To the Editors:
We agree with the letter of Dr. Hussain, indicating that the use of hormone replacement therapy (HRT) in postmenopausal women with epilepsy is an important and complex issue. It merits much more attention in light of the recent demonstration that HRT leads to significant health risks.
We also agree that HRT may be problematic because the commonly prescribed forms of HRT do not utilize the natural ovarian steroids, estradiol, and progesterone. For example, Premarin, an estrogen preparation that is widely used for HRT, is a complex combination of equine steroids, only some of which are actually estrogenic. As Dr. Hussain indicates, medroxyprogesterone acetate, which is often used as the progestin component of HRT, is also problematic. Medroxyprogesterone is a synthetic progestin that acts not only on progesterone receptors but also is a glucocorticoid receptor agonist (Ishida et al., 2002; Cirizza et al., 2006), suggesting that its administration will not induce the same effects as progesterone.
As Dr. Hussain pointed out, a potentially useful strategy is one that would more closely simulate the normal physiology of the female. One reason for this is that synthetic analogs may not exert all of the actions of natural steroid hormones. This is particularly true for progesterone because natural progesterone is metabolized to a neurosteroid (allopregnanolone), but synthetic progestins such as medroxyprogesterone acetate are not. Metabolism to allopregnanolone is potentially important in epilepsy because allopregnanolone has anticonvulsant effects (Rhodes et al., 2004). Therefore, natural progesterone would be likely to have greater anticonvulsant effects than synthetic progestins, and this indeed appears to be the case (Herzog, 1999). Another reason is that hormones may exert different effects if they are administered in a way that mimics normal secretion but not in other ways. For example, it may be important to develop a HRT regimen that involves cyclical rather than continual hormone administration, given that ovarian secretion of gonadal hormones is not constant. Indeed, chronic exposure to gonadal hormones often leads to a loss of effect. Chronic exposure to estradiol, for example, leads to changes in the levels of steroid receptors, and loss of efficacy can result (for further discussion, see Scharfman and MacLusky, 2006). Again, the simulation of normal physiology may be especially important in epilepsy because administration of progesterone for a period of time that matches the duration of the luteal phase appears to provide improved seizure control in women with epilepsy (Herzog, 1986).
There is an additional factor that should be emphasized: HRT is administered to women with epilepsy who are distinct physiologically and pharmacologically from the woman of reproductive age, because the women are older, have had a longer history of seizures, and have experienced menopause. Age influences seizure susceptibility, and it also changes pharmacodynamics, so it is likely it will influence the effects of HRT. A long history of recurrent seizures may also influence the efficacy of HRT because of progressive changes to neurons, glia, or the vasculature, which are all potential targets of hormone action. Menopause clearly influences seizures in women with epilepsy (Harden et al., 1999), and the latency from menopause to treatment can be an important variable. For example, the length of time between menopause and the onset of HRT seems to be an important determinant of age-related dementia (Maki, 2006). Therefore, one must not only consider HRT per se, but also the changes in the CNS of the women with epilepsy that are due to chronic seizures, age, and the loss of ovarian function.