Abstract
In women with epilepsy, seizures can be influenced by variations in sex hormone secretion during the menstrual cycle. The proconvulsant effects of estrogen have been demonstrated in both animals and humans, whereas progesterone has been found to have anticonvulsant properties. Catamenial epilepsy affects approximately one-third of women with epilepsy. This type of epilepsy has generally been defined as an increase in seizure frequency beginning immediately before or during menses. However, three distinct patterns of catamenial epilepsy have been described: perimenstrual, periovulatory, and luteal. The diagnosis of catamenial epilepsy can be made through careful assessment of menstrual and seizure diaries and characterization of cycle type and duration. A variety of therapies for catamenial epilepsy have been proposed, including acetazolamide, cyclical use of benzodiazepines or conventional antiepileptic drugs (AEDs), and hormonal therapy. However, evidence for the effectiveness of these treatment approaches comes from small, unblinded series or anecdotal reports. Larger multicenter trials, as well as further investigation of the pathophysiology of the disorder, are needed to identify the most effective treatment for women with catamenial epilepsy.
The term “catamenial” is derived from the Greek word “katamenios,” meaning monthly. In ancient times, the cyclical nature of epileptic attacks was attributed to the cycles of the moon. Galen believed that the effect of the moon on the periodicity of epileptic seizures depended on “the greater or smaller share it received from the sun; its effects were weak at half moon, but strong at full moon.”1 Antyllus, an older contemporary of Galen and one of the greatest surgeons of antiquity, wrote, “ … the moon rather moistens. And for this reason it makes the brain relatively liquid and the flesh putrid and renders the bodies of people who live in a clear, cold air moist and dull and for the same reason stirs up heaviness in the head and epilepsies.”1 During the Middle Ages, a vapor arising from the uterus was believed to induce epileptic attacks.
At a meeting of the Royal Medical and Chirurgical Society in 1857, Sir Charles Locock first described the relationship between epileptic seizures and the menstrual cycle.2 Hysterical epilepsy (from the Greek “hystera,” meaning “uterus”) “was confined to women and observed a regularity of return connected with the menstruation.” In 1881, Gowers described the first series of menstruation-related seizures, affecting 46 of 82 women.3
The menstrual cycle.
The normal menstrual cycle is depicted in figure 1. The average interval between menstrual periods is 28 days during the reproductive years, increasing at either end of reproductive life. Cycles between 24 to 35 days are considered normal. By convention, day 1 of a 28-day cycle is the first day of menses and ovulation occurs on day 14, since the duration of the luteal phase is relatively constant at 14 days. However, a luteal phase that is between 11 and 17 days from the luteinizing hormone (LH) surge is considered normal.
Figure 1. The normal menstrual cycle. Reproduced with permission from Speroff et al. 1999.137
The human menstrual cycle is the expression of a complex neuroendocrinologic system known as the hypothalamic–pituitary–ovarian axis. This system regulates the interactions among neurohormones, gonadotropin-releasing hormone (GnRH), pituitary gonadotropins, and the gonadal steroids through a feedback loop mechanism (figure 2). GnRH is synthesized in the mediobasal hypothalamus in neurons within the arcuate nucleus. This hormone is secreted in a pulsatile manner from nerve terminals at the median eminence into the portal system and delivered to the anterior pituitary. Normal menstrual function is dependent on the pulsatile secretion of GnRH within a critical, narrow range of amplitude and frequency. The intrinsic pulsatile activity of GnRH neurons is influenced by a number of central modulators, such as catecholamines and endogenous opioids. In the anterior pituitary, GnRH stimulates the pulsatile secretion of follicle-stimulating hormone (FSH) and LH. LH and FSH pulses are characterized as low amplitude and high frequency during the midfollicular phase and high amplitude and low frequency during the midluteal phase. This pulsatile secretion is critical for correct follicular development which, in turn, is responsible for a normal luteal phase. The pituitary gonadotropins regulate the production of the gonadal steroids estrogen and progesterone, which modify the release of the gonadotropins through feedback on the anterior pituitary cells. FSH also stimulates the release of a peptide, inhibin, by the granulosa cells of the ovary. This peptide inhibits FSH but not LH secretion. There are three biologically active estrogens: estradiol, estrone, and estriol. In nonpregnant women, estriol is considered a metabolite of estradiol and estrone that comes from peripheral conversion rather than direct ovarian production. Estradiol is the most potent and is highly lipophilic, making it capable of crossing the blood–brain barrier.
Figure 2. The hypothalamic–pituitary–ovarian axis.
During the follicular phase, FSH promotes follicle maturation in the ovary and stimulates the biosynthesis of estradiol. LH stimulates the production of androgens by the theca cells in the ovary which are aromatized into estradiol by the granulosa cells. These cells express aromatase activity under the influence of FSH. The pituitary has a complex feedback relationship with estradiol. Initially, estradiol has a negative feedback action on pituitary secretion of gonadotropins. As estradiol increases during the midcycle, a critical concentration is achieved for a critical period of time that has a positive feedback action on the pituitary, resulting in a surge of LH. The LH surge is responsible for the completion of oocyte maturation, initiation of the events that lead to ovulation, and conversion of the follicle into the corpus luteum. This marks the end of the follicular phase and precedes ovulation by approximately 36 hours. Ovulation itself consists of a gradual release of an oocyte. During the luteal phase, the dominant follicle evolves into the corpus luteum and progesterone production increases. If pregnancy does not occur, the corpus luteum regresses, progesterone and estradiol production declines, and an increase in FSH release heralds a new cycle of folliculogenesis. The abrupt withdrawal of estradiol and progesterone results in endometrial hemorrhage and tissue sloughing.
Abnormal FSH secretion during the follicular phase leads to diminished follicular development and to inadequate formation and function of the corpus luteum, a condition known as the inadequate luteal phase (ILP).4,5⇓ The corpus luteum is defective in progesterone production, although the estrogen-producing function remains unimpaired. This disorder has a variety of etiologies, including primary ovarian defects, central defects of the hypothalamic–pituitary axis, primary metabolic defects, and specific defects in luteal cell steroidogenesis. Menstrual cycle duration is variable, and cycles may be unusually short or long. ILP cycles occur in as many as 29% of women. This is an uncommon cause of infertility.6
The menstrual cycle is characterized by a dynamic interaction of brain neuroendocrine glands and the reproductive end organs. The marked fluctuations in gonadal steroids and peptides throughout a menstrual cycle clearly have a direct effect on brain function.
Pathophysiology.
Hormonal influences. Although a variety of mechanisms of catamenial epilepsy have been proposed, the literature strongly indicates that hormonal influences are the best established. Seizures are influenced by the physiologic variation in sex hormone secretion during the menstrual cycle. Both estrogen and progesterone exert significant effects on seizure threshold. In various animal species, dense populations of estrogen receptors are found in the hypothalamus and limbic system, including the amygdala, hippocampal formation, and entorhinal cortices.7-10⇓⇓⇓ Progesterone receptors have a more widespread distribution, with notably high density in the brainstem and spinal cord.11 The effects of ovarian steroids on neuroendocrine and reproductive function are mediated through several cellular mechanisms. Estrogen and progesterone affect neuronal function by activating specific intracellular receptors that modulate transcription and protein synthesis.12 At least two estrogen receptors have been identified: ERα mRNA is widely distributed in the CNS and reproductive organs of both females and males, whereas ERβ mRNA is more widely distributed in the female brain.12 In addition, the ovarian sex steroids influence electrical excitability, synaptic functioning, and neuronal morphology.
The proconvulsant effects of estrogen have been demonstrated in a variety of animal models and in humans. Intravenous and topical administration of estrogen induces epileptiform activity and seizures in normal and lesioned animals.13-15⇓⇓ Estradiol facilitates kindling in the amygdala,16,17⇓ hippocampus,18 and neocortex,19 and increases audiogenic seizures in female rats.20 Estradiol potentiates pentylenetetrazole (PTZ)- and kainic acid-induced seizures21-24⇓⇓⇓ and reduces threshold to maximal electroshock (MES) seizures.25,26⇓ In the rat hippocampus, estradiol significantly decreases seizure threshold during estrus27 and exerts glutaminergic effects in the CA1 region that facilitate limbic seizures.10 Binding for the N-methyl-d-aspartate (NMDA) receptor complex28 and dendritic spine density of hippocampal CA1 pyramidal cells increase,29 thereby enhancing sensitivity of CA1 neurons to NMDA receptor-mediated synaptic input. Estradiol increases synaptic connectivity among individual presynaptic boutons and multiple postsynaptic CA1 pyramidal cells, potentially increasing the synchronization of epileptic activity in the hippocampus.30
Recent studies suggest that estradiol also regulates inhibitory function in the CA1 region of the hippocampus. Levels of mRNA for glutamic acid decarboxylase (GAD), the rate-limiting enzyme for GABA synthesis, increase in GABAergic neurons in the CA1 pyramidal cell layer after treatment with estradiol; progesterone reverses this effect.31 Treatment of hippocampal cultures with estradiol decreases GAD immunoreactivity and concomitantly reduces the amplitude and frequency of GABAergic miniature inhibitory postsynaptic currents and increases excitatory postsynaptic currents.32 This decrease in GABAergic activity leads to an increase in pyramidal cell spine density through disinhibition of pyramidal cells.32
The anticonvulsant properties of progesterone were reported in rats more than six decades ago.33 In experimental seizure models and in humans, the effects of progesterone appear to directly oppose those of estrogen. Progesterone and its metabolites enhance inhibitory responses to GABA in a manner similar to the benzodiazepines (BDZs) and barbiturates, by acting at a specific steroid recognition site on the GABA/BDZ receptor–chloride ionophore complex.33-38⇓⇓⇓⇓⇓ Progesterone depresses the amplitude and frequency of interictal spikes generated by application of penicillin to the cerebral cortex of cats39 and has a protective effect on chemically induced convulsions in dogs.40 Progesterone slows kindling,17,41⇓ elevates the electroshock seizure threshold,25,26,42,43⇓⇓⇓ and is as effective as phenobarbital (PB) and phenytoin (PHT) against PTZ-induced seizures in rats.42 Progesterone protects female rats against kainic acid-induced seizures22 and reduces dendritic spines on hippocampal CA1 pyramidal cells after estrogen administration.29 At significantly lower concentrations than PB, progesterone potentiates recurrent inhibition in the CA1 region of the rat hippocampus, leading to neuronal hyperpolarization.44 Increased susceptibility to audiogenic seizures is observed after progesterone withdrawal in rats.45 In contrast to these anticonvulsant effects, progesterone was found to enhance spike–wave discharges in a genetic model of absence epilepsy.46
Several investigators have recently shown that the anticonvulsant properties of progesterone are due to conversion to its 5α-reduced metabolite, 3α-hydroxy-5α-pregnan-20-one (3α,5α-THP, allopregnanolone). This compound is a neuroactive steroid that, like the barbiturates, enhances the frequency of opening of the GABA-Cl-ionophore and possesses anxiolytic properties similar to those of the benzodiazepines.47-50⇓⇓⇓ In vitro, allopregnanolone causes a significant depression of spiking in the CA1 region of the guinea pig hippocampus.51 After sustained exposure to progesterone, withdrawal of 3α,5α-THP increases anxiety and seizure susceptibility,49 decreases the behavioral response to lorazepam, and produces insensitivity to the potentiating effects of lorazepam on GABA-gated Cl currents in the CA1 region of the rat hippocampus.52 These effects have been attributed to an alteration in the expression of GABAA receptor subunits, because 3α,5α-THP enhances the transcription of the gene encoding the α4-subunit of the GABAA receptor.50,52⇓
Recently, Reddy et al.53 proposed the first model of catamenial epilepsy. After induction of persistently elevated progesterone in female rats, allopregnanolone was abruptly withdrawn by blocking its conversion from progesterone using the 5α-reductase inhibitor finasteride. Plasma allopregnanolone levels were reduced by 86% within 24 hours, whereas progesterone levels were unaffected. Acute withdrawal of allopregnanolone produced an increase in PTZ-induced seizure susceptibility, but long-term treatment with finasteride did not. In the same model, administration of allopregnanolone and other neurosteroids before finasteride treatment effectively protected against PTZ-induced seizures.54 This was in contrast to PB, which had a modest effect, and to valproic acid (VPA), diazepam, and the partial BDZ receptor agonist bretazenil, which were ineffective. Treatment with the synthetic neuroactive steroid ganaxolone (3α-hydroxy-3β-methyl-5α-pregnane-20-one), an orally active analogue of allopregnanolone, during this period of enhanced seizure susceptibility protected against PTZ-induced seizures.55 A recent case report of a women with catamenial epilepsy, polycystic ovarian syndrome, and anovulatory ILP cycles provides support for the concept that the anticonvulsant effects of progesterone may be mediated in part by allopregnanolone.56 Treatment with progesterone during the luteal phase normalized menses and luteal progesterone levels and produced a significant reduction in seizures. The introduction of finasteride for the treatment of male pattern baldness led to recurrent seizures, which then remitted once the drug was discontinued. Interestingly, two synthetic neuroactive steroids, alphaxalone (5α-pregnane-3α-ol-11, 20-dione) and tetrahydrodeoxycorticosterone, exacerbated seizures in a rat model of generalized absence epilepsy when administered into the ventrobasal nucleus of the thalamus.57 These agents neither prolonged nor shortened seizures when administered to the thalamic reticular nucleus, a finding attributed to the molecular heterogeneity of GABAA receptor subunits in the thalamus.
Although far fewer, studies in humans also demonstrate the opposing effects of estrogen and progesterone on seizure susceptibility. Intravenous estrogen facilitated interictal epileptiform activity in 11 of 16 women with epilepsy (63%). Seizures were provoked within minutes to hours in four cases.58 Conversely, progesterone infusions significantly reduced spike frequency in four of seven women with focal epilepsy.59 In transcranial magnetic stimulation (TMS) studies of healthy women, increased inhibition/decreased facilitation during the early follicular and luteal phases and decreased inhibition/increased facilitation during the late follicular phase are observed, supporting the inhibitory effects of progesterone and the excitatory effects of estradiol in humans.60,61⇓ Using TMS, Herzog et al.62 demonstrated a significant perimenstrual reduction in cortico–cortical inhibition that normalized with progesterone in a woman with seizures related to menses. Serial EEG recordings demonstrated a marked increase in spike–wave discharges during menstruation compared with other phases of the menstrual cycle in a woman with idiopathic generalized epilepsy.63
Bäckström,64 almost three decades ago, was the first to systematically study the relationship between seizures and sex steroid levels. During nine cycles of seven women with epilepsy, estradiol and progesterone levels were measured on alternate days, and seizure frequency and the estrogen-to-progesterone (E:P) ratio were estimated daily. In six ovulatory cycles, a positive correlation between seizure frequency and the E:P ratio was observed, peaking in the premenstrual and preovulatory periods and decreasing during the midluteal phase. This correlation was stronger for generalized motor seizures than for focal seizures but was present in both. In three anovulatory cycles, seizure frequency correlated positively with estradiol levels; progesterone levels were low or undetectable. However, another investigator found no consistent correlation between average daily seizure frequency and hormonal levels in one woman with perimenstrual seizures.65 In nine women with seizures unrelated to menses, there was no correlation between hormone and AED concentrations, suggesting that the relationship between hormones and seizures in women with catamenial tendencies may not be affected by changes in AED concentrations.66
Other studies support the existence of a luteal progesterone deficiency in women with perimenstrual seizures.67-73⇓⇓⇓⇓⇓⇓ Luteal progesterone levels were significantly lower in women with perimenstrual seizures treated with AEDs than in untreated women and controls, whereas levels of LH, FSH, prolactin, and estrogen were similar among groups.67 However, EEG activity was similar during the menstrual and luteal phases in these women.68 In another series, luteal allopregnanolone and pregnanolone (3α-hydroxy-5β-pregnan-20-one) concentrations were comparable in women with catamenial temporal lobe seizures and controls, although the authors contend that plasma levels are a poor reflection of CNS effects.74
Water balance.
Hippocrates first suggested a relationship between fluid balance and seizures by noting that the brains of individuals with epilepsy were “unusually moist.”75 It was later recognized that cerebral edema was associated with seizures in a variety of settings, including eclampsia, uremia, trauma, and acute alcoholism.76 Almost a century ago, direct drainage of subarachnoid fluid was performed with some success for the treatment of epileptic seizures.77 These observations led to a series of experiments in the early twentieth century investigating the effect of water ingestion on seizures in patients with epilepsy. Excessive water ingestion and the antidiuretic hormone vasopressin provoked seizures in these patients.78 These findings were not observed in normal subjects. Negative water balance produced by fluid restriction had the opposite effect.78,79⇓ Because the addition of salt sufficient to prevent dilution also prevented seizures, it was proposed that dilution of the extracellular space producing a state of water intoxication was required to induce seizures. Others later demonstrated that the administration of vasopressin (Pitressin) and copious amounts of water, known as the Pitressin test, provoked seizures and EEG abnormalities in patients with epilepsy.80,81⇓ Jacobsen82 recommended that the test be used to establish the diagnosis of epilepsy when the history was inconclusive. During the same time period, repeated withdrawal of spinal fluid in epileptic patients produced a significant and sustained reduction in generalized motor seizures.76 These early experiments suggested that neuronal cell membrane permeability was defective in epilepsy.
Because an association between menstruation and edema was already recognized, these experiments led to the theory that water imbalance was a primary mechanism of perimenstrual seizures. However, Ansell and Clarke83 found no significant differences in body weight, sodium metabolism, or total body water between 14 epileptic patients, including seven women with perimenstrual seizures, and 10 healthy controls, or between epileptic women with and without catamenial tendencies. Nor was there a correlation between the peak incidence of seizures and the period of maximal edema. Considerable variation in body weight across the menstrual cycle was observed. Premenstrual symptoms were no more common in women with epilepsy than in controls and were equally prevalent among epileptic women with and without catamenial tendencies. Other investigators also failed to demonstrate a correlation between body weight and seizures over the course of a menstrual cycle.64 Interestingly, premenstrual administration of the angiotension-converting enzyme (ACE) inhibitor captopril improved seizure control in a woman with perimenstrual seizures.84 This effect was attributed to the cyclic increase in circulating renin, present at high concentrations in ovarian thecal cells, leading to changes in salt balance over the course of the reproductive cycle.
Antiepileptic drug metabolism.
Gonadal steroids are actively metabolized in the liver. Most enzymes involved in gonadal steroid hormone production involve the cytochrome P450 group of oxidase enzymes. The extent to which drugs that stimulate hepatic metabolism directly affect the serum concentration of sex steroids or the dynamic process of endocrine changes during a menstrual cycle is unknown. Although some women with epilepsy taking AEDs have anovulatory cycles, this may reflect a central phenomenon rather than altered metabolism of gonadal steroids alone. The same appears to be true for the reverse hypothesis. Normally secreted gonadal steroids usually do not significantly influence the clinical efficacy of medications metabolized in the liver. Women with clinically significant fluctuations in AED concentrations induced by the menstrual cycle may represent a small subgroup of women with epilepsy; however, this requires further study. These observations do not apply to pharmacologic steroids, including oral contraceptives. Hepatic enzyme inducers will accelerate the metabolism of sex steroids and lead to oral contraceptive failure. Sex steroid drugs may induce alterations in clearance rates that may also modify therapeutic levels of some medications.
Rosciszewska et al.71 measured serial AED levels across a menstrual cycle in 64 women receiving PHT alone or in combination with PB. Women with catamenial seizures had lower AED concentrations despite receiving higher doses of the drugs. The mean PHT concentration was significantly lower on day 28 of the menstrual cycle in women with catamenial seizures than in those with seizures unrelated to menses (24.5 versus 42.2 μg/mL). Women with catamenial seizures were also more likely to have a 30% or greater reduction in PHT levels (55% versus 14%). Serum PB concentrations did not vary significantly. Estrogen also remained relatively stable across the cycle, suggesting that catamenial seizures are not the result of higher estrogen concentrations during the premenstrual phase. Urinary progesterone metabolites were lower during the premenstrual phase in women with catamenial seizures, supporting the protective effect of progesterone on seizure susceptibility.
Others have described similar fluctuations in AED concentrations across the menstrual cycle.85-89⇓⇓⇓⇓ Phenytoin levels were significantly lower during menses than during the periovulatory period in 16 of 17 women with perimenstrual seizures.85 The mean change was greater (4.1 versus 1.6 μg/mL) in women with perimenstrual seizures than in those with seizures unrelated to menses. Phenytoin clearance increased to a greater extent (50% versus 26%) in women with perimenstrual seizures. The authors suggested that PHT metabolism is slowed at ovulation because of competition from increased levels of circulating steroid hormones. In another series, women with catamenial seizures had a significantly greater difference in PHT concentration between menses and ovulation than women with randomly distributed seizures (3.44 versus 0.91 μg/mL).86 Although the difference did not reach statistical significance, PHT clearance was greater during menses compared with ovulation. In two women with perimenstrual seizures treated with PHT and VPA in monotherapy, PHT levels varied by as much as 20% and VPA levels by as much as 35%, with the lowest levels recorded during the week of menses, correlating with peaks in seizure frequency.87 Fluctuations in salivary PHT and carbamazepine (CBZ) concentrations have been reported in women with perimenstrual seizures.88 Interestingly, among women with seizures unrelated to menses, serum concentrations of PHT, PB, and CBZ were stable across the cycle.66
Definition and incidence.
The reported incidence of catamenial epilepsy varies from 10% to 78%, due largely to methodological differences between studies (table 1). 3,72,83,90-104⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓ Catamenial epilepsy is often vaguely defined as the occurrence of seizures around the time of menses or an increase in seizures in relation to the menstrual cycle. Many of these studies rely on self-reports or seizure diaries over a single cycle or are limited to institutionalized or medically refractory women. In most cases, information regarding seizure type and epilepsy syndrome is not provided. Patient perceptions of how seizures relate to menses are often inaccurate.94,98,101,105⇓⇓⇓ Duncan et al.101 found that 78% of women they studied claimed to have catamenial seizures, but only 12.5% fulfilled the criteria by having at least 75% of seizures over the 10 days beginning 4 days before menses, representing a sixfold increase in average daily seizure frequency. In another series, only 24% of 21 women who reported increased seizures surrounding menses exhibited this tendency.105 Catamenial tendencies may be more common among women with focal epilepsy compared with those who have generalized epilepsy.97,102⇓ However, this has not been adequately studied.
Incidence of catamenial epilepsy
From its earliest description, catamenial epilepsy has referred to an increase in seizures beginning immediately before or during menses.3,106,107⇓⇓ Although this is clearly the most prevalent pattern,97,98,101,104,105,108-110⇓⇓⇓⇓⇓⇓⇓ some women have seizures occurring in a cyclic manner during other phases of the menstrual cycle. Herzog et al.73 described three distinct patterns of catamenial epilepsy in 184 women with refractory temporal lobe epilepsy (TLE) based on a 1-month seizure and menstrual diary and midluteal progesterone levels (figure 3). The perimenstrual (C1) pattern was defined as a greater average daily seizure frequency during the menstrual phase (day −3 to +3) compared with the midfollicular (day 4 to 9) and midluteal (day −12 to −4) phases in ovulatory cycles. The periovulatory (C2) pattern was characterized by a greater average daily seizure frequency during the ovulatory phase (day 10 to −13) compared with the midfollicular and midluteal phases in ovulatory cycles. In the luteal (C3) pattern, seizure frequency is greater during the ovulatory, midluteal, and menstrual phases than during the midfollicular phase in women with ILP cycles. The average daily frequency of the 1,324 focal and secondary generalized motor seizures recorded during normal cycles in 98 women was significantly greater during the perimenstrual and periovulatory phases than during the midfollicular or midluteal phase. In contrast, the 1,523 seizures recorded during ILP cycles of 86 women occurred with a significantly lower average daily frequency during the midfollicular phase than during any other phase. When defined as a twofold increase in seizure frequency during a particular phase of the cycle, approximately one-third of women had catamenial tendencies. With less stringent criteria, 71% of women with ovulatory cycles had perimenstrual or periovulatory patterns (most having both), and 78% of women with ILP cycles showed the luteal pattern.
Figure 3. Three patterns of catamenial epilepsy (see text for descriptions). Reproduced with permission from Herzog et al. 1997.73
In prospective studies investigating the temporal distribution of seizures, a significant increase in seizures is observed during menses compared with other phases.104,108,111,112⇓⇓⇓ Among 35 women with refractory TLE, mean seizure frequency was significantly greater during menses compared with ovulation and the luteal phase in ovulatory cycles.111 In anovulatory cycles, seizures were significantly less common during menses compared with the remainder of the cycle. Nine women with epilepsy studied over 90 cycles demonstrated a median increase of seizures during menses of 63% compared with the second half of the cycle.104 An increase in menses-related seizures above 50% was found in six cases, and a positive correlation between menses and seizures was found in seven of nine women. In another series, seizures occurred exclusively during menses or ovulation in 46% of ovulatory cycles in women with focal epilepsy.108 Seizure frequency appears to be greater during anovulatory cycles than during ovulatory cycles.64,109,111⇓⇓ Average daily seizure frequency was almost 1.5-fold greater and that of generalized motor seizures almost three-fold greater in women with ILP cycles than in women with normal cycles.73
Diagnosis.
The diagnosis of catamenial epilepsy is established by careful assessment of menstrual and seizure diaries and by characterization of cycle type and duration. Ovulation may be documented in several ways. The simplest approach is to document a history of regular periods, premenstrual symptoms, and dysmenorrhea, and a rise in temperature by at least 0.7° F on basal body temperature (BBT) charts. The patient should take her temperature orally in the morning before any other activity. She starts recording from the first day of the period to the onset of the next. The increase in BBT is due to the presence of progesterone during the postovulatory period. More sophisticated documentation of ovulation includes a serum progesterone level of >3 ng/mL or an endometrial biopsy showing a secretory phase endometrium. Histologic evaluation of the endometrium is performed during the midluteal phase. The criteria used to assess the appearance of the endometrium in response to ovulation were first reported in 1950. An ILP is the result of inadequate progesterone secretion or action. This can be suspected by a BBT rise of <11 days, a midluteal progesterone level of <10 ng/mL, or an out-of-phase endometrial biopsy of >2 days. The postovulatory day of the cycle is dated from the day of the LH surge, as measured by urinary kits. The histologic dating and the clinical dating should be within 2 days of each other. The diagnosis of ILP requires three random serum progesterone levels in the midluteal period (days 5–9 after ovulation documented by the LH surge from a urine test) or two sequential endometrial biopsies showing a >2-day lag between the time of the cycle and histologic dating. In practice, many just use a single midluteal progesterone level of <5–10 ng/mL. Ovulation is more difficult to document in very short (<23 days) and long (>35 days) cycles.
Some women with epilepsy appear to be at increased risk for ovulatory dysfunction, which can make it more challenging to characterize ovulatory status. Some women have both ovulatory and anovulatory cycles, necessitating analysis of multiple cycles.94,111,113⇓⇓ In one series,113 anovulatory cycles were identified in 35% of women with temporal lobe seizures compared with 8% of controls and 0% of women with idiopathic generalized epilepsy. In other series of women with TLE, 32% to 39% of cycles were anovulatory; anovulatory cycles were more common in women with very short or long cycles.111,114⇓ Morrell et al.115 recently reported that anovulatory cycles were more common in women with idiopathic generalized epilepsy (27%) than in those with focal epilepsy (14%) and controls (11%). Furthermore, 38% of women using VPA at the time of testing or within the preceding 3 years had at least one anovulatory cycle, in contrast to 11% of women not using VPA within the same time period. In one of the earliest published case series of catamenial epilepsy, 28% of institutionalized women with epilepsy were amenorrheic for at least 3 months and 10% were amenorrheic for at least 10 months.91 Limbic system dysfunction may underlie these findings because electrical stimulation of the amygdala and hippocampus suppresses LH release and ovulation in female rats.116-118⇓⇓
Management.
A variety of approaches have been proposed for the treatment of catamenial epilepsy; however, all are based on small, unblinded series or anecdotal reports. In 1857, Sir Charles Locock described his experience using potassium bromide in 14 women with perimenstrual seizures.2 Treatment for 7 to 14 days before menses was effective in all but one case. In 1909, Gordon106 reported “very satisfactory results” in 23 women with perimenstrual seizures treated with thyroid hormone for 3 weeks per month beginning with menses, followed by bromide for 1 week and a salt-free diet in which meats, sweets, alcohol, and caffeine were prohibited. In early reports, castration was usually not effective in reducing seizures, although seizure control after oophorectomy and hysterectomy is described.106,119⇓ Within the past several decades, various treatments have been proposed, but there are no published placebo-controlled trials. Women with menstrual dysfunction and in whom hormonal therapy is considered should be evaluated by a reproductive endocrinologist or gynecologist.
Acetazolamide.
Acetazolamide (AZ) is an unsubstituted sulfonamide and a potent inhibitor of carbonic anhydrase. On the basis of the observation that starvation, ketosis, and acidosis reduce seizures, the anticonvulsant properties of AZ were initially attributed to the production of metabolic acidosis due to carbonic anhydrase inhibition. However, studies failed to demonstrate a correlation between bicarbonate levels and seizure frequency.120 AZ produces an accumulation of CO2 in the brain that is sufficient to prevent the tonic extensor component of generalized seizures in the MES model.121,122⇓ Increased CO2 in the extracellular space inhibits the spread of neuronal activity and stabilizes the axon membrane by reducing extracellular calcium.
AZ has been used in the treatment of catamenial epilepsy for almost 50 years largely on the basis of anecdotal reports showing efficacy in isolated cases. Poser123 reported an improvement in seizures without adverse effects in an undisclosed number of women treated with 250 to 500 mg daily for 5 to 7 days before and during menses, an experience that he claimed was shared by “many neurologists over many years.” Ansell and Clarke124 reported a marked benefit in two women with perimenstrual seizures treated with 5 mg/kg/day for 3 days surrounding menses. Although diuretic effect was the proposed mechanism of action, body weight, sodium metabolism, and total body water during menses were no different in women with and without catamenial seizures, and total body water was unchanged once seizures were controlled with the drug.83 These results were not confirmed in larger series.125,126⇓ The efficacy of AZ is limited by the development of tolerance.123,124⇓ In a recent retrospective review of 20 women with catamenial epilepsy treated at our institution, seizure frequency was significantly reduced in 40% and seizure severity was reduced in 30% of AZ-treated women.127 The mean daily dose was low (347 mg), which may have led to an underestimation of efficacy and may explain the low incidence of tolerance and adverse effects.
The initial recommended dose of AZ is 4 mg/kg, with a range of 8–30 mg/kg/day in one to four divided doses, not to exceed 1 g per day. Adverse effects include paresthesias, drowsiness, ataxia, nausea, vomiting, malaise, anorexia, fatigue, diuresis, intermittent dyspnea, depression, hyperchloremic metabolic acidosis, dysgeusia, renal calculi, growth suppression in children, and aplastic anemia. Tolerance to its anticonvulsant effect is due to the induction of increased amounts and activity of carbonic anhydrase in glial cells and to the production of additional glial cells.128 Dose escalation may be required to maintain the anticonvulsant effect. Alternate-day or cyclic dosing for several days surrounding the phase of seizure exacerbation may reduce the development of tolerance.
Cyclical AEDs.
Because of the potential for development of tolerance and adverse effects, BDZs are usually reserved for the treatment of acute seizures and status epilepticus. Although intermittent BDZs have been used to treat women with catamenial seizures for years, only clobazam has been studied in this population. Clobazam is the first 1,5-BDZ to be marketed (although not available in the United States) and is purported to have fewer adverse effects than the older BDZs having a 1,4 configuration. Feely et al.129 compared clobazam and placebo in a double-blind crossover study in 24 women with refractory perimenstrual seizures. Clobazam 20 mg/day was administered for 10 days beginning 2 to 4 days before menses during one cycle. Efficacy was defined as seizure freedom in patients with fewer than four seizures per month or >50% reduction in women with more frequent seizures. Clobazam was effective in 14 of 18 evaluable cases (78%), including three women who initially also responded to placebo and three who responded only during a second trial of the drug at a higher dosage (30 mg/day). Six subjects failed to complete the protocol, including two who withdrew because of adverse effects. The most common adverse effects were sedation and depression. Sustained efficacy was realized in 13 women over 6 to 13 months.130 Ten women were free of perimenstrual seizures and two additional women had single seizures during a treatment period ranging from 2 months to 3.5 years. Seizures occurred between menstrual periods in three cases and increased in one woman with generalized epilepsy. Tolerance was not observed in nine women treated for ≥1 year.
The use of conventional AED therapy adjusted during periods of seizure exacerbation has not been adequately investigated. In a single report of a woman treated with VPA in whom serum concentrations varied by 35%, with the lowest level occurring during the week of menses, seizures were reduced from eight to one per month when the dose was adjusted to correct for the variability in serum concentrations.87 Although this approach may be attractive, frequent changes in medical regimens increase the chance of patient error, particularly in the medically refractory population.
Hormonal therapy.
Oral contraceptives.
Oral contraceptives usually consist of an estrogen and a progestin. Most oral contraceptives contain ethinyl estradiol. Few contain mestranol, which is metabolized to ethinyl estradiol to be active. There are a number of progestins, most of which are derivatives of testosterone, although a recently developed progestin, drospirenone, is derived from spironolactone. The androgenic effects of progestins are uncommon at the low doses used in modern oral contraceptive agents. A low-dose contraceptive is one with <50 μg of estrogen. The cardiovascular effects of oral contraceptives are related to the estrogen dose, and 50-μg formulations are rarely used for oral contraception and uncommonly used, in general, for other medical reasons. Most agents are taken orally, although there is a monthly intramuscular formulation as well as a weekly dermal patch. The oral formulations that contain estrogens and progestins are classified as monophasic, biphasic, and triphasic. These refer to the alterations in the dose of estrogens or progestins throughout the cycle, ostensibly to obtain better control of menstrual bleeding. The progestin-only pills are used when estrogens are contraindicated. The effect of the latter formulation in preventing ovulation is significantly reduced compared with the conventional pill. This results in frequent menstrual abnormalities. The efficacy of oral contraceptives is reduced in women who are taking enzyme-inducing AEDs. There is no evidence that any oral contraceptive increases seizure activity.
Isolated cases of improved seizure control in women treated with oral contraceptives have been described.131-134⇓⇓⇓ Given the widespread use of these agents, the paucity of literature in this area is surprising. Among four women with catamenial epilepsy, seizures were reduced in only one case.135 In the only double-blind, placebo-controlled study, the oral synthetic progestin norethisterone was administered to nine women with focal seizures related to menses.136 Both low- and high-dose regimens were ineffective.
Medroxyprogesterone acetate (MPA).
Medroxyprogesterone acetate is a progesterone derivative available in oral and parenteral formulations. Depot-MPA (Depo-Provera) is the most extensively studied progestin-only contraceptive and is typically administered intramuscularly at 3-month intervals at a dosage of 150 mg/injection.137 In women with adequate levels of endogenous estrogen, MPA produces thickening of the cervical mucus and alteration of the endometrium from a proliferative to a secretory state. Depot-MPA produces levels of progestin high enough to effectively block the LH surge, thereby blocking ovulation. Because FSH is not significantly suppressed, estrogen levels remain comparable to those in the early follicular phase of a normal cycle, and symptoms of estrogen deficiency, such as vaginal atrophy and diminished breast size, do not occur. Depot-MPA is an appropriate contraceptive choice for women who are noncompliant or cognitively impaired and for those at risk of estrogenic side effects, including patients >30 years of age, smokers, and women with a history of thromboembolic events or vascular disease. Adverse effects include irregular menstrual bleeding, breast tenderness, weight gain, and depression. Bleeding and spotting decrease in the long term, and 80% of women are amenorrheic after 5 years of treatment.138
MPA reduces seizures in small numbers of women with epilepsy.139,140⇓ Mattson et al.140 treated 14 women with poorly controlled focal (n = 13) or generalized absence (n = 1) epilepsy with oral MPA 10 mg administered two to four times daily. Six women who failed to become amenorrheic were treated with depot-MPA 120–150 mg at 6- to 12-week intervals. A 39% reduction in overall seizure frequency was achieved at a mean follow-up of 12 months. Among seven responders, the mean seizure reduction was 52% (25% to 71%). Seizures were not improved in the only woman with generalized epilepsy. No serious adverse effects were reported. A 3- to 12-month delay in resumption of regular menses was observed after treatment with depot-MPA. Serum AED concentrations were not affected. However, women receiving enzyme-inducing AEDs may require higher doses or shorter dosing intervals because of the enhanced steroid metabolism. MPA has a dose-dependent inducing effect on hepatic microsomal drug metabolism in rats, with female rats more sensitive to this effect than males.141
Natural progesterone.
In contrast to oral synthetic progestins, which have been shown to be ineffective, natural progesterone is considered the treatment of choice for women with ILP cycles. In 1986, Herzog142 was the first to describe the use of natural progesterone in the treatment of epilepsy. Table 2 summarizes his findings in the three studies. Eight women with TLE and anovulatory cycles or luteal phase defects were treated with vaginal progesterone suppositories in doses ranging from 50 to 400 mg every 12 hours during the phase of highest seizure frequency. Vaginal administration was chosen for its ability to achieve adequate midluteal serum progesterone levels while avoiding the reduced bioavailability of oral formulations. Serum progesterone was measured 2 to 6 hours after suppository insertion and the dosage adjusted to achieve midluteal levels from 5 to 25 ng/mL. AED levels were measured during and after progesterone use and dosages were adjusted to maintain therapeutic levels when necessary. Compared with baseline, average monthly seizure frequency declined by 68% during the 3-month treatment period, and 75% of women had fewer seizures. Four women experienced transient fatigue and depression when the dosage was increased beyond minimally effective levels. One nonresponder experienced adverse effects at low therapeutic levels. In all cases, symptoms resolved within 48 hours of dose reduction.
Adjunctive progesterone therapy in women with catamenial epilepsy
In a subsequent study, Herzog143 reported his experience using progesterone lozenges 200 mg three times daily in 25 women with focal seizures of temporal lobe origin exacerbated during the perimenstrual (n = 11) or luteal (n = 14) phases. Subjects were treated surrounding the period of exacerbation and the dose was adjusted to maintain normal midluteal progesterone levels. AED levels remained in the therapeutic range and dosages remained constant. Over a 3-month treatment period, 72% of women experienced a reduction in seizures and average daily seizure frequency declined by 55%. Focal and generalized motor seizures were reduced to a similar degree. Seizure frequency was unchanged in five cases. Reduction in average daily seizure frequency was greater in women with ILP cycles (59%) compared with perimenstrual seizures (49%). Two patients discontinued treatment because of asthenia and depression. In 23 women who continued treatment, the mean reductions of focal and generalized seizures after 3 years were 54% and 58%, respectively, and three women were seizure-free.144
Others have also produced favorable results using intermittent progesterone in women with catamenial seizures.89,145,146⇓⇓ In the largest report, 36 women were treated with sublingual progesterone over a period averaging 18 months.146 Four subjects became seizure-free. Focal seizures were reduced by 68% and generalized seizures by 57%. Seizures increased or were unchanged in 25% of cases. In a single case report, seizure control deteriorated during treatment with progesterone in a woman with generalized absence epilepsy.147 Whether or not the effect of sex steroids on seizure susceptibility differs in women with generalized and focal epilepsy is unknown.
Antiestrogens.
Clomiphene citrate is an estrogen analogue used in the treatment of ovulatory dysfunction in women who desire pregnancy. Clomiphene stimulates the hypothalamic–pituitary axis, increasing gonadotropin pulse amplitude and producing a rise in FSH and LH that induces ovulation.137 The drug is typically administered starting on the fifth day of the cycle at a dose of 50 mg per day for 5 days. If ovulation is not achieved in the first cycle of treatment, the dosage is increased in 50-mg increments in subsequent months. The most common adverse effects include vasomotor flushes, abdominal distension or pain, breast discomfort, nausea, vomiting, and visual disturbances. Ovarian enlargement has been associated with long-term treatment. Herzog148 treated 12 women with temporal lobe seizures and menstrual disorders, including nine women with polycystic ovarian syndrome and three with luteal phase defects. Clomiphene 25 to 100 mg daily was administered on days 5 to 9 of each cycle. The dosage was increased monthly until regular cycles were obtained (cycle duration of 26 to 32 days and ovulatory midluteal progesterone levels achieved) or until pelvic pain and cramps developed, signifying ovarian overstimulation. Seizures were reduced by at least 50% in 10 cases, including five women who experienced >90% seizure reduction. Adverse effects occurred in six cases, all of which were transient and resolved with drug withdrawal. In a young woman with Lennox–Gastaut syndrome and amenorrhea treated with VPA, clomiphene induced menses and reduced seizures.149 The drug also produced a significant reduction in seizures in a male patient with focal epilepsy and oligospermia.150
Androgens.
A case of cerebral endometriosis presenting with focal seizures related to menses was treated successfully with danazol, a synthetic androgenic steroid derivative that induces atrophic changes in endometrial tissue.151 Danazol eliminates the gonadotropin surge of FSH and LH and inhibits steroidogenesis in the corpus luteum. A high-androgen, low-estrogen state is produced, reducing the pain and disease progression of endometriosis. Danazol has significant adverse effects related to its hypoestrogenic and androgenic properties, including weight gain, fluid retention, fatigue, diminished breast size, acne, facial hair growth, emotional lability, and atrophic vaginitis.
Gonadotropin analogues.
Suppression of pituitary secretion of gonadotropins by GnRH agonists is an effective treatment for endometriosis, uterine leiomyomas, precocious puberty, tumors of the breast, pancreas, ovary and prostate containing GnRH receptors, and other sources of menstrual bleeding. Gonadotropin analogues suppress ovarian and testicular steroidogenesis. These agents are available in intramuscular, subcutaneous, and intranasal formulations. Goserelin, nafarelin, cetrorelix, and leuprolide are available in the United States. Triptorelin, a synthetic GnRH agonist, is the only agent in this class studied in women with catamenial seizures. Unlike oral contraceptive agents, triptorelin is metabolized in the brain and does not interfere with hepatic metabolism of AEDs. Bauer et al.152 treated 10 women with refractory perimenstrual seizures and amenorrhea with triptorelin 3.75 mg intramuscularly at 4-week intervals. Three women became seizure-free during a treatment period averaging 12 months. Four patients experienced a reduction in seizures; seizure duration was reduced in one case. Eight women experienced adverse effects including hot flashes, headache, and weight gain. Because long-term use produces bone demineralization by inducing a pseudomenopausal hypoestrogenic state, these agents are probably of limited utility. However, bone loss was reversible in women with endometriosis treated with GnRH agonists for periods as long as 6 months.153 Treatment with the LHRH agonist goserelin markedly reduced the frequency of status epilepticus in a woman with primary generalized epilepsy.154
Neuroactive steroids.
The term “neurosteroid” was originally coined to designate a steroid intermediate, dehydroepiandrosterone sulfate, which was found in the brain at concentrations independent of peripheral stores.12 This term is now used to refer to a variety of steroid hormones that are synthesized de novo in the brain, including progesterone and estrogen. Neuroactive steroids are steroid hormones synthesized in the brain or peripheral organs that are active on neuronal tissue.12 Ganaxolone, 3α-hydroxy-3β-methyl-5α-pregnan-20-one, is a neuroactive steroid that modulates the GABAA receptor complex via a recognition site distinct from those of BDZs and barbiturates. Ganaxolone is a synthetic analogue of allopregnanolone that has been shown to possess anticonvulsant properties in a variety of experimental models, including PTZ- and MES-induced seizures.155
In a series of 20 children with infantile spasms, ganaxolone reduced spasms by 50% in 33% of subjects; an additional 33% experienced some improvement.156 In an inpatient monotherapy trial, 52 patients with medically refractory focal epilepsy who were withdrawn from AEDs during presurgical evaluations were randomized to receive ganaxolone or placebo for up to 8 days.157 The primary end point was duration of treatment before withdrawal, which occurred after four seizures of any type, three generalized tonic–clonic seizures, or status epilepticus. By the third day of treatment, a clear separation on Kaplan–Meier survival curves was observed, with 50% of ganaxolone-treated patients completing the study versus 25% of placebo-treated subjects. Treatment was well tolerated. In the only report in catamenial epilepsy, a moderate improvement in seizures was achieved in two women with perimenstrual seizures treated with ganaxolone 300 mg twice daily from day 21 of the cycle through day 3 of menses.158
Conclusions.
Catamenial epilepsy is a common disorder, affecting one-third of women with epilepsy. Three distinct patterns of seizure susceptibility have been described: perimenstrual, periovulatory, and luteal phase exacerbations. The pathophysiology of this disorder has not been entirely elucidated, although recent studies suggest that the abrupt withdrawal of neurosteroids is an important factor in women with perimenstrual seizures. Although several different treatment approaches have been proposed, none of these approaches have been compared, and evidence of their efficacy is based on small uncontrolled series or anecdotal observations. Further investigation into the pathophysiology of this disorder and larger multicenter trials are required to identify the most effective treatment for this important subset of women with epilepsy.
Footnotes
-
Publication of this supplement was supported by an unrestricted educational grant from GlaxoSmithKline. The sponsor has provided N.F.-S. with honoraria for her participation in this project and has provided her with other honoraria and grant support during her career.
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.