lau TK, Cheung PT, Gong ZQ, Leung TN, Lo YM. Non invasive pre-
natal exclusion of congenital adrenal hyperplasia by maternal plasma anal-
ysis: a feasibility study. Clin Chem 2002;48:778e80.
[48] Bustamante-Aragones A, Vallespin E, Rodriguez de Alba M, Trujillo-Tiebas MJ,
Gonzalez-Gonzalez C, Diego-Alvarez D, et al. Early noninvasive prenatal
detection of a fetal CRB1 mutation causing Leber congenital amaurosis. Mol
Vis 2008;14:1388e94.
[49] Fan C, Gu W, Wang J, Blumenfeld YJ, El-Sayed YY, Quake SR, et al. Noninvasive
prenatal measurement of the fetal genome. Nature 2012;487(7407):320e4.
[50] Talkowski ME, Ordulu Z, Pillalamarri V, Benson CB, Blumenthal I, Connolly S,
et al. Clinical diagnosis by whole-genome sequencing of a prenatal sample.
N Engl J Med 2012;367:2226e32.
N. Babkina, J.M. Graham Jr. / Seminars in Fetal & Neonatal Medicine 19 (2014) 214e219 219