Antiseizure medication treatment during pregnancy

Continuing antiseizure medication treatment during pregnancy is necessary because uncontrolled maternal seizures pose a greater risk to the mother and the fetus than the use of medications. This might need to be re-emphasized as some women with epilepsy may choose to stop their antiseizure medication against medical advice on confirmation of pregnancy, fearing risk to their babies.

Approximately 65% of pregnant women with epilepsy remain equally controlled (15% actually present a reduction in seizure frequency) compared to pre-pregnancy. Thus, most patients who achieve seizure-freedom will likely remain seizure-free while pregnant if the necessary drug adjustments are made prior to conception (04). Women with epilepsy who experience seizures in the year prior to pregnancy are three to four times more likely to continue to have seizures during pregnancy than those who had been seizure-free (80).

The concentration of some antiseizure medication may change significantly during pregnancy and the puerperium, resulting in more frequent seizures or toxicity (68; 01). If the woman is on an antiseizure medication that undergoes substantial clearance changes, therapeutic drug monitoring is recommended before and during pregnancy. The individualized target concentration should be reassessed and maintained with blood levels, when applicable, throughout pregnancy.

If monitoring is not available or feasible, an increase in dose after the first trimester could be reasonably considered; this should be discussed with the patient to ensure understanding of the medical rationale. This approach is particularly useful in women who have epilepsy with generalized tonic-clonic seizures or focal to bilateral tonic-clonic seizures if their epilepsy has been sensitive to changes in antiseizure medication levels before pregnancy or if they entered pregnancy on the lowest effective dose of their antiseizure medications, provided that their antiseizure medication is known to be subject to marked changes in clearance. This is the case for lamotrigine, levetiracetam, and oxcarbazepine (72; 42).

Teratogenicity of antiseizure medication: focus on anatomical teratogenicity and major congenital malformations

The rate of major congenital malformations in offspring of healthy women is estimated as 1% to 2% of live births. The Active Malformation Surveillance Program at Brigham and Women’s Hospital in Boston estimates the background rate to be 1.6% after the exclusion of genetic and chromosomal anomalies (58). Some other registries use the higher rate of 3.2% determined by the Metropolitan Atlanta Congenital Defects Program (18), but this population­based registry uses active case identification from multiple sources, undertakes direct chart review of potential cases, and includes all malformations identified up to the age of 5 years (26).

Several pregnancy registries worldwide are collecting data on antiseizure medication­related major congenital malformations and other pregnancy-related outcomes in women with epilepsy. However, even these registries have important methodological differences in recruitment, ascertainment, inclusion and exclusion criteria, major congenital malformation classification, and follow-up, which may influence the results and prevent meaningful pooling of data. Registries that included follow-up of offspring for 1 year compared to registries that limited follow-up to 2 months after birth can identify an additional 15% in total cases of major congenital malformations (71). Moreover, exposure to antiseizure medication during pregnancy has most often been determined by the mother’s medication dosage rather than plasma level, which would more accurately reflect fetal exposure to the medication. As it remains equally important to understand dose-effect and type of antiseizure medication, information from prospective cohorts and harmonized criteria for evaluation worldwide will be ultimately meaningful.

The reported incidence of major congenital malformations in offspring of women with epilepsy varies significantly by around 20-fold, mainly because of methodological differences. Earlier studies usually relied on small numbers of recruited patients and lacked statistical power. To understand the extent of the difficulties, consider that a total of 722 antiseizure medication­exposed pregnancies is needed to identify a 7­fold increase in the rate of occurrence of a specific abnormality, such as spina bifida, with a frequency of 1 in 1000; or if drug A has a 3% risk for major congenital malformations and drug B doubles the risk to 6%, then 750 patients on monotherapy are needed in each group to reach p < 0.05 at 80% power. Moreover, the total number of pregnancies exposed to the different antiseizure medications in each registry is very discrepant; thus, percentage descriptions should also be regarded in relation to the total pregnancies for each medication. Another important observation is that parental history of major congenital malformations can be associated with an almost three times increase in the odds of major congenital malformations (71), and details on this information might vary across studies. This suggests a genetic predisposition to malformations or an individual’s susceptibility to the teratogenic effects of antiseizure medication (71).

Pregnancy outcome registries in women with epilepsy have shown that the risk of major congenital malformations (and of behavior or cognitive teratogenicity) is highest with valproic acid, a dose-dependent teratogen (69). Teratogenicity risk also increases with exposure to antiseizure medication polytherapy. This increased risk relative to monotherapy may depend more on the usage and dose of valproic acid as part of a polytherapy regimen than on the use of multiple antiseizure medications per se (78; 30). In one study that excluded valproic acid, major congenital malformations rates were found to be higher in the remaining polytherapy pregnancies compared with the monotherapy ones (6.90% vs. 3.64%; odds ratio [OR] 1.96, 95% confidence interval [CI] 1.14–3.39) (79; 42). Among antiseizure medication polytherapy without valproic acid, topiramate is associated with a dose-related increased risk of major congenital malformations (79; 09).

The teratogenicity risk of most prescribed antiseizure medications follows a (still evolving) spectrum, being considered low for levetiracetam and lamotrigine, intermediate for carbamazepine, oxcarbazepine, and zonisamide, and relatively high for topiramate, phenobarbital, and clobazam. The EURAP Epilepsy And Pregnancy Registry found an increase in malformation rates with increasing dosage at the time of conception for carbamazepine, lamotrigine, valproic acid, and phenobarbital (71).

Very little information is available for newer antiseizure medication that arrived in the North American market over the past 15 years. Understanding this risk spectrum is essential, as this guides treatment choices. However, women with focal epilepsies might not always achieve seizure control when using antiseizure medications with a more favorable teratogenicity profile; thus, the safety of women and their babies must be considered when making these decisions. Women with idiopathic generalized epilepsy might not achieve seizure control with drug alternatives to valproic acid, such as levetiracetam and lamotrigine, and there is still insufficient information on perampanel. This consideration is especially important when attempting to replace valproic acid with newer antiseizure medications in women with idiopathic generalized epilepsy.

Over the past 20 years, there have been some changes in trends for antiseizure medication prescription, with decreased use of valproic acid, phenobarbital, and carbamazepine and increased use of lamotrigine and levetiracetam (74). The EURAP International Registry, which studies risk of major congenital malformations at 1 year after birth in relation to prenatal antiseizure medication exposure, showed that the overall prevalence of major congenital malformations with monotherapies decreased from 6.0% in 2000-2005 to 4.4% in 2010-2013 (74). During the same period, there was increased awareness about folic acid and more efficient prescription and enforcement of folic acid supplementation, although routine folic acid prescription and reduction in teratogenicity rates have not been directly related in some studies (71; 42). Altogether, avoidance of antiseizure medications that are more frequently related to teratogenicity likely plays a more significant role in good pregnancy outcomes than practices of folic acid supplementation, but this remains necessary for all women with epilepsy in reproductive age.

Valproic acid. Comparatively to other antiseizure medications, intrauterine valproic acid exposure is associated with an elevated risk for many major congenital malformations, particularly for neural tube defects, including (adjusted OR): spina bifida, 12.7 (95% CI, 7.7-20.7); atrial septal defect, 2.5 (95% CI, 1.4-4.4); cleft palate, 5.2 (95% CI, 2.8-9.9); hypospadias, 4.8 (95% CI, 2.9-8.1); polydactyly, 2.2 (95% CI, 1.0-4.5); and craniosynostosis, 6.8 (95% CI, 1.8-18.8) (39).

The risk of congenital malformations is dose-related, particularly when the dosage is higher than 1000 mg/day (71). Considering that valproic acid is also definitely associated with cognitive impairment in infants exposed to this drug during pregnancy (51; 14; 15), it should altogether be avoided in women with epilepsy of childbearing age. For women with generalized epilepsies, such as juvenile myoclonic epilepsy, very few other antiseizure medications are available or effective. Alternatives are probably restricted to levetiracetam (most likely to be effective and less likely to be teratogenic but associated with mood changes that sometimes preclude continuation of therapy) and lamotrigine (safe and most often chosen as an alternative treatment but potentially associated with worsening of myoclonic seizures). Some women with idiopathic generalized epilepsy might fail to control generalized tonic-clonic seizures other than with valproic acid; in these cases, continuing this medication through childbearing age requires a joint informed decision. If a valproic acid prescription is needed, a maximum dose of 500 mg/day is recommended to lower the teratogenicity risk (27).

The International League Against Epilepsy made the following recommendations on the use of valproic acid (69; 73; 75):

	(1) Whenever possible, valproic acid should be avoided in women with epilepsy of childbearing potential.
	(2) When valproic acid is the antiseizure medication of choice for women with epilepsy of childbearing potential, because of the epilepsy syndrome or types of seizures, as well as failure to control with alternative antiseizure medication, there should be informed consent based on clear knowledge of risks (including anatomical and behavioral teratogenicity). Valproic acid prescription needs to be a shared decision between clinician and patient (and, when applicable, the patient's representatives). Discussions should include a careful risk-benefit assessment of reasonable treatment options for the patient's seizure or epilepsy type.
	(3) For seizure (or epilepsy) types where valproic acid is the most effective treatment, the risks and benefits of valproic acid and other treatment alternatives should be discussed.
	(4) Valproic acid should not be prescribed as a first-line treatment for focal epilepsy.
	(5) Valproic acid may be offered as a first-line treatment for epilepsy syndromes where it is the most effective treatment, including idiopathic generalized epilepsy associated with generalized tonic-clonic seizures.
	(6) Valproic acid may be offered as a first-line treatment in situations where pregnancy is highly unlikely (eg, significant intellectual or physical disability).
	(7) When valproic acid is prescribed to women with epilepsy, there should be regular follow-up for ongoing consideration of the most appropriate treatment regimen, and the choice of valproic acid over other alternative antiseizure medication should be revisited.

Topiramate. Topiramate teratogenicity information from pregnancy registries indicates major congenital malformations between 4.8% (monotherapy) to 11.2% (polytherapy) (33). Currently, there is evidence that topiramate-exposed pregnancies are associated with approximately 11- to 13-fold higher relative risk of oral clefts than unexposed pregnancies (33; 27).

In a French study on the association between intrauterine exposure to 10 different antiseizure medications in monotherapy during the first 2 months of pregnancy, exposure to topiramate was associated with an increased risk of cleft lip (6.8, 95% CI 1.4-20.0) among 23 major congenital malformations evaluated (09).

Upregulation of transforming growth factor-beta one (TGFβ1) expression has been identified in topiramate-treated human embryonic palate mesenchyme cells. Increased expression of SRY-box transcription factor 9 (SOX9), a TGFβ1 target that causes cleft palate when abnormally expressed, was identified in mouse embryonic palate mesenchyme (MEPM) exposed to topiramate, suggesting that abnormalities in the TGFβ1 pathway and SOX9 expression in the embryonic palate underlie topiramate effect on palate development (64).

Carbamazepine. Carbamazepine is an old but potent antiseizure medication for focal epilepsies with drug-resistant seizures. Although carbamazepine has been associated with increased risk of spina bifida (risk of 1% or less) with a dose-related effect, it is still regarded as an antiseizure medication with relatively low structural teratogenicity risk (27; 41) with no clear evidence of neurodevelopmental disorders (53).

A literature review on teratogenicity of carbamazepine was finalized by 2010, including eight cohort studies of 2680 pregnancies with carbamazepine monotherapy and the EUROCAT dataset, including 98,075 major congenital malformations registered over 3.8 million births (38). Taken together, these eight studies identified an overall prevalence for major congenital malformations of 3.3% (95% CI 2.7 to 4.2) after exposure to carbamazepine monotherapy in the first trimester, specifically related to the occurrence of spina bifida (odds ratio 2.6, 95% CI 1.2-5.3), as compared with no antiseizure medication intrauterine exposure and chromosomal controls. Compared with antiseizure medication monotherapy other than valproic acid, carbamazepine monotherapy showed no difference in the risk for spina bifida (1.1, CI 0.4 to 3.6). No difference in risk was observed for the other major congenital malformations evaluated (ie, hypospadias, cleft lip with or without palate, anomalous pulmonary venous return, diaphragmatic hernia).

In 7355 pregnancies from the EURAP Registry, 107 (5.5%) of 1957 pregnancies with carbamazepine exposure were associated with major congenital malformations, a risk that was significantly higher for all doses of carbamazepine (71). Carbamazepine at doses higher than 700 mg/day was associated with increased risk compared with levetiracetam at doses of 250 to 4000 mg/day (OR 2.41, 95% CI 1.33 to 4.38; p = 0.0055) and oxcarbazepine at doses of 75 to 4500 mg/day (2.37, 1.17 to 4.80; p = 0·0169).

However, in a subsequent nationwide cohort study based on the French health care databases, 512 of 1,886,825 pregnancies with intrauterine exposure to carbamazepine were not associated with increased risk for 23 types of major congenital malformations evaluated (09).

Oxcarbazepine, lamotrigine, and levetiracetam. Amongst 7355 pregnancies from the EURAP Registry, the prevalence of major congenital malformations was 10 (3.0%) of 333 for oxcarbazepine, 74 (2.9%) of 2514 for lamotrigine, and 17 (2.8%) of 599 for levetiracetam. Risks of major congenital malformations associated with lamotrigine, levetiracetam, and oxcarbazepine were within the range reported in the literature for offspring unexposed to antiseizure medication (71).

The Levetiracetam Pregnancy Registry was a United States-based prospective study overseen by an independent expert panel, which provided important contributions to the understanding of the antiseizure medication in question and also to the importance of different criteria adopted by various registries worldwide in defining malformations and the impact this carries on harmonizing worldwide data for practical clinical reference (67). This study confirmed previous evidence from other reports on levetiracetam monotherapy that levetiracetam is generally safe during pregnancy (27; 50).

From 491 women with epilepsy, outcomes were studied for 444 live births from 468 pregnancy exposures, with major congenital malformations diagnosed in 46 infants according to the Levetiracetam Registry (10%) but 22 (5%) according to EURAP criteria and only 7 (1.5%) according to the North American Antiepileptic Drug Pregnancy Registry criteria. Altogether, the Levetiracetam Registry expert panel did not find evidence suggestive of significant teratogenicity associated with prenatal exposure to levetiracetam.

Newest antiseizure medication. A 2021 review summarized all studies published since 2000 that reported neurodevelopmental outcomes on the following antiseizure medications: eslicarbazepine, gabapentin, lacosamide, lamotrigine, levetiracetam, oxcarbazepine, perampanel, topiramate, and zonisamide (44). Data for topiramate, gabapentin, and oxcarbazepine were so limited that conclusions could not be drawn. No studies investigated eslicarbazepine, lacosamide, perampanel, or zonisamide.

Lacosamide. One 2021 study reported on a total of seven pregnancies with exposure to lacosamide with total daily doses from 200 to 600 mg/day, with therapeutic drug monitoring indicating significantly lower levels during the second and third trimesters compared to pre-pregnancy levels (87). None of the neonates had major congenital malformations at birth.

Data from an animal study found a high incidence of embryonic lethality and malformations in mice exposed to lacosamide during embryonic development (47). Neonatal mice born to dams treated with lacosamide during gestation displayed psychomotor delay and morphological alterations in the prefrontal cortex, hippocampus, and amygdala that were associated with behaviors associated with schizophrenia spectrum disorders in adulthood.

Brivaracetam. Two studies to date reported on pregnant women with epilepsy taking brivaracetam, totaling five exposures: three monotherapies and two polytherapies (with eslicarbazepine plus perampanel, and with lamotrigine) (59; 45). Like lacosamide, serum concentrations of brivaracetam remained stable throughout pregnancy, whereas perampanel concentration seemed to steadily increase towards the end (45). Two babies were born with minor malformations (59).

Eslicarbazepine. From eslicarbazepine’s global safety database pregnancy reports up to 2017, there were 79 pregnancies with eslicarbazepine exposure: 28 from clinical trials and 51 from 8-year post-marketing surveillance (16). Eslicarbazepine was used in combination with other antiseizure medication in 11 of the 15 pregnancies, for which the outcome was spontaneous abortion and congenital anomaly. There was one maternal death (sudden unexplained death in epilepsy),ten spontaneous abortions, and five cases of major congenital malformations (including talipes, conjoined twins, cytogenetic abnormality) with eslicarbazepine in polytherapy with other antiseizure medication.

Perampanel. Clinical data from one company’s global safety database in 2018 consisted of 96 pregnancies reported in 90 women receiving perampanel, 26 (28.9%) in monotherapy (82). Overall, 43 pregnancies reached full term (all normal live births), 28 did not reach term (induced abortion, n = 18; spontaneous miscarriage, n = 6; incomplete spontaneous miscarriage, n = 2; premature delivery, n = 1; stillbirth [Fallot's tetralogy], n = 1), 18 were lost to follow-up, and seven were ongoing at data cutoff. Adverse events were reported in five full-term neonates (low Apgar score, n = 2; fatal neonatal aspiration, n = 1; cystic fibrosis and congenital deafness, n = 1; poor sucking reflex and shallow breathing, n = 1).

Behavioral and neurodevelopmental teratogenicity of antiseizure medication

The most significant knowledge gained in recent years that impacts the care of women with epilepsy during pregnancy relates to an improved understanding of behavioral and neurodevelopmental teratogenicity of antiseizure medication. Although tools and methods used to assess causal associations or risks related to intrauterine exposure to various antiseizure medications and unfavorable outcomes on neurodevelopment disorders vary across studies, there is an exponential increase in efforts to better delineate how prevalent these are and which antiseizure medications confer higher risk.

Currently, medical evidence can be summarized in a spectrum led on one end by the well-recognized dose-dependent risk for valproic acid and on the other end, lamotrigine being so far considered safe. There is still insufficient information for many newer and older antiseizure medications. One concern is the urge to treat with the safest medication for a future fetus while also considering medication that will be most effective for the mother to achieve seizure freedom for her epilepsy syndrome.

The concept of behavioral or cognitive teratogenicity is more novel than anatomical teratogenicity; evidence continues to grow for the risk of anatomical teratogenicity associated with valproic acid. With a few exceptions, most available studies remain retrospective regarding other antiseizure medications. Still, efforts should be made to build prospective cohort studies as, contrary to valproic acid, the use of other antiseizure medications will remain steady or increase. Thus, physicians must ensure their women patients with epilepsy are aware of current pregnancy registries and studies in which they can decide to enroll as participants for research. At this time, and as expected, there is insufficient information about the newer antiseizure medications (ie, those approved over the past 15 years). This is particularly impactful for women with drug-resistant epilepsies who have failed older medications and might currently be prescribed these new-generation drugs. Also adding a delayed decision for pregnancy, these women with epilepsy are often older and, therefore, at risk for other pregnancy outcomes unrelated to antiseizure medications.

The Neurodevelopmental Effects of Antiepileptic Drugs (NEAD) Study was a prospective observational multicenter study based in the United States that enrolled pregnant women with epilepsy treated with antiseizure medication monotherapy (with carbamazepine, lamotrigine, phenytoin, and valproic acid) from 1999 to 2004. The study aimed to determine if differential long-term neurodevelopmental effects exist for children who had intrauterine exposure to these commonly used antiseizure medications. The Maternal Outcomes and Neurodevelopmental Effects of Antiepileptic Drugs (MONEAD) Study is still actively recruiting women with epilepsy who are pregnant and aims to determine if women with epilepsy (1) have increased seizures during pregnancy and (2) are more likely to deliver via C-section and (3) have increased risk for depression during pregnancy and puerperium. For the children born to women with epilepsy, MONEAD will determine the long-term effects of antiseizure medication exposure in utero and via breastfeeding on verbal intellectual abilities and other neurobehavioral outcomes as well as the occurrence of adverse neonatal outcomes.

The NEAD study showed that at 3 years of age, children who had been exposed to valproic acid in utero had significantly lower IQ scores than those who were exposed to other antiseizure medication (51). After adjustment for maternal IQ, maternal age, antiseizure medication dose, gestational age at birth, and maternal preconception use of folate, the mean IQ was 101 for children exposed to lamotrigine, 99 for those exposed to phenytoin, 98 for those exposed to carbamazepine, and 92 for those exposed to valproic acid. The association between valproate use and IQ was dose-dependent.

Subsequently, verbal versus nonverbal cognitive outcomes were reported in 216 children who completed testing at 3 years of age within the NEAD study (52). As expected, valproic acid was associated with poorer cognitive outcomes. Verbal abilities were lower than nonverbal, with preconceptional folate use being associated with better verbal outcomes, which led the authors to hypothesize that fetal drug exposure could alter normal cerebral lateralization. Performance was negatively associated with valproic acid dose for both verbal and nonverbal domains and negatively associated with carbamazepine dose for verbal performance. No dose effects were seen for lamotrigine and phenytoin.

A total of 224 children completed 6 years of follow-up in the NEAD study (53). Age-6 IQ was lower in children exposed to valproic acid (mean 97) as compared to carbamazepine (105), lamotrigine (108), or phenytoin (108). High doses of valproic acid were negatively associated with IQ, verbal and nonverbal ability, memory, and executive function. IQ improved over time for infants exposed to any antiseizure medication. Mean IQs were higher in children exposed to periconceptional folate.

Learning and memory functions at 6 years of age were evaluated in 221 children exposed to one of these antiseizure medications during pregnancy and compared to a sample of normally developing children (15). Children exposed to valproic acid had, in a dose-related pattern, significantly low performance as measured through the Children's Memory Scale, indicating impairment in their ability to process, encode, and learn both auditory/verbal and visual/nonverbal material. In addition, they exhibited significant difficulty holding and manipulating information in immediate auditory working memory. Children exposed to the other antiseizure medication showed less clear results, and the authors indicated that further research is required to delineate their potential neurodevelopmental risks.

In the most recent report on the Maternal Outcomes and Neurodevelopmental Effects of Antiepileptic Drugs (MONEAD) study, outcomes of children at 2 years of age did not differ between children of women with epilepsy and children of healthy women in relation to maximum antiseizure medication exposure (most were taking lamotrigine or levetiracetam) during the third trimester (54). According to the Bayley Scales of Infant and Toddler Development Third Edition (BSID-III), language domain scores showed no differences between groups on the primary outcome of language domain. However, secondary analyses revealed that in the third trimester, higher maximum antiseizure medication levels were associated with lower scores for the motor domain, and higher maximum antiseizure medication doses were associated with lower scores in the general adaptive domain.

In children of women with epilepsy treated with antiseizure medication monotherapy, periconceptional folate was associated with better cognitive outcomes, with higher FSIQ at 3 and 6 years of age (56). Specifically at age 3, verbal index and receptive language index were higher, whereas at age 6, nonverbal index, expressive language index, and developmental neuropsychological assessment executive function were higher.

From a cohort of French children evaluated up to 5 years of age (mean 3.6 years), including 8848 with a history of intrauterine exposure to antiseizure medication monotherapy, a 4- to 5-fold, dose-dependent, increased risk of neurodevelopmental disorders was associated with valproic acid, in particular during the second or third trimesters of pregnancy (08; 17). Among other antiseizure medications studied (lamotrigine, carbamazepine, levetiracetam, oxcarbazepine, phenobarbital, clonazepam, topiramate, gabapentin, pregabalin), only pregabalin was consistently associated with an increased risk of neurodevelopmental disorders. Slight increases in risk initially observed with lamotrigine and carbamazepine were no longer present when analyzing only children from mothers with no known mental illness; thus, the authors suggest an effect of maternal mental illness or associated factors rather than exposure to these antiseizure medications per se.

The Kerala registry investigated language function in older children (76). Evaluation at 9 to 13 years of age revealed that, in comparison with peers born from healthy women, children born from women with epilepsy presented an average Core Language Scaled Score (CLSS) significantly lower for antiseizure medication monotherapy (4.5 units lower) and polytherapy (7.3 units lower). This was particularly significant for phenobarbital and valproic acid (mono- or polytherapy), whereas carbamazepine and phenytoin were not associated with differences.

In a prospective observational study, 161 children aged 6 or 7 years, identified from the European Registry of Antiepileptic Drugs and Pregnancy database in The Netherlands, were assessed using the Wechsler Intelligence Scale for Children and the Developmental Neuropsychological Assessment (32). Children exposed to valproic acid (n = 22) performed lower on all domains, especially language, than those exposed to carbamazepine (n = 32), lamotrigine (n = 82), or levetiracetam (n = 25). After controlling for maternal IQ and drug dose, the verbal IQ of valproic acid-exposed children was, on average, 9.1 to 13.4 points lower than for the other antiseizure medications. In this study, no significant dose effect was found.

The Norwegian Mother and Child Cohort Study (35) investigated language impairment in children via questionnaires with validated language screening tools answered by their mothers at age 5 (n = 117) and 8 years (n = 121). Compared to children of healthy mothers, there was an increased risk for language impairment for children exposed to antiseizure medication at both evaluated ages, with a protective effect of periconceptional folate supplementation. Higher maternal valproic acid concentrations (measured at 17 to 19 gestational weeks) correlated with language impairment at 5 years of age. In this study, carbamazepine monotherapy was associated with an increased risk of language impairment compared to controls at 8 years of age.

Amongst 913,302 children born in Denmark over a 14-year period (including 580 children exposed to valproic acid), 6958 (0.8%) had intellectual disability, and 14,967 (1.6%) were identified as having intellectual disability with delayed childhood milestones (20). Intrauterine valproic acid exposure was associated with an increased risk of intellectual disability as well as delayed childhood milestones. Increased risk was also associated with prenatal exposure to monotherapy with carbamazepine, clonazepam, and oxcarbazepine but not lamotrigine.

Intrauterine antiseizure medication exposure and risk of autism spectrum disorder and attention-deficit hyperactivity disorder have also been the focus of clinical studies. In a prospective cohort of women recruited from prenatal care settings whose offspring (n = 415) were followed longitudinally until 6 years of age, neurodevelopmental disorders were diagnosed through standard clinical evaluation independently of the research enrollment. There was an increased risk of neurodevelopmental disorders in children exposed to monotherapy or polytherapy with valproic acid compared with control children, with autism spectrum disorder being the most frequent diagnosis. No significant risk increase was found for intrauterine exposure to carbamazepine or lamotrigine (11).

Within the Swedish Pregnancy Registry, first-trimester use of valproic acid was associated with both autism spectrum disorder and attention deficit hyperactivity disorder, with a small and not statistically significant association with carbamazepine (86). No evidence of risk related to lamotrigine has been determined. Contrary to observations in autism spectrum disorder at large, no male predominance has been observed in children born from women with epilepsy with intrauterine valproic acid exposure (31).

A systematic review and network meta-analysis assessed cognitive development and autism or dyspraxia in children exposed to antiseizure medication during pregnancy and breastfeeding (84). Among all antiseizure medications, only valproic acid was statistically significantly associated with more children experiencing cognitive developmental delay than controls (OR 7.40, 95% credible interval (CrI) 3.00 to 18.46). Furthermore, compared with controls, only valproic acid (OR 17.29, 95% CrI 2.40 to 217.60), oxcarbazepine (OR 13.51, 95% CrI 1.28 to 221.40), lamotrigine (OR 8.88, 95% CrI 1.28 to 112.00), and lamotrigine combined with valproic acid (OR 132.70, 95% CrI 7.41 to 3851.00) were significantly associated with increased occurrence of autism or dyspraxia. For the cognitive-developmental delay and psychomotor developmental delay outcomes, children exposed to the combination of carbamazepine, phenobarbital, and valproic acid were at greater odds of harm than those not exposed to antiseizure medication (84).

Folic acid and anatomical and cognitive/behavioral teratogenicity

Although it is recognized that folic acid can significantly improve outcomes related to anatomical and behavioral teratogenicity, many aspects of folic acid supplementation remain suboptimal. Many women with epilepsy lack knowledge about its importance prior to conception, many remain not compliant to its prescription, and some wait to find out they are pregnant to start taking it. Moreover, although sexually active women with epilepsy in reproductive age should be prescribed folic acid even if not planning pregnancy, dedicated time for counseling on pregnancy planning adds time at each visit. Discrepancies exist, with some countries being less proactive in preconception folate prescription (37).

The dose necessary to positively impact these outcomes is still a subject of debate, although more, rather than less, often high doses (up to 5 mg daily) are prescribed. Moreover, its implication in primary and secondary prevention of bad outcomes is still debated. Most studies related to folic acid deficiency addressed neural tube defects as an outcome, but anatomical and behavioral teratogenicity should be considered. Whereas a 0.4 mg daily dose is recommended to all women with epilepsy planning pregnancy, 4 mg daily dose has been recommended for secondary prevention of neural tube defects, and often 5 mg/day is prescribed. The recommended dose to reduce risk for the various neurodevelopmental disorders and major congenital malformations remains unclear.

Metaanalysis of data from three randomized controlled trials of folic acid supplementation for women with a previous pregnancy with a neural tube defect indicate a 70% [95% CI: 35–86] reduction in recurrence with the secondary prevention (07). Eight studies examined the effect of folic acid supplementation on the incidence of first occurrence of neural tube defect, but data for primary prevention could only be compiled from one randomized controlled trial and three cohort studies, with a 62% (95% CI: 49 to 71) reduction.

In the Raoul Wallenberg Australian Register of Antiepileptic Drugs in Pregnancy, no statistically significant effect of folic acid dosage up to 5 mg/day was found in reducing fetal major congenital malformations rates (including spina bifida) in women with epilepsy (2104 pregnancies) taking folic acid either before and during pregnancy or during early pregnancy only (81).

The question of what dose of folic acid should be supplemented is complicated by what antiseizure medication is prescribed, as antiseizure medications have a range of teratogenicity; one needs to keep both in mind when weighing a mother’s seizure control and risk for the fetus. From nearly 2000 pregnancies in women with epilepsy over 2 decades in the Kerala Registry of Epilepsy and Pregnancy in India, high-dose (5 mg/day) folic acid usage during pregnancy had improved over 2 decades (42). Seizure control has improved overall, and the proportion of women with epilepsy who remained off antiseizure medication during pregnancy was reduced. Despite changes in the antiseizure medication prescription over time (preference for newer antiseizure medications instead of phenobarbital, phenytoin, and clonazepam, with no changes in carbamazepine or valproic acid prescription rates), major congenital malformation rates remained unchanged, probably due to continued use of valproic acid and carbamazepine, increased use of topiramate and clobazam, and no change in prescribed polytherapy.

The impact of maternal folic acid supplementation on a child’s cognition, development, and behavior is only starting to be clarified. In a population-based prospective study (Norwegian Mother and Child Cohort Study, with 104,946 18- to 36-month-old children of women with epilepsy), the degree of autistic traits was inversely associated with maternal plasma folate concentrations and folic acid doses, whereas antiseizure medication concentrations were not (06).

The Norwegian cohort was also evaluated for the effect of maternal folic acid supplementation, maternal plasma folate, and antiseizure medication concentrations on language delay. Compared to controls, children of women with epilepsy taking folic acid showed a reduction in the odds ratio for language delay at 18 months (from 3.9 to 1.7) and 36 months (from 4.7 to 1.7) (34). The authors found that the positive effect was significant only when a supplement was used throughout the pregnancy, including the period of 4 weeks pre-pregnancy.

The Neurodevelopmental Effects of Antiepileptic Drugs (NEAD) study on the effect of dietary folate (from natural folate sources plus a folic acid-enriched diet), as compared to folic acid supplements, evaluated IQ from 6 years of age in children born after intrauterine antiseizure medication exposure (65). Importantly, the authors found that whereas folate from food alone was not associated with improvement of age-6 IQ, periconceptional supplement use was associated with a 10.1-point higher age-6 IQ. Folate from food plus supplement also showed that a higher intake of folate was associated with higher IQ. This supports the rationale for folic acid supplementation even when food is fortified with folic acid because the latter alone is not sufficient to improve cognitive outcomes for offspring of women with epilepsy taking antiseizure medications during pregnancy.

A total of 311 children of 305 women with epilepsy treated with antiseizure medication monotherapy from the NEAD cohort were cognitively evaluated at 3 and 6 years of age (56). Periconceptional folate was associated with a higher Full Scale Intelligence Quotient (FSIQ) at both 3 and 6 years of age. Significant effects for other measures included Nonverbal Index, Expressive Language Index, and Developmental Neuropsychological Assessment Executive Function at 6 years of age and Verbal Index and Receptive Language Index at 3 years of age.

Breastfeeding and effect on postnatal outcomes

There is still limited information on the effect of antiseizure medication exposure via maternal milk for newborns of women with epilepsy. Breastfeeding was associated with less impaired development at ages 6 and 18 months compared with those with no breastfeeding or breastfeeding for fewer than 6 months for infants born from women with epilepsy in the population-based prospective Norwegian Mother and Child Cohort Study (83). In the Maternal Outcomes and Neurodevelopmental Effects of Antiepileptic Drugs (MONEAD) prospective cohort study with follow-up in children until 6 years of age, antiseizure medication concentrations were quantified from blood samples collected from infants and mothers at the same visit (5 to 20 weeks after birth) (05). A total of 164 matching infant-mother concentration pairs from 138 infants were analyzed, including the following antiseizure medications: carbamazepine, carbamazepine-10, 11-epoxide, levetiracetam, lamotrigine, oxcarbazepine, topiramate, valproic acid, and zonisamide. In 49%, antiseizure medication concentration in these infants was below the lower limit of quantification. The median percentage of infant-to-mother antiseizure medication concentration ranged from 0.3% to 44.2%. For lamotrigine, maternal concentration was a significant factor associated with its concentration in infants.

New onset epilepsy during pregnancy

Most seizures during pregnancy occur in women with previously diagnosed epilepsy. The prevalence of new-onset epilepsy during pregnancy has not yet been determined. A retrospective study conducted at the Johns Hopkins Medical Institutions considered all women evaluated for concomitant diagnoses of pregnancy and seizures, excluding eclampsia (48). Over a 5-year period, five of 84 pregnant women were assessed for possible seizures, with diagnosis of first unprovoked seizure supported by epileptiform abnormalities on EEG. All women delivered at term with no major complications, and four of five women continued to have seizures after delivery.

In another study from Azerbaijan, over a 6-year period, 12 of 112 pregnant women presented with a first seizure during pregnancy, 50% of them occurring in the first trimester (57). These women with new-onset epilepsy had an increased risk of C-section and neonatal hypoxia compared with women without epilepsy; there was no increased risk as compared to women with epilepsy diagnosed before pregnancy.

Change in seizure frequency during pregnancy

In most women with epilepsy, seizure frequency in pregnancy remains unchanged from before pregnancy; thus, being seizure-free before pregnancy predicts a good outcome. Significant progress has been made in our understanding of the factors that may predict seizure deterioration or improvement and recurrence or remission; however, they have not been fully elucidated. In most patients (69%), epilepsy largely follows the same pattern as before pregnancy. Furthermore, most women (84%) who are seizure-free for at least 9 months before pregnancy do not experience any recurrences (24; 04). In the minority of women whose epilepsy control changes during pregnancy, what happens in the various types of seizures (particularly generalized tonic-clonic seizures) and syndromes is largely unknown. Seizure frequency and severity may vary between different pregnancies in the same patient.

Some studies indicate that changes are more likely to occur during the first and the third trimesters and that seizure frequency tends to revert to pre-pregnancy levels after delivery. Using the first trimester as a reference, seizure control remains unchanged throughout pregnancy in 64% of women, 93% of whom are seizure-free during the entire pregnancy. For those with a change in seizure frequency, 17% have an increase and 16% a decrease in seizure frequency (23). Seizure deterioration is more frequently observed in women with focal epilepsies than with generalized epilepsies (04). In contrast, better seizure control is likely to improve in women with epilepsy with a catamenial epilepsy pattern (12).

In the EURAP Epilepsy Pregnancy Registry, during 66.6% of 3806 pregnancies (in 3451 women with epilepsy), women remained seizure-free throughout pregnancy, of which those with idiopathic generalized epilepsy were more likely to remain seizure-free (73.6%) than women with localization-related focal epilepsy (59.5%; p < 0.0001). Worsening in seizure control from the first to second or third trimesters occurred in 15.8% of pregnancies. Generalized tonic-clonic seizures occurred in 15.2% of the pregnancies. Compared with other monotherapies, pregnancies exposed to lamotrigine were less likely to be seizure-free; they had more generalized tonic-clonic seizures and a greater likelihood of worsening seizure control from first to later trimesters. As expected, pregnancies exposed to lamotrigine most frequently required an increase in drug dose throughout pregnancy due to increased lamotrigine clearance. There were 21 cases of status epilepticus (10 convulsive), none with maternal mortality, and only one with a subsequent stillbirth (04).

Seizure control deterioration in pregnancy

Worsening in seizure control during pregnancy can be attributed to poor compliance, pregnancy-related reduction in plasma drug concentrations, sleep deprivation, fatigue, hormonal changes, or psychological factors.

An important factor in seizure deterioration is inadequate plasma levels of antiseizure medication for some drugs, mainly because of the multiple physiological changes occurring during pregnancy that influence drug disposition, including increased volume of distribution, increased renal elimination, altered hepatic enzyme activity, and a decline in plasma protein concentrations (61; 04; 68). A decrease in plasma albumin and protein binding and the displacement of antiseizure medication by endogenous compounds lead to an alteration in the ratio of total to free drug, which is particularly important for highly protein-bound drugs (40).

For many antiseizure medications, significant increases in clearance and, therefore, decreases in plasma levels are characteristic during pregnancy. There is evidence for significant increases in the clearance of lamotrigine, phenytoin, phenobarbital, and the active monohydroxy derivative of oxcarbazepine during pregnancy. The clearance of levetiracetam increases in the first trimester of pregnancy (85).

The pronounced decline in plasma concentrations of lamotrigine during pregnancy has been associated with deterioration in seizure frequency in more than 50% of women with epilepsy; patients often require significant dose adjustments or additional antiseizure medication (23; 04). In the Australian Pregnancy Registry, the control of convulsive seizures was significantly worse with lamotrigine than with valproic acid during the entire pregnancy and comparatively worse than with carbamazepine during the second and third trimesters (77).

For the same reasons, oxcarbazepine can also be associated with poorer seizure control if no dose changes are made (23; 63).

The American Academy of Neurology and American Epilepsy Society have assessed the changes that occur during pregnancy for the following antiseizure medication (25):

	Lamotrigine. Pregnancy increases the clearance and decreases plasma levels. The decrease in plasma level is associated with increased seizure frequency (one Class I and two Class II studies).
	Carbamazepine. Pregnancy probably causes a small decrease in plasma levels: 9% in the second trimester and 12% in the third trimester (one Class I study).
	Phenytoin. Pregnancy probably causes an increase in the clearance and a decrease in plasma levels (one Class I study).
	Oxcarbazepine. Pregnancy possibly causes a decrease in the level of the active monohydroxy derivative of oxcarbazepine (two Class III studies).
	Levetiracetam. Pregnancy possibly causes a decrease in plasma levels (one Class II study).
	Phenobarbital, valproate, primidone, and ethosuximide. Evidence of a change in clearance or plasma levels of phenobarbital, valproic acid, primidone, or ethosuximide during pregnancy is insufficient to reach a conclusion.

Therapeutic drug monitoring in pregnancy has been formally recommended during pregnancy and puerperium, in particular for antiseizure medications that present changes in bioavailability as described above. The ideal antiseizure medication concentration should be established for each patient before conception, and monitoring should be performed during each trimester, or at least monthly, especially for lamotrigine (62). Empiric postpartum tapering of lamotrigine can minimize the risk of postpartum toxicity (62).

Effect of seizures on mother and the unborn baby

Generalized tonic-clonic seizures may cause severe harm to both mother and her unborn baby but may be preventable with appropriate management. The harmful effects of seizures, particularly generalized tonic-clonic seizures, are multiple and may be severe or even fatal and may be accidental (falls, drowning) or nonaccidental (aspiration, pulmonary edema, cardiac arrhythmias, cardiac asystole). Furthermore, the risk of sudden unexplained death in epilepsy is significantly higher in patients with generalized tonic-clonic seizures than in patients with other seizures. Pregnancy is a particularly vulnerable period, both for the woman and her unborn baby. In pregnancy, a generalized tonic-clonic seizure imposes an increased risk of damage or death to the mother and her unborn baby. Other types of seizures that may be harmful are those associated with autonomic disturbances and cardiac asystole. The consequences of convulsive status epilepticus are much worse than those of brief seizures.

The harmful effects of seizures on the unborn baby also arise from accidental injury or the hypoxic and other effects of the seizure on the mother or directly on the fetus (eg, lactic acidosis, bradycardia). Fetal fatalities may rarely occur during status epilepticus; in one study, only one fetus died in 12 pregnancies of women with epilepsy with convulsive status epilepticus; fortunately, there was no maternal mortality (23). The number of stillbirths is not increased among women adequately treated for epilepsy during pregnancy.

Data from the EURAP registry on the risk of spontaneous abortions and stillbirth associated with antiseizure medications, collected from 7,055 pregnancies exposed to monotherapy (lamotrigine, carbamazepine, valproic acid, levetiracetam, oxcarbazepine, phenobarbital) or polytherapy, identified 632 intrauterine deaths (592 spontaneous abortions and 40 stillbirths). These intrauterine deaths showed neither higher risk to one antiseizure medication in particular nor a relationship with dose at conception. However, occurrence was associated with antiseizure medication polytherapy and parental history of major congenital malformations as well as with maternal age and number of previous intrauterine deaths (70).

Obstetric and other pregnancy related complications

Most pregnancies in women with epilepsy are not associated with severe adverse fetal outcomes, including fetal loss (MONEAD study) (55). However, there have been reports of higher risk for preterm labor and preterm delivery, cesarean section, low birth weight (49; 02; 36; 66), fetal hypoxia, and Apgar score of 7 or lower at 5 minutes in offspring of women with epilepsy as compared to healthy women (57). In a nationwide study from Finland, women with epilepsy were more often hospitalized during pregnancy for accidents or other external causes, premature rupture of membranes, and premature contractions (02). Moreover, an increase in the relative risk of preeclampsia in women with epilepsy has been observed (49; 19).

The risk of seizures during labor and delivery is higher in those who also had seizures earlier in pregnancy; 1% to 2% of women will have a generalized tonic-clonic seizure during labor, and an additional 1% to 2% will have a generalized tonic-clonic seizure within the next 24 hours (03). Women with epilepsy at risk of seizures should be managed in specialized obstetric units with maternal and neonatal resuscitation facilities.

Pregnant women on antiseizure medication during pregnancy, whether for epilepsy or other neuropsychiatric indications, are at a higher risk of delivering prematurely and giving birth to newborns who are small for gestational age (28). The risk may vary by drug; for example, the risk was found to be 7.3% for lamotrigine and 18.5% for topiramate monotherapies.

Puerperium: breastfeeding and sleep deprivation

It is important to discuss with women with epilepsy, prior to delivery, their plan regarding breastfeeding and any support system in place to avoid sleep deprivation and its effect on seizure control. Ideally, women with epilepsy, breastfeeding or not, should aim for a minimum of 6 hours of continuous sleep. This can be achieved with arrangements for bottle supplementation during the night, including pumping maternal milk and allowing a spouse or caregiver to take charge of newborn feeding at least once during the night. Newborn nursing and sleeping routines should be taken into consideration when planning antiseizure medication dose adjustment back to the pre-pregnancy dosage, and if significant sleep deprivation occurs, a slightly higher dose than preconception might be considered to mitigate the risk of seizures at this stage. To avoid peak of dose, it is advisable to plan to breastfeed an hour before the next antiseizure medication dose. If antiseizure medications were increased during pregnancy, the primary route of elimination for each individual antiseizure medication would be the main determinant of the tapering rate to pre-pregnancy dose or slightly above that (72).

Sudden unexplained death in epilepsy and epilepsy-related mortality in pregnancy

Although data are limited, there is probably a 10-fold increase in mortality in pregnancy of women with epilepsy compared to women without epilepsy (22). In a report on 13,978 pregnancies in women with epilepsy, 14 epilepsy-related deaths occurred, of which 11 (79%) were sudden unexplained death in epilepsy: nine occurred during pregnancy, and five were postpartum. Nine (64%) of the deaths were in women taking lamotrigine (seven as monotherapy). It was, thus, estimated that 1 in 1,000 women died from epilepsy (mostly sudden unexplained death in epilepsy) during or shortly after pregnancy. The high proportion of women taking lamotrigine in this study reporting on sudden unexplained death in epilepsy may reflect country-specific (United Kingdom) prescribing practices. However, because lamotrigine levels significantly decrease during pregnancy, and pregnant women on this drug have more difficulties with epilepsy control, this might not be the sole explanation (04). Given the potential risks, every attempt should be made to prevent seizures, particularly convulsive, during pregnancy and postpartum, and maintaining therapeutic lamotrigine levels should be a priority (22). Generalized tonic-clonic seizures or focal to bilateral tonic-clonic seizures could be associated with risks to the mother and the fetus. In generalized tonic-clonic seizures, the resulting status of hypoxia and lactic acidosis shared with the fetus through the placenta could lead to asphyxia.

Maternal mortality (defined as deaths in pregnancy and the first 42 days after termination of pregnancy) in women with epilepsy was more than five times higher compared to the mortality in women without epilepsy in a matched case-control study of pregnant women in Denmark, with five maternal deaths reported from 11,976 pregnancies of women with epilepsy (13).

Brain stimulation during pregnancy

There are still few cases of pregnant women with epilepsy treated with brain stimulation, including vagal nerve stimulation and responsive neurostimulation. A review of outcomes from 44 pregnancies in 38 women with epilepsy derived from 10 publications did not review major congenital malformation rates unexplained by the concomitant antiseizure medication-associated risk (21). A retrospective review of 14 pregnancies in 10 women with epilepsy treated with responsive neurostimulation from 2014 to 2020 in the United States did not identify any major congenital malformations, and obstetric or perinatal complications were within the previously reported rates for women with epilepsy (46). Two pregnant women with epilepsy successfully treated with deep-brain stimulation presented no change in seizure control during pregnancy and gave birth to healthy children (10).