Contraception, Fecundity and Pregnancy in Women with Epilepsy: An Update on Recent Literature

Abstract

Purpose of review:

Caring for women with epilepsy requires specialized knowledge about potential teratogenicity of anti-seizure medications, interactions with hormonal contraception, and pregnancy outcomes.

Recent findings:

There has been an improvement in understanding the cognitive outcomes of infants exposed in utero in recent years. Folic acid supplementation helps mitigate the cognitive teratogenicity of anti-seizure medications. Recent updates provide reassurance that seizure frequency tends to remain stable throughout pregnancy. There is conflicting evidence about the fecundity impact of epilepsy and antiseizure medications in women with epilepsy.

Summary:

Recent research highlights the importance of early counseling about the risks and interactions of contraception, pregnancy and anti-seizure medications. More research is needed to understand fertility in women with epilepsy.

Introduction

The prevalence of women with epilepsy (WWE) of childbearing age in the United States is around 1.5 million, with around 24,000 WWE giving birth each year [1]. There are multiple factors to consider when caring for WWE throughout their childbearing years, including potential effects of epilepsy and anti-seizure medications (ASMs) on contraceptive efficacy, fecundity and pregnancy. The risk of structural teratogenesis has been an important consideration for some time, and, more recently, an improved understanding of cognitive outcomes for children exposed to in-utero ASMs has been particularly important to patient counseling and management in WWE. The aim of this paper is to review the recent literature about contraception, fertility and advances in our knowledge about pregnancy outcomes in WWE.

Contraception

Contraceptive counseling is an essential aspect of the care for WWE. Around half of WWE reported that their pregnancies were unplanned [2]. The majority (~65%) of women on enzyme-inducing anti-seizure medications were unaware of the interaction between this and hormonal contraceptives [3], placing them at increased risk of unplanned pregnancy. A recent study by the Epilepsy Birth Control Registry (EBCR), a web-based survey performed a decade later, demonstrates that nearly 60% of WWE remain at risk for unplanned pregnancies. Nearly 30% did not use highly effective contraception methods and an additional 7% used hormonal contraception while on an enzyme-inducing anti-seizure medication posing a risk of contraceptive failure. Older age, insurance coverage and increased income were associated with highly effective contraceptive use [4]. This highlights the importance of counseling about contraception and pregnancy early.

Intrauterine devices (IUDs) have emerged as the most reliable contraceptive method and are the contraceptive method of choice for WWE [5]. Both copper and hormonal IUD options are good choices for WWE, as even hormonal IUDs that contain levonorgestrel exert effects locally so interaction with ASMs is minimal. There was a failure rate of 1.1% per year with the 52-mg levonorgestrel-releasing IUD [6]. While this is slightly higher than the general population, this remains significantly more effective than other contraceptive methods for WWE. Neurologist counseling about the use of an IUD has a significant influence on a patient’s choice to adopt one [7].

Enzyme-inducing ASMs induce hepatic P450 enzymes, which in turn induces the metabolism of both synthetic estrogens and progestins. There is a risk of contraceptive failure when these medications and hormonal contraceptives have been used. The WHO advises against using combined oral contraceptive pills, the vaginal ring and the transdermal patch while on enzyme-inducing ASM [8]. Progestin implants and depot medroxyprogesterone acetate can be considered in women taking enzyme-inducing ASMs but rare cases of contraceptive failure have been reported with the implant. Barrier methods are not highly effective for women with or without epilepsy [9].

Combined oral contraceptive pills containing synthetic estrogens can reduce lamotrigine serum concentrations by nearly 50% [10, 11, 12]. This leads to fluctuation in lamotrigine concentrations, with levels being significantly higher when measured during the placebo week compared to the weeks on active hormone [13]. A recent case series of 6 women with epilepsy using a vaginal contraceptive ring found a similar pattern, with decreased lamotrigine levels during the active phase of contraception compared to the time off the ring [14]. This has implications for counseling and maintaining a therapeutic lamotrigine level for WWE using these forms of non-oral contraception similar to combined oral contraceptive pills.

The interaction between lamotrigine and combined oral contraceptive pills has been presumed to be due to the synthetic estrogen in the pills. Estrogen induces uridine-diphosphate glucuronosyltransferase enzymes which increases glucoronidation and excretion of lamotrigine [10]. A recent prospective cohort study attempted to investigate the role of progestins in this interaction. Combined oral contraceptive pills containing drospirenon and levonorgestrel reduced lamotrigine concentrations, though there was no difference in lamotrigine concentrations associated with combined pills containing gestoden [15]. The authors suggest that the unique progestin component of each pill, rather than the estrogen component, may be responsible for the decrease in lamotrigine levels in women on combined oral contraceptive pills which would be a major shift in paradigm. However, this study had several limitations. First, the WWE in this study were all on contraceptive pills which include an estrogen component. While there was no difference in overall estrogen exposure as a group, the individual estrogen components within a patient were not evaluated or reported. In addition, this study had a relatively small sample size (N=20). Group effects were reported and compared, but there is a wide range of individual sensitivity to estrogen components which may lead to biasing the results in such a small sample size (N = 20). More work is needed to fully understand the complex interactions between synthetic progesterone and lamotrigine, including investigation of progesterone-only contraceptive methods. A second finding of this study was that there was no effect of lamotrigine on the concentration of any of these progestins. This is important because, previously, there had been a suggestion that lamotrigine can lower plasma concentrations of levonorgestrel without certain effects on contraceptive efficacy [16].

The EBCR study evaluated the transition period from different types of contraceptive methods for WWE. They found that transitioning from a non-hormonal contraceptive to a hormonal method was associated with increased seizure frequency for the 3 months following the change. Conversely, the transition from hormonal to non-hormonal contraceptive methods was associated with a decrease in seizure frequency. When the effect of individual anti-seizure medications was evaluated, only lamotrigine was found to be significant [17*]. This is consistent with prior knowledge of lamotrigine interaction with combined contraceptive methods.

Fecundity

There is conflicting data about fertility rates in women with epilepsy. Data in a cohort of patients in Rochester collected from 1935–1974 found a reduced fertility rate to 85% of expected for WWE [18]. In contrast, a retrospective cohort study in Iceland found no overall difference in the birth rate to WWE and age matched controls except in the case of epilepsy and comorbid cerebral palsy or intellectual disability [19]. The EBCR group retrospectively surveyed nearly 1,000 WWE about reproductive history to evaluate for infertility (where pregnancy was not achieved within 1 year) and impaired fecundity (the percentage of women who were infertile or did not carry pregnancy to live birth). They found an infertility rate of 9.2% and impaired fecundity rate of 20.7% in WWE, which was higher than the ~6.5% and 12.7% of each occurring in the general population [20].

Fertility rates are impacted by numerous influences including social factors, frequency of intercourse, and biological factors. In the first prospective study on this subject, Pennell and colleagues found no difference in sexual activity, pregnancy rates, time to pregnancy or birth rates in WWE compared to control women without epilepsy [21]. A Danish Registry of a cohort of women undergoing assisted reproduction aimed to investigate the success of this procedure in WWE. The chance of chemical and clinical pregnancy after embryo transfer was similar between WWE and controls. Likewise, the live birth rate in both WWE and controls was not significantly different. There was no difference found in the success of pregnancy when treatment regimens of polytherapy or monotherapy were compared [22]. These novel findings provide reassurance that assisted reproductive technology (ART) is an option for WWE who have difficulty conceiving. The effect of assisted reproduction protocols on seizure control is understudied, however. Two cases of seizure exacerbation during treatment have been reported [23]. One occurred during ovarian stimulation in a woman on levetiracetam and another was in a woman on lamotrigine who was taking estrogen valerate for endometrial priming. Larger prospective cohorts are needed to clarify how often this occurs in WWE utilizing ART. Anecdotally, in our practice seizure exacerbation during ART is infrequent and is not a reason to avoid these procedures. We do typically request the hormonal protocols in advance and make temporary dosing changes (mostly to lamotrigine) when exogenous estrogens are used.

Pregnancy

The majority of WWE require ASMs throughout pregnancy. The goal of treating pregnant WWE is achieving a balance between the risks of seizures and potential teratogenic effects of anti-seizure medications. Seizures during pregnancy are associated with risk of preterm delivery and small for gestational age babies, and there is a risk of decreased IQ in children of mothers with frequent convulsive seizures [1].

Seizure Frequency during Pregnancy

Seizure frequency tends to remain stable throughout pregnancy the majority of WWE, with increased seizure frequency occurring in around one third and decreased frequency in a minority of women [24, 25, 26]. The Maternal Outcomes and Neurodevelopmental Effects of Antiepileptic Drugs (MONEAD) study recently published a prospective cohort study evaluating the frequency of seizures during pregnancy. Women with epilepsy who were not pregnant served as a control group in this study. Two epochs were established to compare the frequency of seizures; epoch 1 included pregnancy and the peripartum period (6 weeks after birth) and epoch 2 included from 6 weeks to 9 months post-partum in the pregnancy group. The control group had epochs defined by the same timing (10.5 months for epoch 1 and 7.5 months for epoch 2). A higher frequency of seizures in epoch 1 than epoch 2 occurred in 23% and 25% of the pregnant and control groups, respectively, which was not significantly different. This study also found that anti-seizure medications were changed more frequently in the group of pregnant women than the control group [27*]. No change in seizure frequency was found between pregnant women and non-pregnant controls in a smaller cohort in Italy [28]. The results of these recent studies are reassuring. Frequent therapeutic drug monitoring with adjustment of dosage is important to maintain steady levels. Lack of therapeutic drug monitoring and dose adjustment may be the reason that previous studies have had inconsistent results about seizure frequency throughout pregnancy.

Structural Teratogenesis

It is well established that ASMs are associated with higher rates of major congenital malformations (MCMs). Much of the data has focused on older ASMs such as valproate and carbamazepine or the preferred medications of levetiracetam and lamotrigine, leading to different prescribing practices over time. The International Registry of Antiepileptic Drugs and Pregnancy (EURAP) reported a decrease in the use of valproic acid and carbamazepine and increase in lamotrigine and levetiracetam usage when a cohort from 2010–2013 was compared to a previous cohort from 2000–2005. There was a 27% decrease in the rate of children born with MCMs which likely reflects the move away from prescribing ASMs with higher morbidity [29]. The Kerala Registry of Epilepsy and Pregnancy (KREP) is a prospective registry of WWE at a tertiary epilepsy center in India. KREP noted that nearly 18% of WWE in their cohort continue to take valproic acid during pregnancy, and the rate of MCMs was higher in infants exposed to valproic acid [30].

A levetiracetam-specific registry recently found an MCM rate of 10.4%, and 9.4% in those on monotherapy [31]. This is higher than MCM rates previously reported in other registries. This registry used a modified version of criteria in the Centers for Disease Control and Prevention (CDC)’s Metropolitan Atlanta Congenital Defects Program (MACDP) [32]. In order to better understand the discrepancy, study personnel from EURAP and North American Antiepileptic Drug Pregnancy Registry (NAAPR) reviewed this report. Using the criteria of EURAP or NAAPR, the MCM rates with monotherapy exposure dropped to 3.6% and 1% respectively [31]. This discordance was largely due to the levetiracetam registry classifying abnormalities that would be minor anomalies in the other registries or including abnormalities due to a known genetic condition. This highlights the importance of consistent methodology in reporting pregnancy outcomes.

There are several newer anti-seizure medications with limited data about safety in pregnancy including zonisamide. Animal data has raised concern about the safety of this medication due to developmental abnormalities [33]. In humans, in utero zonisamide exposure has been associated with low birth weight [34]. In the MONEAD study, 1 infant out of 13 exposed to zonisamide had inguinal hernia and pinna malformation, representing 7.7% rate of MCM [35*]. Previously, a small case series found a 7.7% incidence of MCM with zonisamide use: anencephaly and atrial septal defect [36]. Both cases were exposed to polytherapy with phenytoin or phenytoin and valproic acid, making the contribution of zonisamide unclear. Data from the UK and Ireland Epilepsy and Pregnancy Register (UKIEPR) recently found a rate of MCM of 11.5% (3 of 26) in monotherapy cases and 5.8% (5 of 86) in polytherapy cases for pregnant women on zonisamide. The severity of these malformations ranged from as minor (ear skin tag) to severe (anencephaly) [37*]. This study raised concern that zonisamide may be more teratogenic than originally thought. The largest sample of zonisamide exposures was published as supplementary material by the NAAPR. In this cohort of 90 infants, no MCMs were reported [38]. If the data from all these monotherapy exposures are pooled, 4 out of 138 zonisamide exposure have been reported to have MCM. Clearly more data is needed to draw firm conclusions about its teratogenic risk.

Perampanel is an AMPA-type glutamate antagonist used to treat focal seizures. A recent case series reported pregnancy outcomes in 90 WWE taking perampanel. Forty-three (~45%) of these pregnancies reached full term; 5 of these 43 infants had adverse events at birth of varying severity [39]. While these data note a concerning degree of MCMs and poor outcomes, the data was incomplete and further investigation is needed.

It is important to note that antiseizure medication exposure is not the only factor that may raise the risk of MCMs in children of WWE. The Australian Register of Antiepileptic Drugs in Pregnancy (APR) is attempting to identify other non-drug variables that may contribute to MCMs by collecting data on >2,000 pregnancies in WWE. This study found that maternal age >31 years, family history of malformations, preexisting maternal psychiatric illness, and tobacco use were all independently associated with a higher risk of MCM [40,41].

Cognitive Teratogenesis

Some of the most robust recent research about pregnancy in WWE regards cognitive teratogenicity with in utero exposure to ASMs. Previously, the Neurodevelopmental Effects of Antiepileptic Drugs (NEAD) study found that age-6 full scale IQ was lower for children exposed to valproate than carbamazepine, lamotrigine or phenytoin. Moreover, this relationship was dose-dependent for valproate, with poorer cognitive scores associated with higher doses [42]. Children exposed to valproate particularly struggled with attention and visual and auditory working memory [43].

The MONEAD study continues to prospectively evaluate the cognitive outcomes of children born to mothers with epilepsy. Recently the association between maternal anti-seizure medication blood level throughout pregnancy and verbal intellectual abilities in toddlers was reported. The majority of women in this study were on lamotrigine (46%) or levetiracetam (33.2%) monotherapy. There were no differences in the language performance in 2-year-olds born to WWE compared to those born to healthy control women. Maternal IQ, female sex and child birth weight were all independently associated with higher language performance [44*]. Higher ASM blood levels in the third trimester correlated with poorer scores in the motor domain and higher ASM doses correlated with poorer general adaptive scores [44*]. Overall, these results are reassuring about the cognitive safety of lamotrigine and levetiracetam, though data in older children is known to be more closely associated with long term outcomes. The MONEAD study will continue to follow this cohort and repeat neuropsychological testing through 6 years of age, which will aid in our understanding of ASM cognitive teratogenicity.

The KREP study examined the long-term effects of ASM exposure on children born to WWE. Older children (aged 9–13 years) of WWE who were exposed to antiseizure medications were more likely to have language deficits than control children. The effect was worse with polytherapy than monotherapy, and valproic acid and phenobarbitone were particularly associated with poorer language function. No difference was found in the language performance of children exposed to carbamazepine or phenytoin [45*].

There is a recent systematic review of 35 articles on the neurocognitive developmental outcomes in children exposed to newer ASMs. Most cognitive outcomes are similar in children exposed to lamotrigine and levetiracetam and control children. The authors concluded there was limited and conflicting data when it comes to oxcarbazepine, gabapentin and topiramate. No studies have evaluated the neurocognitive outcomes in children exposed to eslicarbazepine, lacosamide, perampanel or zonisamide [46*]. More research into structural and cognitive teratogenesis with the newer ASMs is clearly needed.

Obstetric and Perinatal Outcomes

The MONEAD study prospectively examined a cohort of 351 pregnant WWE for severe adverse events. Severe adverse events occurred more often in pregnant WWE than in healthy controls (7.9% vs 1.9% respectively). While there was a trend towards higher MCMs and fetal loss in WWE, neither of these findings were statistically significant independently. This was likely related to the sample size of the study. Of note, the severe adverse outcomes were not significantly affected by periconceptional folate use, anti-seizure medication type or level [47*]. No differences were found in head circumference, gestational age at birth, 5-minute Apgar scores, or birth weight in infants born to WWE compared with controls. There was a higher risk of low birth weight in infants exposed to topiramate and higher risk of needing the neonatal intensive care unit for infants exposed to oxcarbazepine [48*].

Folic acid

The AAN practice guideline recommends at least 0.4 mg/day acid supplementation for any woman of childbearing age on an ASM [49]. Folic acid supplementation in the periconceptual period was associated with lower risk of MCMs in children exposed to anti-seizure medications in utero in one study [50]. The APR found no change in the rate of MCMs if WWE were taking folic acid. More data is needed to fully understand the role of folic acid supplementation on MCMs in pregnancy. Recent publications from several large cohort studies demonstrate the importance of folic acid supplementation on cognitive development. The NEAD study found that folic acid supplementation was associated with higher IQ scores in children exposed to ASMs [13]. This effect has been found in both 3 year and 6 year IQ, and does not seem to be related to specific ASMs [47*].

The Norwegian Mother and Child Cohort Study (MoBa) is a prospective cohort study based on parental surveys that investigated the relationship of autistic traits and folic acid. The MoBa study found a 5–8 time increased risk of having a child with autistic traits if folic acid was not taken around the time of conception. There was an inverse relationship between maternal plasma folate concentration and the severity of autistic traits. Mothers of children without autistic traits were more likely to start folic acid supplementation before pregnancy, while those with children with autistic traits were more likely to have started folic acid after conception [51]. Children exposed to anti-seizure medication had increased risk of language impairment and delay, but folic acid had a protective effect on language development [52, 53*].

Of note, the EBCR survey of more than 300 WWE found that only around 50% were taking folic acid [4]. Data from the NEAD study was retrospectively analyzed to evaluate if folic acid fortification in food was sufficient to increase childhood IQ. Folate from fortified food alone (without supplementation) was not associated with improved IQ, while periconceptual folate supplementation was associated with a 10-point improvement in age 6 IQ. Total folate intake combined from food and supplemental sources was associated with a higher age-6 IQ [54*]. This study was the first to evaluate the effect of folic acid in food sources, and confirms that supplementation should be recommended to all WWE.

While some folic acid is clearly recommended for WWE of childbearing age and WWE planning pregnancy, there is very limited data to guide the dosing. There is some evidence that too much folic acid may be problematic as evidenced by studies on pregnancy and folic acid in healthy women without epilepsy. In a general population study, 1–2 year old children exposed to high-dose (>5mg) of folic acid had lower mean psychomotor scores compared to children who had been exposed to lower doses [55]. Moreover, attention problems were noted in male children born to women taking >10 mg of folic acid per day during pregnancy [56*]. The optimal dose of folic acid is unknown. Common practice ranges from 0.4mg-5mg a day. Prenatal vitamins typically include 0.8mg (or 800mcg).

Breastfeeding

Breastfeeding promotes bonding, and has been associated with many benefits including decreased risk of breast and ovarian cancer and a lower risk of diabetes, infections and sudden infant death syndrome and positive effects on neurocognitive development in infants [57]. In a recent study, the rate of breastfeeding initiation was 30% lower in WWE than in other pregnant women. Only around half of pregnant WWE received counseling about breastfeeding from their neurologist, but those who did were significantly more likely to breastfeed their infants. A consultation with a lactation consultant improved the likelihood of continuing to breastfeed at 6 weeks’ time [58].

All ASMs can be transmitted into breast milk. Previously, the levels of anti-seizure medication that an infant is exposed to via breastmilk was not well established. Birnbaum et al. recently evaluated the serum anti-seizure medication level in 138 breastfed infants. Mothers were on carbamazepine (N=11), lamotrigine (N = 70), levetiracetam (N = 58), oxcarbazepine (N = 6) or zonisamide (N = 4). Overall, the serum ASM concentration in infants was low, ranging from 0.3 – 44.2% of maternal concentration. Maternal lamotrigine concentration was closely associated with the lamotrigine concentration in infants [59*]. This study supports that the benefits of breastfeeding likely outweigh the risks of infant exposure to anti-seizure medications for women being treated for epilepsy.

Conclusion

The care of WWE of childbearing age requires knowledge about contraception and pregnancy. Hormonal contraceptive methods can interact with anti-seizure medications, at times placing women at risk of unintended pregnancy, breakthrough seizures and teratogenicity. More research is needed though to fully understand the complicated relationship of epilepsy, ASMs and fertility. Anti-seizure medications are associated with teratogenic effects, and recent research has shown there can also be cognitive impacts of exposure to these drugs in utero. Valproate is associated with decreased IQ and language abilities, though data seems to be reassuring so far for newer medications like lamotrigine and levetiracetam. Folic acid supplementation seems to have a protective effect on cognitive teratogenicity. Recent updates highlight the importance of early counseling about folic acid supplementation and pregnancy planning. They also provide reassuring data on fertility and seizure control in pregnancy in women with epilepsy.

Key Points.

• There is limited data available with conflicting findings about the fertility rates of women with epilepsy

• Seizure frequency over time during pregnancy tends to remain stable compared to non-pregnant women with epilepsy

• Recent literature highlights the protective effects of folic acid supplementation on cognitive development for children born to women with epilepsy.