Epigraph Vol. 24 Issue 3, Summer 2022

Sex hormones and epilepsy: Dr. Andrew Herzog

Reported by Bruna Nucera | Produced by Nancy Volkers

Cite this article: Nucera B. Sex hormones and epilepsy: Dr. Andrew Herzog. Epigraph. 2022;24(3):1-7.

Listen below or download the episode.

Find Sharp Waves episodes on SpotifyApple PodcastsGoogle Podcasts, or Amazon Music.

About two-thirds of women with epilepsy have seizure fluctuations associated with their menstrual cycles, with about one-third experiencing at least twice as many seizures during certain phases. Dr. Bruna Nucera talks with Dr. Andrew Herzog about the three patterns of catamenial epilepsy, possible mechanisms and treatments, and how sex hormones otherwise affect both women and men with epilepsy.


Podcast Transcript

Dr. Nucera: Good morning everybody, I’m Bruna Nucera from Merano, Italy. I’m a young epileptologist and today I’m with Professor Herzog, Andrew Herzog, and we will talk about the link between reproductive hormones and epilepsy.

Dr. Herzog: I’m Andrew Herzog, professor of neurology at Harvard Medical School and director of the Harvard Neuroendocrine Unit at Beth Israel Lahey Health.

Dr. Nucera: Thank you so much for being here. You helped to define and classify catamenial epilepsy, in which seizure patterns are related to women’s menstrual cycle. Can you talk a little bit about the research that led up to this?

Dr. Herzog: There has been a line of research that has led to our current concept of catamenial epilepsy, starting with the idea that seizures don’t occur randomly. In the majority of men and women with epilepsy, seizures tend to cluster. Seizure clusters show a temporal periodicity – that can be circadian, ultradian, multidien, and lunar, or annual. When the lunar periodicity in women tends to align itself with the menstrual cycle, it’s called catamenial epilepsy.

A number of investigations since 1883 have shown that seizures tend to predominate in the perimenstrual phase of the menstrual cycle. And in 1956, John Laidlaw did a study and an analysis of 900 menstrual cycles, and found seizures tend to have a predilection not for occurrence only in the late luteal phase, but also in the late follicular phase, which is prior to ovluation. So not just prior to menstruation, but also prior to ovulation. Then in 1976, Bäckstrӧm found that in women who have ovulatory cycles, the seizure frequency tends to vary with the serum estradiol-to-progesterone ratio, and in women with anovluatory cycles who do not make progesterone or make very little, their seizure frequency tends to vary with the estradiol.

Based on this knowledge and our own observations, we proposed in 1997 that there are three patterns of catamenial epilepsy. The C1 pattern occurs perimenstrually, and we arbitrarily assigned it to occur between day -3 and day +3 of the cycle, with day 1 being the first day of menstrual flow. C2 pattern, or preovulatory pattern, which occurs from day 10 of one cycle to day -13, with day -14 being the usual day of ovulation in 95% of women in the general population. In other words, the luteal phase is fixed at 14 days, so ovulation always occurs 14 days before menstrual onset, and it’s the follicular phase that varies in length to account for variations in menstrual cycles.

Then we identified a third pattern that occurs in anovulatory cycles, where progesterone is not made after ovulation. It starts with a preovulatory estradiol surge on day 10, and lasts until day 3 of the next cycle. So we proposed the existence of these three patterns of catamenial epilepsy, then we carried out a prospective study in 184 women, which verified statistically the existence of these three patterns.

Dr. Nucera: In terms of epidemiology, this is common? 

Dr. Herzog: Taubøll, Rościszewska, Ansell and Clarke, John Laidlaw, found that two-thirds of women have more seizures perimenstrually than during the rest of the cycle. We also found this. We found 71.4% had more seizures, according to one of these patterns. But if you want to know how many women have not just more seizures but two times more, that goes down to 35%. If you want to know how many women had three times more seizures during one of these phases, it goes down to 11.2%.

Our statistician made a plot of the proportion of women who had more seizures against the level of seizure exacerbation, expressed as multiples of seizure frequency in the mid-follicular and luteal phases. In other words, we found 71.4% of women had more seizures perimenstrually, 35% had two times more, 11.2% had three times more. And when he plotted this curve, he found a reverse S-shaped curve. And the reverse S-shaped curve starting at 71.4% and dropping down to 11.2% suggests there were many women with epilepsy that had some hormonal sensitivity to their seizures, and a minority of women who had marked sensitivity.

When you have reverse S-shaped curves, the best way to mathematically distinguish between these two populations – some sensitivity and marked sensitivity – is to take the point of inflection of the curve. The point of inflection for each of these three patterns fell between 1.5 and 2. For perimenstrual it was 1.69. We used this point of inflection as the cutoff for identifying a catamenial exacerbation in a cycle, and for our NIH progesterone trial we required that women show a catamenial pattern in 2 out of 3 consecutive cycles. That’s the long answer!

Dr. Nucera: It’s a long story.

Dr. Herzog: This is how the current concept developed, and these are pretty well accepted definitions internationally.

Dr. Nucera: After the diagnosis, what about the treatment of this kind of epilepsy? What do you know about the efficacy of progesterone in catamenial epilepsy?

Dr. Herzog: We have evidence now that treatment with cyclic synthetic progesterones has not benefited women with catamenial epilepsy significantly. We know from the NIH phase III clinical trial that progesterone was not superior to placebo when treating women with focal-onset epilepsy overall. A prespecified post-hoc analysis, however, showed that progesterone was highly effective in women who had the perimenstrual or C1 pattern of seizure exacerbation, and the level of 50% responders correlated with the level of catamenial seizure exacerbation.

The C2 and C3 patterns did not respond, but the study was designed to start treatment on day 14 of the cycle, and C2 and C3 patterns start with a preovulatory estrogen surge on day 10, so it was in hindsight not designed to treat more than half of the women with catamenial epilepsy.

Dr. Nucera: What is the linkage between hormones and catamenial epilepsy? Are the hormonal changes causing the epilepsy, or are they exacerbating an underlying process?

Dr. Herzog: Hormones can both potentiate kindling, and also be a factor in seizure exacerbation. So they can promote epileptogenesis or retard epileptogensis, depending on which hormone, estrogen or progesterone, but mainly they are modulatory factors that influence an underlying process.

There’s clear evidence that these three patterns of catamenial epilepsy relate temporally very closely to changes in hormone levels or hormonal ratios. The C1 or perimenstrual pattern occurs when there’s a rapid withdrawal of progesterone and hence allopregnanolone, a potent GABA modulator, premenstrually. That starts around day 3. The level starts dropping a week before menstruation, but in the last two days of the cycle there is a massive drop, from 20 to 30 picograms per milliliter (pg/ml) down to 1. This is like suddenly withdrawing from phenobarbital or benzodiazepines or alcohol, and will produce seizures. That’s the C1 pattern.

The C2 pattern aligns with the massive preovulatory estradiol surge, which starts around day 10 of the menstrual cycle. Estradiol starts surging from a level above 40 pg/ml to 250, 300, sometimes over 400 pg/ml. Estradiol, being a proconvulsant, can start the C2 pattern, whereas the withdrawal of progesterone triggers the C1 pattern.

The C3 pattern is a combination of the preovulatory estradiol surge starting on day 10, and then during the entire luteal phase, in the absence of ovulation, you’ll have a high estradiol-to-progesterone ratio, which can exacerbate seizures.

Dr. Nucera: What happens when after, for example, a hysterectomy or menopause? When there are hormonal changes?

Dr. Herzog: I think there’s clear evidence that women with catamenial epilepsy are very sensitive to changes in hormonal state. Claudia Cagnetti in Italy showed compelling evidence that women with catamenial evidence showed a substantial reduction in seizures during pregnancy, much more than women with epilepsy that doesn’t show a catamenial pattern.

Cynthia Harden showed that women with catamenial epilepsy tend to experience an exacerbation of seizures in the years leading up to menopause, but they experience a reduction in seizures after menopause, post menopause. My impression is, I suspect the perimenopausal exacerbation of seizures occurs when women transition from ovulatory cycles, usually in their 40s, to anovulatory estradiol predominating cycles, and that post menopausally—about a year after the last menstrual period, when menses stop and hormonal production of estradiol stops—then they experience an improvement.

Dr. Nucera: Why are reproductive disorders, such as polycystic ovarian syndrome (PCOS) and abnormal menstrual cycles, more common in women with epilepsy, compared with women who don’t have epilepsy?

Dr. Herzog: The hypothalamus regulates the pituitary, which controls ovarian folliculogenesis and ovarian hormone production. The hypothalamus is modulated by supra-hypothalamic influences, mainly from the limbic system and especially the temporal limbic system, most notably the amygdala, a common site of epilepsy origin or epilepsy involvement. It has massive direct connections to those parts of the ventromedial hypothalamus which are involved with the regulation, production and secrtion of gonadotropin releasing hormone (GnRH) pulses.

In a study in 2003, women with unilateral left or right temporolimbic epilepsy, if you do EEG monitoring and blood samples every 10 minutes from 8 am to 4 pm, the women with left temporal epilepsy have more GnRH pulses than normal controls, and women with right temporal epilepsy have fewer GnRH pulses.

If you increase GnRH pulse frequency, that induces the pituitary to secrete more LH and less FSH, follicle stimulating hormone. If you have less FSH, then you have failure of ovarian follicle maturation and then these partially developed follicles don’t ovulate; they accumulate in the ovary as small ovarian cysts. If an ovarian follicle doesn’t mature, it doesn’t make aromatase. So the testosterone the follicle makes is not converted to estradiol and it’s secreted as testosterone. So you have hyperandrogenic anovulation. This is the simplest definition of PCOS.

When you have fewer pulses, lower pulse frequency, you get decreased gonadotropin secretion, failure of follicle maturation, and a failure of the menstrual cycle – so you get amenorrhea. So you get hypothalamic amenorrhea with right epilepsy, and PCOS with left epilepsy.

With non-temporal epilepsy, the situation changes. For example, right frontal epilepsy is associated more with PCOS.

Dr. Nucera: What is the role of anti-seizure medications? What is their effect on female reproductive health?

Dr. Herzog: Valproate is generally considered to be the anti-seizure medication (ASM) with the most adverse effects. Valproate retards folliculogenesis in animal studies, and likely does so also in humans. You get more anovulatory cycles with valproate. It’s also an aromatase inhibitor, so it blocks the testosterone conversion to estradiol, so you get this hyperandrogenic chronic anovulation, which is PCOS development, with valproate.

Valproate is also the most teratogenic of the anti-seizure medications, moreso than topiramate, which may be second. Lamotrigine and levetiracetam tend to have low congenital malformation rates, comparable to those in women in the general population.

Dr. Nucera: And the newer anti-seizure medications? For example, perampenel?

Dr. Herzog: We don’t have much information on perampanel. It has a unique mechanism of action, blocking noradrenergic release, but I don’t think we have much information about it.

Dr. Nucera: What do we know about male patients? What happens in males with epilepsy?

Dr. Herzog: Testosterone is interesting, insofar as it can have dual, opposing effects on seizure threshold, at least theoretically. Testosterone can be aromatized to estradiol, which would have proconvulsant effects. It also can be reduced to dihydrotestosterone and further reduced to androstenediol, which like allopregnanolone is a potent positive allosteric modulator of the GABA-A receptor and therefore is an anti-convulsant.

Men tend to develop hypogonadism, and that has negative effects on libido, and reduced potency, and quality-of-life assessments and depression scales show that low testosterone in men with epilepsy is also associate with reduced energey, competitive drive, and mood. Testosterone supplements, in our 2010 Epilepsy and Behavior publication of 40 men with epilepsy and hypogonadism, testosterone supplements markedly and significantly improved scores of libido, potency, energy, competitive drive, and mood. Interestingly, testosterone also reduced seizure frequency in that study by 40%.

Enzyme-inducing ASMs lower bioavailable testosterone. There are two mechanisms, at least two. Enzyme-inducing medications induce hepatic production of sex hormone binding globulin, or SHBG. That tightly binds testosterone, and the loosely bound or albumin bound testosterone and free testosterone are the bioavailable testosterone. 

The SHBG production reduces bioavailable testosterone. A less well known explanation, which I think is a major factor, is that enzyme-inducing ASMs also induce aromatase, which increases the conversion of testosterone to estradiol. Although estradiol in men makes up only 1% of the total levels of steroids, it exerts 50% of the negative feedback on the hypothalamo-pituitary axis, thereby preventing the mechanism that regulates bioavailable testosterone levels to function effectively.

Most anti-seizure medications, including carbamazepine, oxcarbazepine, phenobarbital, phenytoin, valproate, and levetiracetam, have been associated with abnormal semen analysis, lower sperm count, lower sperm motility, and dysmorphic sperm forms. Lamotrigine has had the least effect on hormones and on sperm formation. So that seems to have the most favorable profile.

With regards to fertility, most studies have suggested that men with epilepsy have lower fertility rates, lower numbers of offspring. These are mostly old studies from the 1980s and 1990s, but a recent Finnish study shows only about two-thirds of the expected offspring. On the other hand, an Icelandic study found no differences in offspring compared with the general population. With regard to fetal malformation there is no evidence in that regard, that male use of ASMs leads to malformations in offspring.

Evidence also shows that epilepsy itself is a factor – so there are social factors, the stress of epilepsy, lower marital rates, there are anti-seizure medication effects on fertility because of sperm abnormalities, and there is evidence that epilepsy itself may be associated with more semen abnormalities before an anti-seizure medication is introduced.

There’s also evidence that men with right temporolimbic epilepsy have more hyposexuality than men with left temporal limbic epilepsy. It’s a complex issue, many factors.

Dr. Nucera: What is most important for physicians to undrstand regarding reproductive hormones and epilepsy?

Dr. Herzog: I think the most important takeaway is there is a reciprocal interaction between hormones and epilepsy such that reproductive steroids or their neuroactive metabolites can influence seizure thresholds and seizure occurrence, and that epileptiform activity in turn can affect the pulsatile secretion of GnRH by the hypothalamus and thereby influence ovarian function, leading to disruption of ovarian function, which leads to reproductive endocrine disorders.

In the NIH Progesterone Trial, we found that during the 3-month baseline, 92 women had both an ovulatory and an anovulatory cycle, at least one of each, and these women had 29.5% more seizures in their anovulatory cycles than in their ovulatory cycles. So the formation of reproductive endocrine disorders, which can be promoted by epilepsy, also poses a risk for seizures.

Finally, if there’s a role for hormones in the occurrence of seizures there’s also a role for hormones in treatment. This has to be qualified by stating that in susceptible individuals. The NIH progesterone trial and the more recent trials have failed to show superiority of these neuroactive steroids over placebo because they did not select for suscpetible individuals. When we looked at the logical treatment – namely, women with premenstrual seizure exacerbation, which is due to progesterone and intense allopregnanolone withdrawal—the progesterone treatment was successful.

Likewise, in post-partum depression (PPD), brexanolone (a positive allosteric modulator of GABAA) was successful because PPD occurs in women who show an emotional change secondary to a massive hormonal change in the ninth month of pregnancy—namely, the rapid withdrawal of progesterone and allopregnanolone.

Currently, susceptibility is identified by recording seizures and menses, and if two of three cycles show catamenial epilepsy, we make that designation. This is very cumbersome. This year in Epilepsia we published a study showing that susceptibility of seizures to hormonal effects may represent heritable traits which extend to a broader range of clinical sensitivities. We showed that in women with epilepsy who had a positive family history of alcohol use disorder, the alcohol use disorder was a highly specific predictor of catamenial epilepsy—it had 92% specificity,but had a sensitivity only of 28.8%.

So if a woman with epilepsy and intractable seizures has a positive family history of alcohol use disorder, it’s worth investigating for catamenial epilepsy, which may respond to hormaonal treatment. But we currently need a much more efficient biomarker and hopefully, work will be done to identify a heritable gene or other chemical factor.

Dr. Nucera: Thank you so much for discussing this topic with me. Anything else?

Dr. Herzog: We need to do follow-up studies on the NIH Progesterone Trial. It was a post-hoc analysis which found that perimenstrual seizure exacerbation responds to progesterone, and I think the NIH would like to have a phase III randomized, prosective, placebo-controlled study where the primary outcome looks at whether women with perimenstrual seizure exacerbation respond better to progesterone than placebo.

For the C2 and C3 patterns, you have to suppress the preovulatory estradiol surge. Currently, the only way to do that is to suppress the entire menstrual cycle. It’s been shown in one open-label study that medroxyprogesterone can do this. There’ve also been case reports showing that a GnRH analogue can suppress the menstrual cycle and benefit seizures.

Our own experience for C2 and C3 patterns – these women didn’t do well with starting progesterone supplements earlier, and but suppressing their menstrual cycle with a GnRH analogue and then supplementing each day, with estradiol in a patch and progesterone in the form of lozenges or capsules, women have normal hormonal expsure during the month without the preovulatory estrogen surge and the big drop premenstrually in progeterone. You can maintain a stable estradiol and progesterone regimen after suppressing their menstraul cycle and they’ll have normal hormonal exposure and, in our experience, less seizures.

Dr. Nucera: To a young epileptologist like me what do you suggest for growth, professional growth in this field?

Dr. Herzog: I think – it’s interesting that the field was much ignored. In fact, when we started in the 1980s, the consensus was, based on previous studies, that there is no such thing as catamenial eilepsy. I think our 1997 paper showed the reason why there was no consensus was beause each investigator used a different deifnition. Most just looked for more seiures perimenstrually, but ignored other patterns. I think this formula with the reverse S-shaped curve is a scientific basis for how to designate women with catamenial epilepsy.

Neurologists don’t like hormones. Even in our NIH progesterone trial – these were all experienced investigators in doing phase III studies, but to explain to women about hormones and the menstrual cycle and folliculogenesis—this was not part of their training. Yet the body does not distinguish neurological aspects from endocrine aspects from gynecological aspects. 

I think being able to be conversant in more than just the neurological aspect is important. They say “Well we send the patient to the gynecologist.” But for the gynecologist, epilepsy makes up only 1% of their practice. What we need is to develop neurologists who are conversant in more than just their isolated subspeciality and who will answer basic questions, like do hormones affect seizures?

In 2010 we formed a net-based survey – the Epilepsy Birth Control Registry, where over 1,600 women participated. We published like 18 articles on the safety and efficacy of various forms of contraception in the setting of various types of anti-epileptic drugs. Thes are the basic questions that neurologists who are intrested in gender specific or sex specific issues should address. There is a future in this, but again even in getting NIH grants, I think hormones are a fringe.

For years I’ve worked on women with epilepsy more than men, and I think men with epilepsy are doing better overall since there are fewer enzyme-inducing  drugs used. But they are still in use and the recognition of hypogonadism is very important – those men develop low energy, depression, drive. I think it’s important to recognize the issues in men with epilepsy.

Dr. Nucera: Thank you very much.

Dr. Herzog: I appreciate the opportunity to participate and talk about this. Thank you.