Clinical Pharmacokinetics

, Volume 41, Issue 8, pp 559–579

Treatment of Epilepsy in Women of Reproductive Age

Pharmacokinetic Considerations


    • The Ohio State University College of Pharmacy
  • Gail D. Anderson
    • University of Washington College of Pharmacy
Review Article Special Populations

DOI: 10.2165/00003088-200241080-00002

Cite this article as:
McAuley, J.W. & Anderson, G.D. Clin Pharmacokinet (2002) 41: 559. doi:10.2165/00003088-200241080-00002


Although epilepsy affects men and women equally, there are many women’s health issues in epilepsy, especially for women of childbearing age. These issues, which include menstrual cycle influences on seizure activity (catamenial epilepsy), interactions of contraceptives with antiepileptic drugs (AEDs), pharmacokinetic changes during pregnancy, teratogenicity and the safety of breastfeeding, challenge both the woman with epilepsy and the many healthcare providers involved in her care. Although the information in the literature on women’s issues in epilepsy has grown steeply in recent years, there are many examples showing that much work is yet to be done. The purpose of this article is to review these issues and describe practical considerations for women of childbearing age with epilepsy. The article addresses the established or ‘first-generation’ AEDs (phenobarbital, phenytoin, primidone, carbamazepine, ethosuximide and valproic acid) and the ‘second-generation’ AEDs (felbamate, gabapentin, lamotrigine, levetiracetam, oxcarbazepine, tiagabine, topiramate, vigabatrin and zonisamide).

Although a relationship between hormones and seizure activity is present in many women, good treatment options for catamenial epilepsy remain elusive. Drug interactions between enzyme-inducing AEDs and contraceptives are well documented. Higher dosages of oral contraceptives or a second contraceptive method are suggested if women use an enzyme-inducing AED. Planned pregnancy and counselling before conception is crucial. This counselling should include, but is not limited to, folic acid supplementation, medication adherence, the risk of teratogenicity and the importance of prenatal care. AED dosage adjustments may be necessary during pregnancy and should be based on clinical symptoms, not entirely on serum drug concentrations.

Many groups have turned their attention to women’s issues in epilepsy and have developed clinical practice guidelines. Although the future holds promise in this area, many questions and the need for progress remain.

Epilepsy is a common neurological disorder. Yet, unlike some other neurological diseases, it is manageable and seizures can be well controlled in many cases or even eliminated in others. Even so, a recent report has estimated that 25% of patients with epilepsy experience uncontrolled seizures.[1] Pharmacotherapy with antiepileptic drugs (AEDs) is the mainstay of therapy for patients with epilepsy. The 1995 annual cost for the nearly 2.3 million patients in the US with epilepsy was estimated at $US12.5 billion.[2]

Epilepsy affects men and women equally, and it is estimated that nearly 1 million American women of childbearing age have epilepsy.[3] There are many women’s health issues in epilepsy, especially for those of childbearing age. These include menstrual cycle influences on seizure activity, interactions of oral contraceptives and AEDs, pharmacokinetic changes during pregnancy, teratogenicity of AEDs, and the safety of breastfeeding while taking AEDs. These issues challenge both the woman with epilepsy and the many healthcare providers involved in her care.

The information in the literature on women’s issues in epilepsy has grown steeply in recent years. The purpose of this article is to review these issues and describe practical considerations for women of childbearing age with epilepsy. Although there are many unanswered questions regarding the postmenopausal woman with epilepsy, that topic is beyond the scope of this review.

In some sections of this review, information on the established or ‘first-generation’ AEDs has been separated from information on the newer or ‘second-generation’ AEDs. The first-generation group includes phenobarbital, phenytoin, primidone, carbamazepine, ethosuximide and valproic acid. Second-generation AEDs include felbamate, gabapentin, lamotrigine, topiramate, tiagabine, levetiracetam, zonisamide and oxcarbazepine, all released since 1990.

1. Need for Information and Education

Four recent papers highlight the need for education of healthcare professionals in women’s issues in epilepsy. Krauss et al.[4] performed a national survey of neurologists and obstetricians and showed that many do not have accurate information about interactions between oral contraceptives and AEDs. In their survey, less than 5% of respondents accurately knew the pharmacokinetic drug interaction between six common AEDs (carbamazepine, ethosuximide, phenytoin, phenobarbital, primidone and valproic acid) and oral contraceptives.

Morrell and colleagues reported the results of the Epilepsy Foundation’s survey of healthcare professionals on their knowledge about women’s issue in five areas: hormone-sensitive seizures, fertility and contraception, pregnancy and contraception, sexuality and bone density.[5] Their data also demonstrate deficiencies in knowledge of women’s issues by professionals most likely to be involved with the care of women with epilepsy. Of the 3535 respondents, most knew that some AEDs interact with oral contraceptives, but most did not know which AEDs were the culprits. Half of the respondents were not certain of the frequency of birth defects caused by AEDs. The good news is that a majority of their respondents were willing to learn about these important issues.

Crawford and Lee surveyed women members of the British Epilepsy Association and asked them about the information they received from healthcare professionals on contraception and pregnancy.[6] Of the 1855 questionnaires returned, 51% claimed not to have received any advice about possible drug interactions between AEDs and contraception. When asked about the advice they had been given about pregnancy, a total of 59% claimed they had not received any advice or had not discussed pregnancy with anyone. A mere 7% reported having been given advice about the teratogenic effects of AEDs. This study demonstrates that women want, and need, more information and counselling about key women’s issues in epilepsy.

Fairgrieve et al.[7] reported some disquieting statistics on their cohort of pregnant women with epilepsy. For example, less than 50% of the women planned their pregnancy, nearly 25% reported contraceptive failure and less than 15% took folate appropriately. These data provide further evidence that there is much to be done to educate both the healthcare professional and the patient.

2. Menstrual Cycle Influences on Seizure Activity

Steroid hormones exert pharmacological effects in the central nervous system by various actions.[8,9] Many effects of steroid hormones are thought to be due to an interaction with intracellular receptors and gene-controlled changes in protein synthesis. Nevertheless, several investigators have reported that certain naturally occurring steroids modulate the GABAA receptor complex with a rapid onset of action, suggesting a mechanism other than a change in protein synthesis.[1013] These potent steroids include a metabolite of progesterone, one of the main steroids secreted by women throughout their reproductive years. Benzodiazepine binding to the GABAA receptor complex is enhanced by more than 50% in the presence of the ring-A reduced metabolite of progesterone, pregnanolone (3α-hydroxy-5α-dihydroprogesterone).[1417] The major blood and neuroendocrine metabolic conversion of progesterone is reduction by 5α-reductase and subsequent oxidoreduction to form pregnanolone. This neurosteroid, which increases the affinity of the GABAA receptor complex, can be formed endogenously by neurons and glial cells.[18,19]

Several observations suggest that steroids play a physiological role in regulation of brain excitability, including a role in epilepsy.[10,2022] Sex steroid hormones have been shown to significantly modify neuronal excitability; estrogen lowers the seizure threshold and progesterone increases seizure threshold. Pregnanolone has been shown to decrease epileptic activity in mice[23,24] and cats.[25] In the latter model, it was shown to be more potent than the benzodiazepine clonazepam. In rats, it has been suggested that neurosteroids are intricately involved in the sex differences in seizure susceptibility.[26]

Hormones can play a significant role in seizure activity.[2734] Some women with epilepsy experience more frequent or severe seizures before or during menstruation.[27,28,30,31] ‘Catamenial epilepsy’ refers to seizures influenced by the menstrual cycle.

2.1 Incidence

Estimating the number of women affected by this hormonal influence on seizure activity is difficult. Literature reports vary widely. A study of women in our outpatient epilepsy clinic found that approximately 50% of women self-reported an influence of their menstrual cycle on seizure activity.[35] The problems arise in defining catamenial.[36,37] There are many definitions, which may explain the variability in reported incidences of catamenial epilepsy.[28] Herzog et al. described the existence of three patterns of hormonal influence on seizure activity. [37] They examined seizure and menstrual data for one cycle from 184 women (aged 18 to 45 years) with intractable complex partial seizures, and divided the women’s cycles into ovulatory and anovulatory cycles based on midluteal serum progesterone concentrations. From their observations, the authors propose three patterns: (a) perimenstrual and (b) periovulatory during normal ovulatory cycles, and (c) the entire second half of the cycle (i.e. luteal) in anovulatory cycles. Although more than 70% of women in either category (anovulatory or ovulatory) had seizure exacerbation consistent with one of the three patterns, approximately one-third showed at least a doubling in average daily seizure frequency. Thus, they propose a 2-fold or greater increase in seizure exacerbation as a reasonable definition of catamenial epilepsy.

2.2 Mechanism

Some believe changes in hormones across the menstrual cycle to be the major factors responsible for catamenial seizure exacerbation.[28] Seizures can occur when the ratio of estradiol to progesterone is high during the days before menstruation and ovulation.[31] It is hypothesised that endogenous neurosteroids, including pregnanolone, may decrease when progesterone concentrations fall before menstruation, reducing the inhibitory effects of GABA and causing seizure exacerbation. Although no differences in mean plasma concentrations of neurosteroids were observed during the luteal phase between 15 women with catamenial epilepsy and 15 healthy controls,[38] measuring the peripheral concentrations of these neuroactive progesterone metabolites may not be indicative of the CNS situation.

Recently, a study by the same investigators examined the temporal relationship of serum pregnanolone after a seizure.[39] They reported data on ten seizures in seven different women of childbearing age with complex partial seizures immediately after the seizure and at 15 minutes and 6 hours after the event. They showed increasing serum concentrations of the neurosteroid within 15 minutes after the seizure, and then a decline at 6 hours in most patients. Although they speculated that there is a role of the neurosteroid in neuronal control, the neurosteroid concentration-time pattern within the CNS is yet to be discovered.

Another potential explanation for cyclical seizures is pharmacokinetic changes in AED metabolism across the menstrual cycle. In general, investigators have demonstrated a lack of menstrual-related effects on hepatic enzymes involved in the metabolism of AEDs. Of all the cytochrome P450 (CYP) isozymes, CYP3A4 is the primary isozyme regulated by hormonal factors. CYP3A4 is involved in the metabolism of over 50% of drugs and is also involved in the metabolism of endogenous steroids, including testosterone, progesterone, cortisol, androstenedione and estradiol.[40] In spite of fluctuating levels of steroid hormones during the menstrual cycle, no menstrual-related effects have been demonstrated clinically with a variety of CYP3A4 substrates, including carbamazepine,[41] alfentanil,[42] alprazolam,[43] nitrazepam[44] and midazolam.[45] At this time, only phenytoin, a CYP2C9 and CYP2C19 substrate, has been found to have decreased serum concentrations during menses.[46,47]

2.3 Treatment

Although relationships between hormones, AEDs and seizure activity have been shown, the associations are incompletely understood and unfortunately, there are no effective treatments available for these women. Backstrom et al. have shown that the number of seizures were significantly decreased during the luteal phase when progesterone concentrations are elevated as compared with the follicular phase.[48] Also, spike frequency was decreased in women with partial epilepsy after progesterone administration by continuous infusion.[49]

Attempts to treat catamenial epilepsy with medroxyprogesterone acetate or progesterone have been reported. A preliminary report showed an average 30% reduction in seizure frequency in seven women with epilepsy who developed amenorrhoea when using depot medroxyprogesterone acetate.[29] Problems in this small sample of women included irregular breakthrough bleeding and a lengthy delay in the return of normal menstruation after the cessation of therapy was reported. Herzog conducted a non-blind trial with natural progesterone lozenges 200mg three times daily as add-on therapy for 25 women with catamenial epilepsy.[32] Eighteen women (72%) showed a decline in average daily seizure frequency by more than 45% during 3 months of cyclical progesterone therapy as compared with the 3 months before therapy. The same author has recently published a 3-year follow-up study showing sustained success rates with oral progesterone.[50]

The few studies of adjunctive hormonal treatments for women with catamenial epilepsy identify some value and the need for further double-blind placebo-controlled studies. An antiepileptic drug in development, ganaxolone, may be beneficial to women with catamenial epilepsy because this drug is a synthetic derivative of the neurosteroid pregnanolone.[51,52]

3. Background to the Optimal Choice of Agents in Women with Epilepsy

Patients with epilepsy experience a wide variety of endocrine-related problems affecting pituitary, adrenal, thyroid, bone and sexual function. Both the disease state and the drug therapy have been implicated. Unfortunately, there is very little or no information regarding the incidence of endocrine disorders or the effect on sexual or reproductive function with the second-generation AEDs. As many of the endocrine effects may be due to the effect of the AED on the metabolism of endogenous hormones, the decreased propensity of the newer AEDs for drug interactions suggest that they may be better tolerated in women.[53]

3.1 Sexual/Reproductive Function

The effects of AEDs on serum sex hormones has mainly been studied in male patients; however, several studies have been published recently evaluating female patients. Patients with epilepsy have reduced fertility and suffer from hyposexuality more frequently compared with the average population.[5456] A larger number of women with epilepsy have anovulatory cycles as compared with control individuals.[48,57] As reproductive endocrine disorders have been shown to be associated with temporal lobe epilepsy[58] and primary generalised epilepsy,[59] it is difficult to determine the role of the AED. However, therapy with phenytoin, carbamazepine, phenobarbital and valproic acid is associated with a variety of changes in plasma concentrations of the circulating hormones (table I) and therefore may contribute to the disorders.
Table I

Effect of antiepileptic drugs on endogenous steroid hormones in female patients

AED therapy is associated with the onset of early puberty in both female and male patients.[63] Case reports describe a female child on valproic acid with pubertal arrest[64] and an infant girl with increasing gynaecomastia and galactorrhoea which resolved 3 days after valproic acid was stopped.[65] In a cross-sectional study, Isojarvi et al.[62] found that valproic acid was associated with an increased incidence of polycystic ovaries and hyperandrogenism with menstrual disturbances in 45% of patients, compared with 25% of patients on carbamazepine and 8% receiving all other AEDs. In this study, women receiving valproic acid had elevated plasma testosterone and dehydroepiandrosterone (DHEA) levels and a trend toward a decreased estradiol level, with no significant difference in luteinising hormone (LH) or follicular stimulating hormone (FSH). When 16 female patients with either polycystic ovaries or hyperandrogenism were switched to lamotrigine, their total number of polycystic ovaries, body mass index, fasting serum insulin level and testosterone level all declined significantly.[66] However, as valproic acid itself causes significant bodyweight gain and increased body mass indices, which can independently increase the risk of polycystic ovary, it is not a clear causal relationship. In contrast to the reports by Isojarvi et al., retrospective studies by Murialdo et al.[60] and Bauer et al.[67] did not find an increased incidence of polycystic ovary disease in valproic acid-treated women, in spite of their having higher bodyweight and increased body mass index compared with women treated with phenobarbital or carbamazepine. A prospective study of the endocrine adverse effects of valproic acid was strongly recommended, and a study is in progress in the UK.[68]

One study has addressed the effect of replacing carbamazepine with oxcarbazepine on serum sex hormones in male patients. Steroid hormone binding globulin (SHBG) concentrations decreased and DHEA sulphate concentrations increased with no change in the serum free and total testosterone, FSH, LH or prolactin, suggesting that oxcarbazepine may have less effect than carbamazepine.[69] In addition, as reported in a review article by Morrell,[53] lamotrigine and gabapentin do not appear to alter SHBG or adrenal or gonadal steroids.

3.2 Bone Disorders

Women are at a high risk of bone disease because of their smaller body size and the effects of the loss of estrogen at menopause. The risk of falls during seizures also can increase the risk of serious bone fractures in women with epilepsy. AEDs may contribute to increasing risk factors. In a prospective study evaluating the risk of hip fractures in women greater than 65 years old,[70] women currently taking AEDs were at a 2-fold higher risk of hip fracture. However, the epilepsy itself may be responsible for an increase in falls.

The mechanism of the effects of AEDs have not been established. Patients with epilepsy receiving phenobarbital, phenytoin and carbamazepine have hypocalcaemia, hypophosphataemia, increased serum levels of alkaline phosphatase activity and parathyroid hormone, and decreased serum levels of active vitamin D. This disorder has been called AED osteomalacia.[71] In a study of 226 outpatients with epilepsy, Gough et al.[72] determined that the incidence and extent of osteomalacia caused by AEDs are ranked as follows: polytherapy > phenobarbital > phenytoin = carbamazepine. Valproic acid was not associated with AED osteomalacia.

AED therapy is associated with decreased bone mineral density.[73] Duration of therapy with phenytoin and/or carbamazepine correlated with the bone mass density at lumbar spine and femoral neck region in 59 patients.[74] The hepatic enzyme-inducing properties of the AEDs has been shown to increase the metabolism of active vitamin D to inactive metabolites, which may be one contributing factor. However, decreased bone density also occurs with normal vitamin D metabolism, suggesting that the AEDs may also have a direct effect on bone cells.[75] Valproic acid, which is not an enzyme inducer, is also associated with decreased bone mineral density in children.[76,77] There is no information on the effects any of the new AEDs on bone function.

3.3 Lipid Abnormalities

Heart disease is the leading cause of death for women. Epidemiological data has linked elevated total and low-density-lipoprotein cholesterol (LDL-C) and reduced high-density-lipoprotein cholesterol (HDL-C) to the development of coronary heart disease in men and women.[7880] Epidemiological studies have also demonstrated that mortality due to atherosclerosis-related heart disease is lower in patients with epilepsy treated with AEDs than in the general population.[81]

A summary of the known effects of AEDs on lipids is given in table II. The increase in HDL-C by phenytoin, carbamazepine and phenobarbital has been suggested to be a positive factor.[82] As with the other endocrine effects of the AEDs, the hepatic enzyme-inducing properties of the first-generation AEDs is suggested to be involved. There is a correlation between the HDL-C concentration and hepatic microsomal enzyme activity and CYP content in liver biopsies.[83] In contrast, the increases in total cholesterol and LDL-C suggest a possible negative cardiovascular effect. Of the first-generation AEDs, only valproic acid does not increase HDL-C. However valproic acid does decrease total cholesterol and LDL-C.[82,8486] There is no information available on the new AEDs; however, the decreased hepatic enzyme induction by the second-generation AEDs suggests that the effects on lipids may be less.
Table II

Effect of antiepileptic drugs on lipid profile

4. Drug Interactions with Oral Contraceptives

Oral contraceptives are widely used among women of childbearing age to prevent unplanned pregnancies. Ethinylestradiol and mestranol, a prodrug that is inactive until converted to ethinylestradiol, are the two synthetic estrogens used in oral contraceptive products. The initial formulations contained doses of 50 to 100µg, but recent formulations contain less than 50µg with the majority of women using products containing 35µg of ethinylestradiol.

Ethinylestradiol undergoes significant first-pass metabolism with bioavailability of 40 to 50%, but interindividual variability ranges from 10 to 75%.[93,94] First-pass metabolism is due to both sulphation by gut metabolism and then glucuronidation or hydroxylation in the liver. The hydroxylation reaction is catalysed by CYP3A4. In two healthy individuals, the fraction of ethinylestradiol metabolised by CYP3A4 to either the 2- or 4-hydroxylated derivatives was determined as 36 and 64% of the dose.[95] The hydroxylated metabolites then undergo sequential conjugation with sulphate or glucuronide. Enterohepatic recirculation of conjugated ethinylestradiol, with hydrolysis to unconjugated ethinylestradiol, results in reabsorption of active ethinylestradiol. Drug interactions that interrupt the cycle can have profound effects on ethinylestradiol activity.

The synthetic progestational components of oral contraceptives are levonorgestrel, norethindrone, norethisterone, desogestrel, norgestimate and gestodene. They are also eliminated by hepatic metabolism. Unlike the estrogens, the progestogens do not undergo extensive first-pass gut or hepatic metabolism, nor do they undergo enterohepatic recirculation. Bioavailability is between 80 and 100% for the different progestogens. The isozymes responsible for their metabolism have not been elucidated. However, drug interaction studies suggest that CYP3A4 is involved.

Phenytoin, phenobarbital, primidone (via phenobarbital) and carbamazepine are broad-spectrum inducers of both the hepatic CYP system and the enzymes that form glucuronide conjugates, uridinediphosphate glucuronosyltransferases (UGTs). Valproic acid and ethosuximide are not enzyme inducers. Of the new AEDs, topiramate, felbamate and oxcarbazepine are weak inducers of only one of the CYP isozymes, CYP3A4, which metabolises oral contraceptives. Lamotrigine, gabapentin, tiagabine and zonisamide are not enzyme inducers. As shown in table III, the ability of the AEDs to induce CYP3A4 results in increased clearance and corresponding decreased plasma concentrations of the estrogen component and also of the progestational component of oral contraceptives. Therefore, valproic acid, ethosuximide, lamotrigine, gabapentin, tiagabine and probably zonisamide should not affect the efficacy of oral contraceptives in women concurrently taking these AEDs.
Table III

Effect of antiepileptic drugs on cytochrome P450 (CYP) 3A4 and metabolism of oral contraceptives

If a woman is receiving an AED with CYP3A4 enzyme-inducing properties (see table III) and is started on an oral contraceptive, she should be notified that the effectiveness of the oral contraceptive may be reduced. She should be told to report to her healthcare provider if any breakthrough bleeding occurs during the menstrual cycle. In women taking an enzyme-inducing AED, oral contraceptives containing 50µg of estrogen or an alternative or additional birth control method should be recommended.

The interaction is not limited to oral contraceptives; contraceptive failure has also been reported with implantable contraceptives and should be predictable with the new patches.[107,108]

5. Pregnancy

Because the woman and foetus are already exposed to the AED when pregnancy is confirmed, planned pregnancy is vital for women with epilepsy. This is especially important because of the potential for unplanned pregnancies from the interactions between AEDs and oral contraceptives discussed in section 4.

5.1 Optimal Prepregnancy Conditions

Although seizure control is desirable for all patients with epilepsy, it is especially favourable for a woman’s seizures to be well controlled before conception. Patient counselling should begin when pregnancy is first considered. Monotherapy is preferred whenever possible since the relative risk of birth defects dramatically increases with AED polytherapy.[109,110] Monotherapy also improves patient compliance, as does having a better understanding of birth defects caused by AEDs. The dosage of an antiepileptic drug should be at the lowest effective level to reduce the possibility of birth defects.[111] The gradual discontinuation of AEDs may be considered if a woman has been seizure-free for at least 2 years and has a normal EEG.[112] This subject is addressed in more detail in the practice parameter published by the American Academy of Neurology.[113] A report by Betts and Fox documents the value of a proactive pre-conception counselling service for women with epilepsy.[114] Items addressed in their pre-conception programme include the possibility of withdrawing AEDs, supplementation with folic acid and, more controversially, substitution of ‘lesser risk’ AEDs for current therapy. They compared their outcomes with a control group of women who were already pregnant at presentation and showed their programme to be very beneficial in preventing adverse outcomes.

5.2 Pharmacokinetic Changes in Pregnancy

Physiological changes in pregnant women may have an impact on the pharmacokinetics of AEDs (see table IV).
Table IV

Effect of pregnancy on the pharmacokinetics of the antiepileptic drugs

5.2.1 Absorption

Decreased gastric tone and motility due to pregnancy has not been shown to interfere with absorption of AEDs, but there have been few studies. Nausea and vomiting experienced by 40% of women during early pregnancy could interfere with the ability to ingest medications. Alternative means, such as rectal use of syrups or suspensions, may be necessary to provide coverage. Rectal administration of phenytoin,[120] carbamazepine,[121] valproic acid[122] and lamotrigine[123] has been demonstrated to provide adequate bioavailability.

5.2.2 Distribution

In pregnancy, plasma volume increases by 40 to 50%, predominantly because of increased total body water. As volume of distribution changes will only affect loading doses, this will only be a factor if therapy needs to be initiated parenterally with an AED during pregnancy. The binding capacity of albumin is decreased during pregnancy, resulting in decreased protein binding for highly protein bound drugs. The unbound fractions of phenobarbital, phenytoin and valproic acid increase with the decreased concentrations of albumin.[124126] In theory, for drugs predominantly metabolised by the liver with a restrictive clearance, decreased protein binding without changes in intrinsic clearance (metabolism) should result in a decrease in total drug concentrations, while unbound drug concentrations should remain unchanged. This includes carbamazepine and valproic acid. For drugs with both increased hepatic metabolism and decreased protein binding, for example phenytoin and phenobarbital, both total and unbound plasma concentrations will decrease, but not necessarily proportionately.

5.2.3 Hepatic Metabolism

There is evidence that pregnancy causes a differential effect on hepatic CYP isozymes (table IV). Hepatic metabolism of phenytoin increases during pregnancy[127] and both total and unbound plasma concentrations of phenytoin fall by 50 to 60% by the third trimester.[116] There is no significant decrease in unbound plasma concentrations for either carbamazepine or valproic acid; however, total plasma concentrations fall due to the decrease in albumin concentrations.[115,116,119] Although carbamazepine is metabolised by several CYP isozymes, CYP3A4 is the predominant isozyme responsible for the hepatic metabolism. Valproic acid is predominantly eliminated by glucuronidation and β-oxidation with only minor metabolism by CYP-dependent pathways. Therefore, it appears that the CYP2C enzymes (CYP2C9 and CYP2C19) may be more affected by pregnancy than CYP3A4. Both phenobarbital and valproic acid are metabolised by multiple pathways. Phenobarbital forms a glucoside conjugate, is renally excreted unchanged (25%), and undergoes CYP2C9- and CYP2C19-dependent p-hydroxylation (20%). Studies conducted by Chen et al. show that both total and unbound drug concentrations of phenytoin and phenobarbital are lower during the third trimester of pregnancy compared with the post-pregnancy period at least 1 month after delivery.[125] Of the second-generation AEDs, total lamotrigine plasma concentrations have been reported to decrease, but there is no information on unbound plasma concentrations.[118]

5.2.4 Renal Excretion

Renal function increases during pregnancy. Theoretically, total and unbound plasma concentrations of drugs that are predominantly renally excreted unchanged (gabapentin, levetiracetam, vigabatrin) will decrease. However, none of these are significantly protein bound, so total concentration is equivalent to unbound concentration. There have not been any reported studies of gabapentin, levetiracetam or vigabatrin. As plasma concentrations are not routinely monitored for these drugs, dosages should be adjusted based on clinical need.

5.3 Therapeutic Drug Monitoring

Due to the decreased protein binding, total plasma concentrations of phenytoin, phenobarbital, carbamazepine and valproic acid do not reflect unbound or active plasma concentrations. For phenytoin, measurement of unbound plasma concentrations can be used for therapeutic drug monitoring (TDM). For all the other AEDs, adjusting dosages on the basis of total plasma concentration is not recommended. Dosages should be adjusted based on clinical need only, and may not be necessary if the seizures are well controlled before pregnancy. In a study of 42 pregnant women receiving the first-generation AEDs, 49% did not require an adjustment in dosage, 38% required an increase in dosage and 6% required a decrease in dosage.[115]

5.4 Impact of Pregnancy on Seizure Activity

Pregnant women should be informed that the frequency of their seizures may change during pregnancy. Examining the results of a series of studies reviewed by Hauser and Hesdorffer, seizure frequency was unchanged in 50 to 70% of pregnant epileptic women, whereas 23 to 46% of women experienced an increase in frequency of seizures.[128] This increase is not associated with any particular seizure type or how long the woman has had seizures. Among other reasons, missing doses of AEDs, especially those with a short half-life, may be responsible for worsening of seizures during pregnancy.[129]

6. Teratogenicity

6.1 Incidence

The risk of birth defects in children of healthy women is about 2 to 4%, which rises to 4 to 6% in women with epilepsy taking one antiepileptic drug.[129] Women with epilepsy have a greater than 90% chance of having a healthy baby. Exposure to AEDs during the first trimester, especially the third to eighth week post-conception, presents the greatest risk of congenital malformation.[109,110,130] Although genetic susceptibility may play an important role in teratogenicity, the ‘fetal anticonvulsant syndrome’ has been described with phenytoin, carbamazepine, phenobarbital and valproic acid.[109,110,130,131] This syndrome includes lip and palatal malformation, congenital heart disease, and facial and digital anomalies. Neural tube defects such as spina bifida are specific to first trimester exposure to valproic acid and carbamazepine.[109,110,130,132] The risk of this defect is 1 to 2% with valproic acid and less than 1% with carbamazepine.[130] Table V lists the commonly used AEDs and the pregnancy categories assigned by the US Food and Drug Administration (FDA). Table VI provides a summary of studies reporting the incidence of AED-associated major congenital malformations in contrast with a control group.
Table V

US Food and Drug Administration-assigned pregnancy categories for the common antiepileptic drugs
Table VI

Summary of major congenital malformations in children exposed to antiepileptic drugs in utero

Investigators from Japan, Italy, and Canada have published two papers from a prospective analysis of 983 babies born to women with epilepsy in those countries over 13 years. Kaneko et al. reported an incidence of congenital malformation with drug exposure of 9.0% as compared with 3.1% in babies not exposed in utero to AEDs.[134] The incidences of malformation for mothers exposed to a single AED were: primidone 14.3%, valproic acid 11.1%, phenytoin 9.1%, carbamazepine 5.7% and phenobarbital 5.1%. There was no significant difference in the incidence of malformations after exposure to any single AED. As in other studies, their data show that malformations increase with polytherapy and high total daily doses. They also found a significantly higher risk of congenital malformation in babies born to mothers with a valproic acid dosage above 1000 mg/day and serum concentrations above 70 mg/L. Data on tobacco and alcohol consumption were not recorded.

In the second paper from this patient cohort, Battino et al. evaluated the risk of delay in intrauterine growth in babies of women with epilepsy.[135] The overall frequencies of head circumferences and neonatal weight below the tenth percentile were no different from expected. An association was found between small head circumference and exposure to polytherapy, or to monotherapy with phenobarbital or primidone. Also, differences in intrauterine growth were observed between countries, suggesting that genetic or environmental conditions may explain some of the heterogeneity of previous studies.

In an attempt to more closely examine specific exposure to AEDs, a recent report has focused on children with the fetal anticonvulsant syndrome.[143] Moore and colleagues reported their findings from 57 children with this syndrome. The majority (60%) of the children in their cohort were exposed to valproic acid alone, and one or more aberrant behaviours was reported in 46 of the 57 children. Another recent report provides more data by describing two siblings with valproic acid embryopathy.[144]

Hvas and colleagues reported their findings on their cohort of women with epilepsy.[140] They compared length of gestation and anthropometric measures in 193 pregnancies from 145 women with epilepsy with those in 24 094 pregnancies in women without epilepsy. Children of women with epilepsy who took AEDs during pregnancy had lower birthweight, body length and head circumference as compared with children of women without epilepsy. The outcomes were not only influenced by drug therapy, but also by lifestyle. Women who smoked, but did not have epilepsy, had a lower risk of pre-term delivery than did women with epilepsy who smoked.

A recent study examined the impact on physical and cognitive function in children born to mothers with epilepsy but who were not on AEDs during pregnancy.[145] This study demonstrated no difference in outcomes of children born to mothers with or without the disease, thus adding more evidence that AED therapy is the main teratogen.

6.2 First-Generation Drugs

Many different mechanisms for the teratogenicity of AEDs have been postulated. Phenytoin, phenobarbital and carbamazepine are metabolised via CYP-dependent oxidation. Oxidative intermediates are formed and further metabolised via hydroxylation by epoxide hydrolase, a hepatic cytosolic enzyme. Lower levels of this enzyme in foetuses as compared with adults may cause the accumulation of oxidative intermediates. The formation of oxidative intermediates is believed to be partly responsible for birth defects.[109,146148] Valproic acid inhibits the metabolism of oxidative intermediates and of folic acid.[130,148,149] The teratogenic effect is worsened with the addition of valproic acid to phenytoin, phenobarbital or carbamazepine. Combination therapy with valproic acid should be avoided during pregnancy, if possible.[148]

Interference with folic acid metabolism has been widely accepted as a mechanism of teratogenesis.[149,150] Folic acid is involved with the biosynthesis of DNA and RNA, and with the metabolism of certain amino acids. The precise role of folic acid in neural tube closure remains unknown. Valproic acid can cause neural tube defects and interfere with folate metabolism by inhibiting glutamate formyltransferase.[149]

The multicentre Medical Research Council (MRC) Vitamin Study, which included 1817 pregnant women carrying foetuses at high risk of neural tube defects, demonstrated that a 4mg daily dose of folic acid can prevent neural tube defects in pregnant women.[151] Ironically, women with epilepsy were excluded from this study. Investigations conducted by Dansky et al. have found that spontaneous abortion and developmental anomalies were significantly associated with folate deficiency caused by AED use.[150] In contrast, a case-control study of women receiving AEDs during pregnancy demonstrated that administration of folic acid did not reduce the risk of non-neural tube defects including oral clefts, cardiovascular and/or urinary tract defects.[152]

6.3 Second-Generation Drugs

Although the teratogenicity of the first-generation AEDs is well documented, conclusions regarding the teratogenic potential of the recently marketed AEDs cannot be made because of limited clinical data and inherent difficulty extrapolating the available preclinical animal studies.

6.3.1 Preclinical Studies

Even though animal reproductive studies are not sensitive and selective predictors of human teratogenicity, there is evidence to suggest that the second-generation AEDs may be less teratogenic. Intrauterine growth retardation and delayed skeletal ossification is found with both the first and second-generation AEDs and may be a consequence of maternal toxicity. The congenital malformations associated with the first-generation AEDs, including neural tube, orofacial, cardiovascular and urogenital defects, have not been found with the second-generation AEDS in the preclinical studies. In animal studies, growth retardation and limb agenesis has been demonstrated for topiramate[153] and growth retardation has been demonstrated with tiagabine.[153]

6.3.2 Clinical Data

There is little clinical data on the second-generation AEDs in pregnant women. An international lamotrigine exposure pregnancy registry, which included data from 1992 until September 1998, showed 8 birth defects in 123 first trimester exposures, or 6.5%.[154] The birth outcomes of this study do not show significant differences from birth outcomes in the general population.

Kondo et al. reported on the teratogenicity of zonisamide in 22 Japanese women.[155] Twenty-six offspring were exposed to the drug. In only four cases were the mothers on zonisamide monotherapy. Twenty-four deliveries were made and two artificial abortions were performed. Two malformations (anencephaly and atrial septal defect) were observed with zonisamide polytherapy, and none with monotherapy. Maternal serum zonisamide concentrations were below the therapeutic range in both cases and neither mother had any seizures during pregnancy. It is unknown whether either mother was receiving folic acid supplementation.

6.4 Folic Acid Supplementation

A recent exhaustive review assessed the effects of supplementation with multivitamins or folate on the prevalence of neural tube defects.[156] In their analysis of 6425 women, supplementation with folate reduced the incidence of neural tube defects, and did no harm. Mulitivitamins alone offered no protective effects. The appropriate dose of folic acid for women with epilepsy remains unknown. Further studies are required to confirm the dosage.

Notably, a case has been reported of a failure of folic acid to prevent a neural tube defect in a baby born to a mother taking valproic acid 2000 mg/day.[157] The woman reportedly took 4 mg/day of folic acid preconceptually and for the first trimester. The authors point out that no evidence exists that folic acid (in any dose) protects against neural tube defects caused by valproic acid.

6.5 Vitamin K Supplementation

The babies born to women with epilepsy and taking enzyme-inducing AEDs are at risk of haemorrhage due to decreased vitamin K-dependent clotting factors. Women taking carbamazepine, phenobarbital, primidone, and/or phenytoin should receive vitamin K 20mg every day from 36 weeks of gestation until delivery, and babies should also receive intramuscular vitamin K 1mg at birth.[158]

7. Breastfeeding

The benefits of breastfeeding are well-established, and include nutritional and cost advantages over formula, protection against infectious diseases and reduction in infant mortality as well as many other infant and maternal benefits.[159] In addition, the importance of the role of breastfeeding in enhancing the maternal bonding interaction between mother and infant is an important consideration. When a woman is receiving long-term medication, the risk of drug exposure to the infant needs to be weighed against the benefits of breastfeeding.[160]

All drugs transfers into milk to some extent. Drug in maternal blood distributes into breast milk by simple diffusion. The passage of drug into breast milk is determined by the physicochemical characteristics of the drug, including molecular weight, pKa, degree of lipophilicity and extent of plasma protein binding. The extent of protein binding of the drug is the single most important predictor of drug passage into milk.[161,162] For many drugs, the milk to plasma ratio (M/P) has been determined. The M/P is often used to calculate an exposure dose to the infant. As shown in table VII, for the AEDs there is a large interindividual variability in the ratio, presumably due to differences in volume and composition of the milk, and the ratio is not useful for predicting infant exposure. Table VII also provides a summary of the relationship between AED passage into milk and what is known about the accumulation in infants based on maternal protein binding characteristics. Unlike the M/P ratio, the percentage of drug protein bound in maternal plasma does provide valuable information on infant exposure during breastfeeding. There have been two recent reviews on AEDs and breastfeeding,[163,164] and the reader is referred to these publications for complete references on data in table VII.
Table VII

Passage of antiepileptic drugs into breast milk and accumulation in infants based on plasma protein binding characteristics

As shown in table VII, of the first-generation AEDs, only use of ethosuximide, primidone and phenobarbital by the mother result in measurable plasma concentrations in the infant. For phenytoin and valproic acid, maternal protein binding is high enough to decrease passage of the drug into the milk and limit the infant exposure. For carbamazepine and its active metabolite, carbamazepine epoxide, measurable concentrations can be found in the infant. However, serum concentrations in the infant are less than 20% of the maternal plasma concentrations, and significantly below the therapeutic range of carbamazepine. Therefore, carbamazepine is considered ‘probably safe’ for breastfeeding, but the infant should be followed for undue tiredness, poor suckling or vomiting.[160,164]

Except for tiagabine, the second-generation AEDs have low to moderate protein binding and they probably pass into breast milk and are absorbed by the infant. However, maternal and infant exposure information is only available for lamotrigine. The protein binding characteristic of lamotrigine (55% bound) would predict that the infants would attain clinically significant concentrations of lamotrigine. As predicted by the protein binding, a recent study of 10 women receiving lamotrigine during pregnancy demonstrated that the infants attained lamotrigine plasma concentrations of approximately 30% (range 23 to 50%) of the maternal plasma concentrations after breastfeeding for 2 to 3 weeks.[165] This is consistent with a previous case report.[166]

In summary, for the majority of the first-generation AEDs (carbamazepine, phenytoin, valproic acid), breastfeeding results in negligible plasma concentrations in the infant. Except for tiagabine, the protein binding characteristics of the newer AEDs suggest that passage into the milk will result in attainment of therapeutic plasma concentrations in the infant. For the second-generation AEDs, breastfeeding should be done cautiously and, if possible, the infant should be monitored for excess AED plasma concentrations and toxicity.

8. Women’s Health Resources

8.1 General

Information for patients and healthcare workers providing care to women with epilepsy is abundant. The Epilepsy Foundation in the US and the British Epilepsy Association are national organisations that work for people affected by seizures through research, education, advocacy and service. The Epilepsy Foundation has an ongoing programme entitled ‘Women and Epilepsy Initiative’. The Epilepsy Foundation can be reached by phone (1-800-332-1000) or the world wide web ( The British Epilepsy Association ( also has a focus on women’s issues. The American Epilepsy Society is a professional group that is dedicated to supporting individuals affected by epilepsy through research, education and advocacy. The American Epilepsy Society can be reached by phone (1-860-586-7505) or the world wide web ( In addition, the objective of the International League Against Epilepsy is to advance and disseminate knowledge concerning epilepsy worldwide.

Shafer has described a comprehensive list of topics to cover when counselling women with epilepsy,[167] and a new book in the ‘Brainstorms’ series by Schacter et al.[168] is now available.

8.2 Clinical Practice Guidelines

Best practice guidelines have recently been published by the American Academy of Neurology[169] and the United Kingdom’s Women with Epilepsy Guidelines Development Group.[6] The American College of Obstetrics and Gynecology have also published an educational bulletin on seizure disorders in pregnancy.[170] A Scottish national guideline on management of pregnant women with epilepsy, developed by a multidisciplinary group in accordance with rigorous methodology recommended by the Scottish Intercollegiate Guidelines Network and published in 1997, is available via the web.[171]

The publication of these clinical practice guidelines challenges healthcare professionals to implement change. Examples provided throughout this review show that much work is yet to be done. Also, as pointed out in a recent editorial by Wiebe, none of these guidelines and few others address issues on the cost of change in clinical practice and implementation, nor provide estimates of cost effectiveness.[172]

9. Practical Considerations

Women of childbearing age who have epilepsy face many challenges. These include menstrual cycle influences on seizure activity, interactions between contraceptives and AEDs, influence of disease and therapy on pregnancy, teratogenicity of AEDs, and the safety of breastfeeding. Many recommendations can be made on the basis of the available literature.

We suggest that our women patients who report a catamenial influence keep a seizure calendar that also incorporates their menstrual cycle in order to properly identify any pattern to their seizure activity. Treatment options for catamenial epilepsy are limited, although progesterone supplementation is a possibility. It is also important to note that women with irregular menstrual cycles, midcycle spotting and cycles shorter than 23 days or longer than 35 days may be having anovulatory cycles. Catamenial patterns may not be evident during anovulatory cycles.

Drug interactions between enzyme-inducing AEDs and contraceptives are well documented. Higher dosages of oral contraceptives or a second contraceptive method are suggested if women with epilepsy use an enzyme-inducing AED.

Planned pregnancy is highly recommended and counselling before conception is crucial. Prepregnancy counselling should include, but is not limited to, folic acid supplementation, optimal control of seizure activity, monotherapy with the lowest effective dosage of AED, and medication adherence. Information about protection against neural tube defects by folate should be made more widely available. Because of the high rate of unplanned pregnancies in our clinic, we suggest that each woman of childbearing age should take at least 1 mg/day of folic acid. Pregnant women without epilepsy should consume 0.4 mg/day of folic acid, but 4 mg/day of folic acid was administered in the MRC vitamin study. The exact folic acid supplementation dosage for women on AEDs is not known. Most prenatal vitamins contain 1mg of folic acid. Regardless, folate supplementation should be given to any woman considering conception and early in pregnancy to prevent birth defects. Information should be provided to the patient about the risk of teratogenicity and the importance of prenatal care.

Pregnant women with epilepsy are advised to receive prenatal care and good nutrition, while avoiding sleep deprivation, alcohol and caffeine. If these steps are followed, the risk of birth defects can be reduced. Some women may themselves discontinue, or reduce their dosage of, medication in the hope of preventing birth defects. Although this may reduce the risk of teratogenicity, it will increase the risk of recurrent seizures, especially generalised seizures that are more dangerous to the developing foetus. Ultrasound scans or α-fetoprotein determination during weeks 16 to 18 can detect neural tube defects.[111]

Dosage adjustments of AEDs may be necessary during pregnancy, and should be based on clinical symptoms, not solely on serum drug concentrations.

Additionally, patients should be encouraged to join a support group. In our clinic, we work very closely with the local affiliate of the Epilepsy Foundation. As pointed out by Russell et al.,[173] there is a need for epilepsy services to be accessible to obstetricians. In our clinic we work closely with a high-risk obstetrics clinic.

Providers are encouraged to enrol their patients into the AED pregnancy registry so more data can be accrued, especially for the second-generation agents. The AED pregnancy registry is the first North American registry for pregnant women taking any AED.[174] The registry can be contacted through its US toll-free telephone number (1-888-233-2334 or 1-888-AED-AED4) or on the world wide web (

The mother with epilepsy should be thoroughly educated about the care of her newborn. A comparison between mothers who received preconception and postdelivery counselling and those who did not revealed several accidents with baby care in the group not receiving counselling.[175] Some practical points include avoiding exhaustion, sharing care of the newborn at night, never bathing the child while alone, and feeding and changing the baby on the floor.

10. Conclusion

Although the future holds promise for many women’s issues in epilepsy, there remain many unanswered questions. These include the efficacy of drugs in women with catamenial epilepsy and the teratogenic effect, or lack of it, for the newer AEDs. The data summarised in this review suggest that there are special considerations in the selection of pharmacotherapy for women of childbearing age with epilepsy.


No sources of funding were used to assist in the preparation of this manuscript. There are no potential conflicts of interest directly relevant to the contents of this manuscript.

Copyright information

© Adis International Limited 2002