Herzog et al. [29] studied 50 consecutive women with temporal lobe epilepsy. Fifty-six percent had amenorrhea, oligomenorrhea, or abnormally long or short menstrual cycle intervals, and 68% of these women had clearly identifiable reproductive endocrine disorders with polycystic ovarian syndrome (PCOS) in 20%, hypoandrogenism in 12%, severe premature menopause in 4%, and functional hyperprolactinemia in 2%. The estimated frequency for hypoandrogenism in the general population of women is 1.5%. since there was no significant relationship between these menstrual disorder and the use of AEDs, the authors concluded that epilepsy itself may be a significant factor, which stimulated future investigations.

Hypogonadotopic hypogonadism has been found to be more common in women with right temporal lobe epilepsy compared to left temporal lobe epilepsy [36,46]. Herzog et al. [36] studied 30 women with unilateral temporal lobe interictal epileptiform dischages and reproductive disorders. EEG laterality was significantly different for women with PCOS (left sided discharges) and hypogonadotropic hypogonadism (right-sided discharges). Kalinin et al. [46] studied 80 women with temporal lobe epilepsy. They also reported that PCOS was more commonly associated with left TLE, and hypogonadotropic hypogonadims with right TLE. They also reported that a catamenial patter was more common in the women with left TLE.

Other studies supporting the effect of epilepsy itself on reproductive dysfunction include the finding that women with idiopathic generalized epilepsy have higher pulse frequency GnRH secretion than normal controls [8]. Quigg et al. [69] studied LH secretion patterns in 10 men with TLE and in control men. They reported that the pulsatile secretion of LH was abnormal in the TLE group; interictal effects consisted mainly in loss of circadian fluctuations in LH burst amplitude, whereas postictal effects consist of altered burst timing.

A study of 40 men with refractory localization-related epilepsy by Kuba et al. [50] in the Czech Republic investigated sexual dysfunction and hormonal profiles. At least one of the types of sexual dysfunction was found in 55% of patients, with erectile dysfunction in 15%, orgasmic dysfunction in 15%, and loss of sexual desire in 40%. In the group of men with sexual dysfunction, there was an increase of FSH and SHBG, and a decrease of DHEAS and FAI in comparison with patients with normal sexual function. An increase in estrogen was also seen in men with erectile dysfunction. The observation was made that all patients with orgasmic dysfunction were being treated with CBZ, but when all patients with at least one type of sexual dysfunction were analyzed, a higher proportion were on VPA than CBZ.

Testosterone supplementation can be an effective treatment for hypogonadal men with epilepsy and sexual dysfunction. It may improve not only sexual function, but mood and even seizure frequency [33].

Morrell and Flynn [56] studied sexual function and hormones in women with epilepsy 18 to 40 years with localization-related epilepsy, primary generalized epilepsy (PGE), and nonepilepsy controls. Questionnaires were administered examining sexual experience, arousability, anxiety, and depression (Beck Depression Inventory). Endocrine assessment was performed during the early follicular phase. Compared to the controls, women with LRE had significant higher sexual dysfunction scores, lower mean arousal, and higher depression scores. Mean arousal scores were also lower in the PGE group. Women on EIAEDs, combined into one group, had statistically higher sexual dysfunction and lower sexual arousal compared to controls Overall, estradiol negatively correlated with sexual anxiety, and dehydroepiandrostone (DHEAS) negatively correlated with sexual dysfunction and positively correlated with sexual arousal.

Another study reported similar adverse findings of altered hormone blood levels in association with lower sexual quality of life in men with epilepsy, but interestingly, not women with epilepsy [55]. Seventy-nine patients with different types of epilepsy or no epilepsy completed the Derogatis Interview for Sexual Function-Self-Report Inventory and had blood drawn for measurement of hormone studies. In men, increasing sex hormone-binding globulin (SHBG) levels and duration of epilepsy decreased sexual quality of life. SHBG levels were related to enzyme-inducing antiepileptic drugs (EIAEDs) and age. Women did not demonstrate an association between any hormone levels and sexual quality of life.

Another study reported that the reduced rate of childbirths held in married women with epilepsy (69% of expected number of live-born children) but did not hold in married men with epilepsy [13]. A population- based cohort study in Finland (1985–2001) of patients with newly-diagnosed epilepsy (n=14,077) reported that childbirth rates were lower than in a reference cohort among men (hazard ratio = 0.58 (95% CI 0.54-0.62)) and women (HR=0.88 (0.83–0.93)) [2]. However, a comparable study from Iceland suggested no difference in childbearing rates for women with epilepsy [60]. Geographic and cultural differences are likely substantial factors and these will influence self-directed decisions. The “Ideal World” survey of Epilepsy Action UK women found that 33% of the women were not considering having children because of their epilepsy [12].

The definition of PCOS has been a source of debate and varies between Europe and North America, with a more strict criteria adopted by North America. Minimal diagnostic criteria were defined by a 1990 NIH Consensus Panel: 1) menstrual irregularity, 2) biochemical or clinical evidence of hyperandrogenism, 3) and exclusion of other diseases that could cause female hyperandrogenism (such as congenital adrenal hyperplasia) [54]. These North American minimal criteria do not include the need to identify the polycystic component of the ovaries by ultrasound. The signs and symptoms expressed by women with PCOS vary substantially (Table 1) [4]. There is a familial tendency with some recent discoveries of potential underlying genetic contributions with variants of luteneizing hormone (LH), the pathway for androgen biosynthesis and of regulation of expression of the insulin gene [4]. Women with PCOS are at long term risk of developing diabetes and cardiovascular disease.

One of the most influential factors that affect the expression of PCOS is weight; weight gain causes a worsening of symptoms, and weight loss improves the endocrine and metabolic profile. Yet women with PCOS are prone to eating disorder, possibly due to a link with leptin. This is one of the reasons why the interplay between epilepsy, the use of VPA, weight gain, and PCOS has been so difficult to tease out. The underlying epilepsy itself likely contributes to the expression of PCOS via effects on hypothalamic control of the menstrual cycle with the tendency toward more anovulatory cycles and oligomenorrhea. It is likely that both the underlying epilepsy disorder and AED treatment play roles in the development of PCOS [54].

Herzog, et. al. [41] was one of the first groups to publish on the finding of PCOS in women with epilepsy. He reported that PCOS occurs significantly more often in women with TLE than in the general female population and is associated with predominantly left-sided lateralization of interictal epileptic discharges. These findings suggest that the underlying epilepsy contributes to the development of PCOS via the hypothalamic-pituitary-gonadal axis. Bilo, et al. also reported a higher incidence of reproductive endocrine disorders including PCOS, but in women with idiopathic generalized epilepsy syndromes [7]. Neither group found an association with the type of AED being taken by the women.

Several studies by Isojarvi et al. reported that the higher incidence of PCOS in WWE was specifically associated with VPA use. One of the probable reasons for discrepancies in different studies is due to the differing definitions for PCOS. A study of 238 WWE in Finland on a variety of AEDs included vaginal ultrasonography and serum sex-hormone concentrations [42]. Unlike the North American criteria for PCOS, they reported on findings of polycystic ovaries as an isolated finding as well as elevated testosterone concentrations. The VPA monotherapy group included findings of polycystic ovaries (43%) and elevated testosterone concentrations without polycystic ovaries (17%). Notably, 80% of the group with that began VPA treatment before 20 years old had polycystic ovaries or hyperandrogenism

A later study by Isojarvi, et al. [44] enrolled 16 women taking VPA for epilepsy that had polycystic ovaries or hyperandrogenism, They were converted to lamotrigine (LTG) monotherapy and followed for one year (n=12 completers). The additional findings of weight gain, hyperinsulinemia, and lipid profiles were monitored. During the first year after drug conversion, the number of polycystic ovaries, BMI, fasting serum insulin and testosterone concentrations decreased, and the HDL/total cholesterol ratios increased.

A recent examination of 148 women with epilepsy (WWE) by epilepsy type and AED use by Lofgren, et.al. [51] reported a higher prevalence of reproductive endocrine disorders in women with idiopathic generalized epilepsy than control subjects, but the specific findings of hyperandrogenism, polycystic ovaries, and polycystic ovary syndrome were more prevalent in WWE on VPA than in WWE taking other drugs or control women. The use of VPA and younger age predicted the development of hyperandrogenism.

Morrell, et al. [57] studied 447 women prospectively with randomization to initiating 12 months of treatment with either VPA or LTG. More women in the VPA group than the LTG group developed ovulatory dysfunction (54% vs. 38%; p = 0.010) and more women in the VPA group than the LTG group developed PCOS (9 vs 2%; p = 0.007). Development of hyperandrogenism was more frequent with VPA than LTG among those initiating treatment at age younger than 26 years (44% vs. 23%; p = 0.002) but was similar if treatment was started at age 26 years or older (24% vs. 22%).

Another study in children with new-onset epilepsy [82] was performed prospectively with baseline thyroid testing prior to drug administration, followed by testing at 3, 6, and 12 months after initiation of CBZ or VPA; age-matched subjects served as controls. The epilepsy group included subjects with partial epilepsy and subjects with generalized epilepsy. Thyroid function tests were normal at baseline in all subjects. After 3 months and persisting until 12 months, CBZ-treated patients demonstrated significantly lower serum T4 and fT4 levels compared to baseline evaluation and control subjects, but normal serum T3 and TSH response to a TRH test. Some of the CBZ-treated children were withdrawn from CBZ for other reasons, and reevaluation 6 months later revealed renormalization of their values. All TFT values remained normal in the VPA group. Even with the alterations in the CBZ group, the authors concluded that the patients are euthyroid and therefore, thyroid hormone alterations are not associated with clinical or subclinical hypothyroidism.

Gomez, et al. [23] studied 96 patients on a variety of first-generation AEDs. They observed decreased T4 serum levels and free T4 index below the normal ranges (in 25 and 14 patients, respectively), decreased rT3, but normal levels of free T4, T3, and TSH. They concluded that the TSH was the best measurement to reflect the euthyroidism of these patients.

Studies of other AEDs, especially the enzyme-inducing AEDs, have demonstrated somewhat similar findings of a tendency toward lower serum concentrations of thyroxine (T4), free thyroxine, and thyroid-binding globulin while maintaining TSH at normal or slightly elevated levels [33]. Thyrotropin response to thyrotropin-releasing hormone is either unchanged or slightly increased as well [43,81].

Another study of 35 patients on long-term PHT or CBZ [81] reported that the mean concentrations of T4, FT4, FT3, and rT3, but not T3, of these patients were significantly lower than those of 19 controls of similar age and sex distribution. The mean serum TSH concentration was slightly but significantly higher in patients than in controls, but the serum TSH response to TRH was not significantly increased. Several other metabolic studies in these patients did not show other clinical signs of functional hypothyroidism and they assessed response to thyroxine treatment in a smaller subset of subjects. These authors concluded that, on the basis of all data from the cross-sectional and thyroxine treatment studies, patients receiving AEDs chronically are eumetabolic and do not need thyroxine supplementation.

Herzog et al. [40] studied a cohort of 184 women with medically-refractory localization-related epilepsy and were able to define three distinct patterns of catamenial epilepsy based on statistical evaluation of seizure frequency in women during their reproductive years. The definitions are based on Day 1 as the first day of menstrual flow and ovulation is presumed to occur 14 days before the subsequent onset of menses (Day −14). Seizures and menses were charted and midluteal serum progesterone levels were obtained on day 22. Progesterone concentrations >5 ng/ml were considered ovulatory. Cycles were divided into four phases: menstrual (M) = −3 to +3, follicular (F) = 4 to 9, ovulatory (O) = 10 to −13, and luteal (L) = −12 to −4 (Figure 3). Average daily seizure frequency for each phase was calculated and compared among phases, separately for ovulatory and anovulatory cycles. The seizures recorded during ovulatory cycles occurred with significantly greater average daily frequency during the menstrual and ovulatory phases than during the follicular and luteal phases, in patterns labeled perimenstrual (catamenial 1) and periovulatory (catamenial 2). For the anovulatory cycles, seizures occurred with lower frequency during the follicular phase than during all other phases, in a pattern labeled catamenial 3 by the authors (Figure 3). The authors went on to plot of the percentage of women with greater seizure frequency according the three patterns of catamenial epilepsy and evaluated points of inflection of the S-shaped distribution curves for optimal distinction between women with high and low susceptibility to hormonal influence. Based on these findings, these authors proposed a twofold increase in average daily seizure frequency during these phases in one of the three patters for the designation of catamenial epilepsy. In their patient cohort, one-third of women showed at least one of the three patterns of catamenial epilepsy.

A follow-up study by Herzog et al. [39] in 87 women found that 39% of women enrolling in a progesterone treatment trial demonstrated a catamenial pattern. This study assessed 87 women over a few menstrual cycles rather than just one cycle. Catamenial epilepsy designation was made if two of three cycles showed at least one of the three catamenial patterns. Another study by El-Khayat et al [16] of 42 Egyptian women with epilepsy (and 21 controls) evaluated the pattern of seizure occurrence and hormone levels. Additionally, pelvic ultrasound was performed near the time of ovulation. A catamenial pattern of seizures occurred in 31% of patients (53.8% C1 pattern; 46.15% C3 pattern). Patients with C3 pattern showed lower progesterone levels in the midluteal phase compared to patients with noncatamenial pattern, to those with C1 pattern or to controls. Patients with the C1 pattern had lower progesterone levels than controls in the menstrual phase. The patients with a catamenial pattern also had higher estradiol/progesterone ratios.

A later open-label trial by Herzog et al. [34] of 25 women with medically-refractory temporal lobe epilepsy included women with either anovulatory or ovulatory cycles and with the associated catamenial patterns. Progesterone was administered as oral lozenges, 200 mg tid, in relation to the pattern of seizure exacerbation. For perimenstrual exacerbation (C1 pattern) they were given for days 23 to 25 of each menstrual cycle and then tapered to discontinuation by day 28. For luteal phase exacerbation (C3 pattern) progesterone lozenges were given for days 15 to 25 of each menstrual cycle followed by the same taper schedule. The dosage of progesterone dosage was adjusted if needed to produce physiologic luteal range progesterone serum levels between 5 and 25 ng/ml 4 hours after taking a lozenge. Over 3 months of treatment, overall average monthly decline was 54% for complex partial and 58% for secondarily generalized tonic clonic seizures. Overall, 72% of the patients experienced a reduction of seizures. The investigators noted that progesterone was more efficacious when administered during the entire second half of the cycle for both groups with the different catamenial patterns, followed by gradual taper over 3 to 4 days. This observation formed the basis for future treatment trials.

Fifteen women from the treatment group in this open-label study were followed for over 3 years and continued to show improved seizure control [35]. Three of the women maintained seizure freedom and four had seizure reduction of >75%, while another eight had 50–75% seizure reduction. It is encouraging that despite a likely primary action at the GABA_A-BDZ receptor complex (via conversion to allopregnanolone), tolerance to adjunctive progesterone does not appear to occur and the efficacy is maintained.

The findings from these open label trials, including details about which regimen seems to most effective for all patterns of catamenial epilepsy, have led to a prospective, double-blind, randomized, placebo-controlled, multicenter trial of use of cyclic adjunctive progesterone lozenges. Enrollment just ended and analysis of the data is pending. The population enrolled included women with medically-refractory focal-onset seizures, and included a group with a non-catamenial pattern and another group with any of the three catamenial patterns [39].

Progesterone lozenges can be obtained from almost all compounding pharmacies. It can be initiated as it is used in the treatment trials as 200mg TID on days 15 to 25, with a taper to 100 mg TID for 2 days, then 50 mg TID for one day, then off. However, it is often helpful to later individualize the dose based on response and progesterone levels with a target concentration of 20–40 ng/mL drawn 4 hours after dosing. Some women may require an even slower taper or addition of topical progesterone cream during other phases of the menstrual cycle to prevent withdrawal effects. Progesterone is also available in a micronized form in an oral capsule preparation that may be effective, but it has not been studied formally. One concern by experts is the possible large first-pass effect of hepatic metabolism which occurs following swallowing of the capsule. In addition to possible sedation, and changes in mood, progesterone can also be associated with breast tenderness, weight gain, irregular vaginal bleeding and sometimes constipation. Higher progesterone dosages may be required to achieve luteal-range levels in women who take EIAEDs.

The use of adjusting conventional AED therapy during periods of seizure exacerbation has not been studied in a systematic manner or reported on, but is occasionally used by practitioners for women with catamenial seizure patterns. Intermittent benzodiazepines have been used for treatment of catamenial seizures, despite very few studies. One group reported on use of clobazam in a small group of patients a few decades ago, with moderately successful results [17,18]. Clobazam, 20 or 30 mg/day was given for 10 days around menstruation in successive menstrual cycles to 13 women who had responded favourably to this drug in the earlier short-term placebo controlled cross-over study. A little over 1/3 did very well seizure freedom around menstruation with retained benefit for over 3 years, another 1/3 had a moderate response, and another 1/3 had seizures or side effects leading to discontinuation of the clobazam. The most common adverse effects were sedation and depression.

Clobazam is is not available in the United States but can be obtained from pharmacies outside the US. Despite the lack of rigorous trials, it is a fairly popular treatment regimen for women with catamenial epilepsy in Europe, and is purported to be better tolerated than other benzodiazepines.

Another option is use of contraceptive hormones to eliminate the effect of cycling endogenous hormone concentrations. Again, despite the fact that his may be used in clinical practice as a treatment strategy, clinical trials are lacking. There have been isolated case reports, but a paucity of systematic trials or even observational studies reported in the literature [19]. Taking this theoretical benefit one step further, there may be an additional advantage of use of continuous dosing OCPs, with rare to no weeks of placebo administration.

Medroxyprogesterone acetate (MPA) is a synthetic progestin and is available as an oral form and as an intramuscular injection every 3 months. This suppresses normal ovulatory hormone cycling, and thus could theoretically improve catamenial epilepsy. Mattson et al. [53] studied 14 women with medically-refractory epilepsy for the effects of MPA. They were given the oral or depot-intramuscular form to achieve amenorrhea. Half the women were responders with seizure reduction and the group overall had a 39% reduction in seizure frequency. As is common for all women receiving depot-MPA, they had a 3- to 12-month delay in resumption of regular menses after treatment. In addition to this problem, in the general population, reported side effects are common with irregular menstrual bleeding, breast tenderness, weight gain, and depression. Another concern for women with epilepsy, especially if they are on an EIAED, is that depot-MPA is associated with increased risk for osteoporosis.

Supported in part by NIH NINDS RO3NS063233