|
It is important to emphasize that random fetal anomalies
are remarkably common and represent a high background rate against
which to compare any teratogenic effects specific to psychotropic
agents. The baseline incidence of major congenital malformations
in newborns in the US is approximately 2% (59). Basic formation
of major organ systems takes place early in pregnancy and is virtually
complete within the first 12 weeks after conception. However, pregnancy
is often not diagnosed for 6 to 8 weeks, during which time critical
steps in major organ development have already occurred. Teratogens
are agents, including drugs, that interfere with this process to
produce malformations of varying severity. Each organ system appears
to be vulnerable to teratogenic effects during relatively specific
and limited time periods during the first trimester (60).
It is important to point out that dating the age of the embryo
differs from gestational dating by 2 weeks, in that gestation is
dated by clinical convention from the last menstrual period [LMP
= first day of menses], whereas embryonic age is dated from conception.
Since the date of conception can be difficult to determine, gestational
dating is preferred clinically. For example, formation of the heart
and great vessels takes place from 5 to 10 weeks from last menstrual
period, equivalent to embryonic ages 3 to 8 weeks, and formation
of the lips and palate is typically complete by gestational weeks
8 to 14 (embryonic ages 6 to 12 weeks). Folding and closure of the
neural-tube to form the brain and spinal cord occur within the first
5 to 6 weeks of gestation, or as early as 3 to 4 weeks of embryonic
age—often well before pregnancy has been diagnosed. Exposure
to a toxic agent before 2 weeks of gestation, or within the first
week after conception usually results in a nonviable, blighted conceptus
(60).
Since the early 1970s, there has been concern about an association
between prenatal exposure to lithium and risk for major congenital
anomalies. Reports from an early International Register of Lithium
Babies, based on a voluntary physician-reporting system, describe
an excess of cardiovascular malformations, and particularly Ebstein’s
anomaly, in lithium-exposed newborns (61,63,64). Ebstein’s
anomaly is characterized by right ventricular hypoplasia and downward
displacement of the tricuspid valve, often with varying septal defects.
The risk for this malformation in infants with first-trimester lithium
exposure was initially proposed to be 400 times higher than the
background baseline rate of about 1/20 000 live births found in
the general population (11,61,63,64). However, despite the fact
that the reliability of this initial estimate was highly suspect
in view of almost certain selective reporting of adverse outcomes
to such registries, this risk estimate influenced clinical practice
for the next 2 decades.
More recent, controlled epidemiologic studies suggest a real, but
more modest, teratogenic risk of Ebstein’s anomaly following
first-trimester lithium exposure (11,62,65–70). Based on a
pooled analysis of the data, Cohen and others estimated the risk
for Ebstein’s anomaly following first-trimester exposure to
be between 1/1000 (0.1%) and 1/2000 (0.05%) births (11). Based on
relatively well-designed studies, rates of other congenital cardiovascular
defects among lithium-exposed infants have varied from 0.9% to 12%
(11,67,69). Although the estimated risk of Ebstein’s anomaly
in lithium-exposed infants is 10 to 20 times higher than in the
general population, the absolute risk is small (0.05% to 0.1%),
and lithium arguably remains the safest mood stabilizer for use
during pregnancy. Nevertheless, the FDA fetal risk rating for lithium
is D. Prenatal screening with a high-resolution ultrasound and fetal
echocardiography is recommended at or about weeks 16 to 18 of gestation
to screen for cardiac anomalies (11,62,65,70).
While reintroduction of lithium after the first trimester is not
associated with increased risk for major malformations, additional
risks from exposure later in pregnancy include reports of neonatal
toxicity in offspring exposed to lithium during labour and delivery.
These include several cases of muscular hypotonia with impaired
breathing and cyanosis, often referred to as “floppy baby”
syndrome (62,65,70–72). Isolated cases of neonatal hypothyroidism,
nephrogenic diabetes insipidus, and polyhydramnios have also been
described (70,72).
Based on these case reports of toxicity in infants born to lithium-treated
mothers, some authors have recommended discontinuing lithium several
days or weeks prior to delivery to minimize the risk of neonatal
toxicity (11,65,70,73,74). However, there is a low incidence of
neonatal toxicity with lithium exposure, and this practice carries
significant risk, since it withdraws treatment from patients precisely
as they are about to enter the postpartum period. A recent naturalistic
survey found no direct evidence of neonatal toxicity in newborns
whose mothers received lithium either during pregnancy or during
labour and delivery (46).
Limited information is available regarding behavioural outcomes
of children exposed to lithium in utero, but a 5-year follow-up
of 60 children exposed to lithium during the second and third trimesters
of pregnancy found no evidence of significant behavioural problems
(74). A preliminary report of 13 children (average age 3.5 years)
of women with BD who had been exposed to lithium in utero and 11
children (average age 3.3 years) of women with BD not exposed to
medication in utero found no significant differences in neurobehavioural
outcome, using blinded and well-validated neurocognitive assessments
(75). The small sample, however, precludes conclusions about lithium
exposure and long-term neurobehavioural sequelae.
Compared with lithium, anticonvulsants may pose a much more serious
teratogenic risk. All commonly used older anticonvulsants have been
associated with teratogenicity, and the risk for major birth defects
in infants born to women receiving anticonvulsants is 2 times greater
than that in the general population (76). Although most information
about the reproductive safety of anticonvulsants derives from patients
with epilepsy rather than BD, recent findings suggest that exposure
to certain anticonvulsants, rather than the presence of a seizure
disorder, is the relevant variable (77). Fetal exposure to anticonvulsants
has been associated not only with relatively high rates of neural
tube defects (NTDs), such as spina bifida, but also with multiple
anomalies, including craniofacial abnormalities (also known as the
“anticonvulsant face”), congenital heart disease, cleft
lip or palate, growth retardation, and microcephaly (76–79).
Factors that may increase the risk for teratogenisis include high
maternal serum anticonvulsant levels and exposure to more than a
single anticonvulsant (77,82–84).
|
|
The lowest effective dosage should be used, and given in frequent
divided doses over the course of the day (85). Anticonvulsant levels
should be monitored closely, with the dosage adjusted appropriately
(82). Prenatal screening for congenital malformations (including
NTDs and cardiac anomalies), using fetal ultrasound at 18 to 22
weeks of gestation, is recommended (79,84,85). The possibility of
fetal NTDs should be evaluated with maternal serum alphafetalprotein
(MSAFP) and ultrasonography. In addition, 4 mg daily of folic acid
is recommended before conception and in the first trimester for
women receiving anticonvulsants, even though it is unknown whether
supplemental folic acid can attentuate the risk of NTDs in the setting
of anticonvulsant exposure (79,84,85).
First-trimester exposure to carbamazepine is associated with risk
of NTDs estimated to be about 1.0% (86). Infants exposed to carbamazepine
prenatally are also at increased risk for craniofacial abnormalities,
microcephaly, and growth retardation (77,86). It is not known whether
the new derivative, oxcarbazepine, is associated with similar fetal
risks (88). However, among the anticonvulsants used to treat BD,
valproic acid and its various derivatives and preparations, including
divalproex, may be even more serious teratogens, with rates of NTDs
in the range of 1.0% to 5.0%, or about a two- to tenfold increase
in risk above the general-population base rates of about 0.5% (89,90).
These risks are of particular concern because formation of the neural
tube occurs within the first month of gestation, often before the
pregnancy has been diagnosed. Prenatal exposure to valproate has
also been associated with characteristic craniofacial abnormalities,
cardiovascular malformations, limb defects, and genital anomalies,
as well as other central nervous system (CNS) structural abnormalities,
including hydrocephalus (76–78,89,90).
Information about possible untoward neurobehavioural effects of
anticonvulsant exposure is very limited. There is no evidence to
suggest increased risk for mental retardation following antenatal
exposure to anticonvulsants, but subtle cognitive effects have been
suggested, including after
second-, or even late third-trimester, drug exposure (91–94).
These subtle deficits may be correlated with the presence of midface
hypoplasia (93).
Information about the reproductive safety of newer anticonvulsants
sometimes used to treat BD—including lamotrigine, gabapentin,
oxcarbazepine, and topiramate—remains very sparse (95). Most
of the available information is limited to a few case reports pertaining
to such drugs, given alone or often in combination with other anticonvulsants,
and almost always to pregnant women with epilepsy. A pregnancy registry
was established recently by the manufacturer of lamotrigine, with
a preliminary suggestion that the risk of all malformations following
prenatal exposure to lamotrigine monotherapy during the first trimester
averaged 2.5% (96). Data from a UK registry suggest a similar risk
with exposure to lamotrigine alone early in pregnancy (97). These
registries have not shown a consistent excess of any specific form
of birth defect, but the numbers of pregnancies accumulated so far
remain small. Other efforts are under way to accumulate unbiased
information regarding teratogenic risks across a broad range of
anticonvulsants in pregnancies enrolled prospectively. For example,
the North American Antiepileptic Drug Pregnancy Registry was recently
established as a way of collecting such information rapidly and
efficiently (its toll-free telephone number is 888-233-2334). The
registry will release its findings only after information on neonatal
outcome has been collected from at least 300 monotherapy exposures.
It is estimated that this number will provide sufficient statistical
power to detect at least a twofold excess of major birth defects.
At this time, given the sparse data on the fetal safety of the newer
anticonvulsants proposed for use in BD, it is difficult to justify
their use as first-line agents during early pregnancy.
Switching from a prolactin-elevating antipsychotic agent, such
as risperidone or an older neuroleptic, to a modern agent without
such effects can increase the risk of becoming pregnant (98). Early
case reports described limb malformations following first-trimester
exposure to haloperidol (99,100), but several other studies have
not demonstrated teratogenic risk associated with any of the older
typical neuroleptics of either low or high potency (101,102). Nevertheless,
a metaanalysis of available studies noted a suggestive elevation
of overall risk for congenital malformations following first-trimester
exposure to low-potency neuroleptics; however, no specific type
of malformation was identified (50). In clinical practice, high-potency
neuroleptic agents such fluphenazine, haloperidol, perphenazine,
and trifluoperazine are recommended because they have lesser autonomic,
sedative, and cardiovascular side effects than do the low-potency
agents. (50,52).
Information on the reproductive safety of newer ATPs remains very
sparse. There are no adequate human studies to evaluate the risk
for potential teratogenicity of clozapine, olanzapine, risperidone,
quetiapine, or ziprasidone. There are perhaps 5 published case reports
of women treated during pregnancy with the oldest ATP agent, clozapine;
they yield no evidence of major congenital malformations (103–106).
In addition, the original manufacturer of clozapine has collected
information on at least 29 babies exposed to clozapine before birth
(107). Of these, 25/28 were healthy, and 4/28 had problems, including
neonatal convulsions, Turner’s syndrome, collar-bone fracture,
facial deformity, congenital hip dislocation, and blindness. However,
the possible significance of these findings as evidence of teratogenic
actions of clozapine is not clear. The manufacturer of olanzapine
also established a registry that includes at least 96 reports of
outcomes following prenatal exposure to this ATP (95,108). There
was only 1 case of a major malformation (1/96) and 7 other instances
of temporary perinatal complications. Experience to date with all
these registries for atypical antipsychotic agents remains insufficient
to provide for adequate assessment of fetal safety.
Several case reports have documented transient extrapyramidal symptoms
(EPS), including motor restlessness, tremor, hypertonicity, dystonia,
and parkinsonism in neonates exposed to neuroleptics during pregnancy
(109,110). These problems have typically been of short duration
and have been followed by apparently normal subsequent motor development
(111). Risks for potential neurobehavioural or cognitive effects
from prenatal exposure to older neuroleptics have also been considered,
but the available data remain limited and inconclusive. A longitudinal
study that evaluated general intelligence and behaviour of children
exposed to low-potency neuroleptics in utero found no evidence of
dysfunction or developmental delays up to age 5 years (111).
|
