Epilepsy and antiepileptic medications

Classification of antiepileptic drugs

The following agents are older antiepileptic drugs: carbamazepine , clobazam , clonazepam , ethosuximide , phenobarbital , phenytoin , sultiame and VPA .

The newer antiepileptics include eslicarbazepine , felbamate , gabapentin , lacosamide , lamotrigine , levetiracetam , oxcarbazepine , pregabalin , rufinamide , tiagabine , topiramate , vigabatrin , and zonisamide .

2.10.1

Antiepileptic therapy

2.10.2

Antiepileptic and contraceptive drugs

Certain antiepileptic drugs can lead to contraceptive failure. Carbamazepine , phenobarbital , primidone , phenytoin , felbamate , oxcarbazepine and topiramate are inducers of the hepatic cytochrome P450–3A4 enzyme system. This enzyme system is involved in the metabolism of estrogen and progestogen. The resulting increased metabolism of contraceptives can give rise to unwanted pregnancies ( ). Thus it is not recommended to rely on hormonal therapy, including hormonal contraception, as even (as occasionally recommended) doubling the hormonal contraceptive dose will not guarantee prevention of pregnancy. An intrauterine device with local progestogen delivery would be preferable, or perhaps an intrauterine pessary, although somewhat less effective. Only if these methods cannot be tolerated, should a higher dose of hormonal contraception be considered, realizing the possible limitations of its reliability. In such cases, two daily doses of a low-dose monophasic formulation taken continuously for 3–9 months for each long-term cycle can be considered. Other recommendations aim for oral contraceptives with a higher dose to inhibit ovulation.

Effectiveness of hormonal contraceptives is not known to be impaired with the concomitant use of benzodiazepines , ethosuximide , gabapentin , lamotrigine , levetiracetam , pregabalin , tiagabine , VPA , vigabatrin and zonisamide .

However, estrogen-containing contraceptives can activate the degradation of lamotrigine, leading to an increased tendency for seizures without proper dosage adjustment, and, if appropriate dosing is initiated, after discontinuation of estrogen-containing contraceptives, there is potential for toxic side effects as the lamotrigine concentration increases ( ).

In sensitive individuals, sex hormones can increase (estrogens) or decrease (progestogens) the disposition for seizures. This is of relevance in cycle-dependent seizures.

2.10.3

Epilepsy and fertility

Epilepsy and antiepileptic drugs can decrease fertility. Thus there is, at the moment, an unclarified association between temporal lobe epilepsy and VPA therapy on one side and polycystic ovary syndrome (PCOS) on the other side. PCOS may cause anovulatory infertility and is found in 10–25% of epileptic women, and at even higher rates in patients taking VPA, while its prevalence in the general population is 5–10%. Adiposity with hyperinsulinism or insulin resistance seems to play a part in PCOS. Thus antiepileptic medications that enhance weight gain, such as VPA , carbamazepine , gabapentin and vigabatrin need to be viewed critically.

2.10.4

Frequency of seizures in pregnancy

Seizures may occur more frequently during pregnancy because the dose of antiepileptic medication is inappropriately decreased (e.g. intentional reduction or discontinuation of medication to protect the child), sleep disturbances, and higher clearance ( , , , , ). When antiepileptic drugs with higher clearance ( lamotrigine , carbamazepine , oxcarbazepine , levetiracetam , topiramate ) are dosed according to the patient’s blood level, the risk of seizures does not appear to increase (e.g. ). Further, it has been shown that the risk for seizures during pregnancy is only about 10%, where patients experience no seizures during at least 9 months prior to the pregnancy ( ).

2.10.5

Risk of malformations

Although well-established anticonvulsives belong to a class of medications that are among the most prescribed, and best investigated drugs with possible or proven teratogenicity, the determination of the individual risk of a patient remains a challenge ( ).

The classic drugs VPA , carbamazepine , phenobarbital , and phenytoin are proven teratogens, yet the malformations seen in numerous studies differ markedly. There is agreement that the highest risk has been observed with VPA. The teratogenic risk of newer antiepileptics has not been defined in reliable studies that confirm their risk or safety. Only for lamotrigine are sufficient data on hand to assure good tolerance by the unborn.

Well over 100,000 women with epilepsy have been analyzed in major studies in previous years, and from established registries for epilepsy and pregnancy in Europe, EURAP ( www.eurap.org ) and UK Epilepsy and Pregnancy Register ( ), in Australia ( ) and North America ( http://aedpregnancyregistry.org/ ). Major malformations were seen in 1.2–11% when monotherapy was used ( http://aedpregnancyregistry.org/ ; ). When combination therapy with several antiepileptic agents is used, the risk is on average higher than with monotherapy ( , , ), and clearly greater than 10% in combination therapy that includes VPA. These values are up to four times higher than those seen in corresponding control groups of healthy pregnant women. A meta-analysis encompassing more than 65,000 pregnancies from 59 studies and epilepsy registries, found that 17.6% of infants born to women with epilepsy using monotherapy with VPA had malformations (95% CI 5.25–30.03) ( ). For carbamazepine, the rate of affected infants was 5.7% (95% CI 3.71–7.65). The malformation rates for both substances were significantly higher than for infants born to non-exposed mothers. The rates were not significantly increased for lamotrigine, phenobarbital, and phenytoin, while no analysis was conducted for other agents due to low numbers of exposed cases.

A study from England with 277 pregnant epileptic women ( ) confirms the exceptionally high risk of VPA in mono- and combination therapies. While on average the major malformation rate of offspring of antiepileptic treated women was 6.6%, it was 11.3% when VPA was used alone and 16.7% when used in combination. Differentiating the daily dose above and below 1000 mg of monotherapy, the malformation rates were 16.0 and 7.1%, respectively; this difference was not significant due to the relatively low number of cases. Surprisingly, once the VPA–related cases had been excluded from the analysis, the authors found a malformation rate of only 3.0% for the cohort of pregnant women treated with all the other antileptic drugs. Carbamazepine was not found to have a significant elevation in risk for major malformations.

The North American Antiepileptic Drug Pregnancy Registry ( http://aedpregnancyregistry.org/ ) as of the Spring 2012 newsletter ( http://www2.massgeneral.org/aed/newsletter/Spring2012newsletter.pdf ), currently contains data on over 7,000 pregnancies and indicates that the risk for major malformations was 2.0% for monotherapy with lamotrigine, 3.0% for carbamazepine, 2.9% for phenytoin, and 5.5% for phenobarbital. Here, too, VPA clearly had a higher risk with 9.3% of monotherapy exposed pregnancies resulting in a major malformation.

In contrast to other published results, determined from the Australian Pregnancy Register of Antiepileptic Drugs in Pregnancy that combination therapies had a lower risk of malformation, and that the risk among polytherapy-treated pregnancies is, to a major degree, determined by the presence of VPA. According to their data, co-medication with lamotrigine can actually lower the risk of malformations in comparison to VPA given alone at the same dose. The authors discuss that lamotrigine enhances the elimination of VPA by degradation and glucuronidation. Accordingly, they challenge the general advice to use a monotherapy, especially when such treatment is not satisfactory. These results indicate that the avoidance of VPA is paramount.

Different malformation rates and relative risks in various studies can be explained by the methodological specifics of each study, the definition of (major) malformations, age of the child at the pediatric examination, as well as the composition of and malformation rates in the relevant control groups.

Two recent studies examined recurrence risk in women who took the same antiepileptic drug during two or more pregnancies. Using the Australian Register of Antiepileptic Drugs in Pregnancy, evaluated 2637 births among 1243 women, and found that the rate of malformation in the subsequent pregnancy was 35.7% in women who had a malformed infant in the first pregnancy. In contrast, women who had a non-malformed infant in the first pregnancy had a malformation rate of 3.1% if they stayed on the same drug (odds ratio 17.6; 95% CI 4.5–68.7). The recurrence risk was highest in those who had a malformed infant and remained on VPA. In the second study using the United Kingdom Epilepsy and Pregnancy Register, evaluated 1534 pregnancies among 719 mothers who stayed on the same drug through two or more pregnancies, and found that those whose first child was malformed had a 16.8% risk of having a second child with a malformation, compared to a 9.8% rate among women whose first child did not have a malformation (relative risk 1.73, 95% CI 1.01–2.96). The recurrent risk rose to 50% in a third pregnancy if a mother had two previous malformed children while on the same drug. In this study, VPA and topiramate were associated with higher recurrence risks.

2.10.6

Typical malformations and other anomalies

It is not possible to map certain malformation patterns to individual antiepileptic drugs save for a few exceptions ( ). There are malformations that are typical for VPA, such as neural tube defects, primarily lumbar spina bifida, and preaxial limb defects such as that of the radius. Specific developmental anomalies are described in the sections about the individual antiepileptic drugs.

The use of classical antiepileptic drugs primarily increases the risk of those anomalies that are also spontaneously more common ( ). These include cardiac defects, cleft lip and palate (frequency each about 2%), neural tube defects (for VPA and carbamazepine 1–2%), anomalies of the urinary tract, especially hypospadias, skeletal anomalies such as club foot or hip dysplasia, and eye anomalies (ptosis, iris coloboma).

The term fetal anticonvulsant syndrome (FACS) has been applied to the constellation of minor malformations and other adverse effects that are similar across the carbamazepine, phenytoin , and barbiturate embryopathies . A syndrome refers to all symptoms beyond the major malformations, such as dysmorphism, growth restriction, microcephaly, and mental dysfunctions. The milder anomalies and dysmorphisms as well as functional deficits include:

• ▪

  Midface hypoplasia (short nose, low and wide bridge of the nose or hypertelorism, epicanthus, long upper lip).

• ▪

  Anomalies of the distal phalanges (small nails, short terminal phalanges of the fingers, finger-like thumb).

• ▪

  Growth restriction.

• ▪

  Microcephaly (especially with phenytoin and with antiepileptic combination therapy).

• ▪

  Mental developmental dysfunctions, behavior problems, and suggestions of autism-like symptoms especially with VPA.

The detection of signs of dysmorphism is not always easy; it is based on subjective evaluation differences and sometimes requires a radiologic proof ( , ).

Typically only some and not all malformations or dysmorphisms will be present. A recent small study has suggested that tooth enamel defects may be more common in children prenatally exposed to antiepileptic medications ( ). In 38 children prenatally exposed to one or more anticonvulsant medications compared to 129 unexposed children in Denmark, 11% versus 4% had diffuse opacities, and numerous white opacities in the primary dentition, while 34% of exposed children versus 12% of unexposed had numerous white opacities in the permanent dentition.

Lymphocytes of the cord blood of offspring from mothers treated with antiepileptic drugs (primarily VPA and carbamazepine), demonstrated a significant increase of DNA damage as evidenced by presence of sister chromatid exchanges. Neither cytotoxic effects nor inhibition of cell-division kinetics were observed with these drugs ( ).

2.10.7

Pregnancy complications

Pregnancy complications can be increased in women taking antiepileptic drugs; however, data are contradictory. In a comprehensive evidence-based analysis of more than 285 studies, concluded that there was no substantially higher risk (>2) for caesarean sections, no moderately increased risk (>1.5) for premature uterine contractions and delivery, but possibly a markedly higher risk for premature uterine contractions and premature delivery for antiepileptic medication users who are also smokers. Data were considered insufficient to estimate the risk for preeclampsia, pregnancy hypertension, and spontaneous abortion. Possibly there is a higher risk for an Apgar value of <7 at 1 minute and a likely increased risk for SGA (small for gestational age). In contrast, a Norwegian investigation with about 2,900 pregnancies with epilepsy found that women using antiepileptic drugs (1/3 of the cohort) displayed more frequently mild preeclampsia, premature delivery, and children with a birth weight <2,500 g, a head circumference below the 2.5 percentile, or lower Apgar scores. The authors also found that, independent of the specific antiepileptic therapy, intrauterine growth restriction and caesarean section were more common ( , ). A Swedish study saw a smaller head circumference primarily with carbamazepine, and to some degree also with VPA, but not with other antiepileptic drugs ( ). Finally, a recent US study of 440 women with a seizure disorder were found to be no more likely to have a growth restricted infant or to experience stillbirth, preterm delivery or preeclampsia than women without a seizure disorder ( ).

2.10.8

Mental development dysfunction

CNS dysfunction is more common in children with midface hypoplasia. examined 57 children with a fetal anticonvulsant syndrome (FACS) and detected about 80% in behavior anomalies, speech impediments, and learning disorders; in 60%, two or more autistic features. examined neurodevelopment in 4–5 year old children born to mothers with or without epilepsy, who did or did not use antiepileptic medications in pregnancy, using parental questionnaires that were completed by 1,117 parents in Denmark. Behavioral problems as measured by the Strengths and Difficulties Questionnaire were more common in children born to mothers who took antiepileptic medication than either untreated eplileptic mothers or non-epileptics. When comparing different antiepileptic medications, developmental problems were noted primarily after prenatal exposure to VPA, and especially for autism spectrum disorders ( , , , , , ). When evaluating the relative effects of prenatal antiepileptic exposure in children at 3 years of age, found that all four of the commonly used anticonvulsants (VPA, carbamazepine, lamotrigine, or phenytoin) impair verbal versus non-verbal abilities. In a subsequent analysis at 4.5 years of age, the deficits persisted for all four drugs; however, there was statistically significantly lower performance on mean IQ in VPA exposed relative to carbamazepine , lamotrigine , or phenytoin ( ). At the 6-year follow-up mark, findings persisted, with VPA exposed children doing less well on measures of verbal and memory ability compared to those exposed to other drugs, and in particular at high doses of VPA ( ). In that study, mean IQs were higher in children prenatally exposed to periconceptional folate supplements.

2.10.9

“Damage mechanisms”

Different hypotheses have been proposed to explain the teratogenic effect of antiepileptic drugs; these are primarily derived from experimental studies. Thus a number of mechanisms can be invoked as possible explanation:

• ▪

  Carbamazepine , phenobarbital , and phenytoin can interfere with the uptake of folic acid or alter its metabolism by stimulating the cytochrome-P450 enzyme system. VPA inhibits glutamate formyltransferase and also decreases the production of folinic acid. A genetically determined deficiency of methylentetrahydrolate reductase may possibly be of relevance.

• ▪

  VPA inhibits the gene expression of histone deacetylase (HDAC). This enzyme participates in the control of the structure of nucleosomes. An HDAC deficiency results in a hyperacetylation of embryonal proteins particularly in the area of the caudal neural tube. Thus it represents a mechanism for the development of spina bifida that is independent of the folic acid pathway ( ). Experiments also demonstrated an inhibition of HDAC by topiramate and the main metabolite of levetiracetam ( ).

• ▪

  VPA causes changes in gene expression that regulates cell growth (e.g. brain-derived growth factor (BDGF) and nerve growth factor (NGF)) and the corresponding receptors.

• ▪

  A deficiency of the microsomal enzyme epoxide hydrolase in the mother and the embryo, leads to the accumulation of teratogenic epoxide metabolites in the presence of agents such as carbamazepine or phenytoin ( , ). These epoxide metabolites are produced by the enzyme monooxygenase that is linked to cytochrome-P450 and can bind to macromolecules. It can thus interfere with cell function and even lead to cell death ( ).

• ▪

  Phenytoin decreases the mRNA expression of several growth factors (e.g. TGF-β, NT3 and WNT1) ( ).

• ▪

  Phenytoin inhibits the potassium channel resulting in hypoxia and a subsequnet reoxygenation ( ).

• ▪

  Phenytoin has been implicated in the enhancement of gene expression of retinoic acid receptors in connection with a retinoic acid deficiency ( ).

• ▪

  VPA lowers the intracellular pH, for example in the limb buds (cited in ).

Clinical observations of familial accumulation of typical anomalies after use of antiepileptic drugs and results of gene sequence analyses, suggest that a genetic disposition is required for the teratogenic effect of damaging medications. An interplay of external (medication-related) and genetic factors was first discussed about 25 years ago in relation to a dizygotic pair of twins that had been exposed to phenytoin. One twin was healthy, while the other demonstrated the typical phenytoin anomalies although the intrauterine environment was the same ( ). Individual genetic patterns of metabolism can also explain the differences seen in a trizygotic triplet pregnancy, where the mother took phenytoin and phenobarbital. The three children displayed various degrees of intrauterine growth restriction as well as hypoplasia of the midface and the distal phalanges. One member of the triplet set also had a cleft lip and palate and another craniosynostosis ( ).

2.10.10

Folic acid and antiepileptic drugs

While a high dose of folic acid supplementation is recommended when folic acid antagonists are used in pregnancy, the proof of effective protection against the embryotoxic and teratogenic effects of antiepileptic medications, has not been demonstrated ( , ). A study conducted at the National Health Services Maternity Hospitals in Liverpool and Manchester, UK ( ), and the UK Registry for Epilepsy and Pregnancy ( ), also did not find a protective effect of a higher folic acid dose when compared to a standard dose. Generally, folic acid prophylaxis is recommended for all women when planning a pregnancy and during the first trimester. US recommendations, in a country with food fortification, are 0.4 mg/day ( Chapter 2.18 ). The general recommendation for epileptic patients who want to conceive, is that they be treated with folic acid supplements at a dose of 0.8 mg/day from before conception until the end of organogenesis (gestational week 10). The absence of any additional effectiveness of higher doses argues against greater supplementation. Additionally, it needs to be considered that folic acid enhances the drug metabolism of the hepatic hydroxylases so that the concentration of antiepileptic medication could be lowered in the mother. Provided nutritional balance is present, this and the lack of proof for additional benefit argue against the continuation of folic acid intake past the first trimester in epileptic women.

2.10.11

Vitamin K and antiepileptic drugs

Independent of the maternal medication, newborns and especially premature infants exhibit a deficiency of vitamin K that needs to be substituted right after birth to prevent bleeding problems. In addition, carbamazepine , ethosuximide , oxcarbazepine , phenytoin , phenobarbital , primidone , topiramate , vigabatrin and zonisamide belong to a group of medications that induce enzymes leading to a decrease in vitamin K-dependent clotting factors. The prothrombin precursor PIVKAII (protein induced by vitamin K absence or antagonist II) represents an indirect marker and can be elevated in the newborn ( ).

It has often been recommended that when a mother uses drugs that counteract vitamin K, she should be given vitamin K1 during the last 4 weeks of gestation, initially 10 mg/day, and in the last 2 weeks 20 mg/day. The effectiveness of this regimen remains controversial ( , ).

did not find a higher rate of bleeding complications in 667 newborns whose mothers had taken antiepileptic drugs (among them 463 caramazepine, 212 phenytoin and 44 phenobartital), in comparison with 1,324 children of healthy mothers. The mothers had not been given vitamin K during pregnancy, but all children received 1 mg vitamin K1 (preferably intramuscularly) at birth. In another study of about 200 children of mothers with antiepileptic therapy who had not received vitamin K prophylaxis during pregnancy, no increase in a bleeding tendency was apparent in the newborns exposed to antiepileptic drugs compared with the control infants ( ).

Vitamin K is absorbed orally as well as parenterally, but just after delivery the oral route may be unreliable, so that an IM injection of 0.5–1.0 mg vitamin K1 is recommended. This therapy appears to be superior to the oral administration, particularly for the prevention of late bleeding problems (after two weeks) ( ). If oral prophylaxis is chosen, it needs to be ascertained that the newborn actually swallows the dose.

2.10.12

Is epilepsy teratogenic?

According to the current state of knowledge, observed malformations in newborns are the result of antiepileptic therapy and not the epilepsy itself. However, this distinction is difficult to demonstrate conclusively, as only in minor forms of epilepsy can treatment be foregone. Some authors observed higher rates of malformation when mothers suffered a grand mal seizure during the first trimester ( ). described in a very small cohort a significantly increased risk of malformation when epilepsy was not treated (4/31 = 13%). Most other investigations did not find teratogenic effects, either with untreated epilepsy or with grand mal seizures during pregnancy. No distinguishing link has been illustrated between the duration of the antiepileptic treatment prior to pregnancy and the pregnancy outcome ( ). An analysis of the Belfast UK Epilepsy and Pregnancy Registry detected a 3.5% rate of major congenital malformations in 239 pregnancies of mothers whose epilepsy had not been treated, while the rate on average was 3.7% for those treated with monotherapy ( n = 2598) and 6.0% for those treated with polytherapy ( n = 770) ( ).

evaluated 10 studies in a meta-analysis covering 400 pregnancies of mothers whose epilepsy was not treated. They did not detect a teratogenic effect of epilepsy itself, but indicated that untreated epilepsy tends to occur in women with a less severe form of the disease, and with a lower frequency of seizures. Data from the Finnish Birth Registry were evaluated by which showed 26 malformations in 939 pregnancies corresponding to a unsuspicious malformation rate of 2.8%. examined 57 children of mothers who reported a history of epilepsy but were not treated nor suffered seizures during pregnancy. These children showed no impairment of intellectual development and no dysmorphism of face and fingers that are often seen after anticonvulsive therapy in pregnancy. , however, reported that verbal IQ was significantly more often lower (<70), where more than five generalized tonic-clonic seizures had occured during pregnancy, irrespective of any antiepileptic treatment.

2.10.13

Carbamazepine

Carbamazepine has structural similarities to tricyclic antidepressant medications and is used for grand mal seizures, focal and complex focal seizures, as a phasic prophylactic and mood stabilizer, and for trigeminal neuralgia. As with other antiepileptic agents, the anticonvulsive effect of carbamazepine is explained by its membrane-stabilizing ability.

Carbamazepine is well absorbed after oral administration, binds readily to proteins, and has a plasma half-life of 1–2 days. In the fetus, 50–80% of the maternal concentration is attained. The ratio of concentration to dose of carbamazepine decreases down to 40% during the third trimester and can necessitate a dose increase for adjustment ( ).

The effectiveness of oral contraceptives can be reduced by the marked induction of the cytochrome P450 enzyme ( Chapter 2.10.2 ).