- Top of page
Purpose: Neuroactive sex steroids influence neuron excitability, which is enhanced by estradiol (E2) and decreased by progesterone (Pg). In epilepsy, the production, metabolism, biologic availability, and activity of sex hormones may be affected by seizures themselves or by antiepileptic drugs (AEDs). This cross-sectional observational study was aimed at evaluating the relationships between sex steroids, seizure frequency, and other clinical parameters in women with partial epilepsy (PE) on AED treatments.
Methods: Serum E2, Pg, sex hormone binding globulin (SHBG) levels, free E2 (fE2), and E2/Pg ratios were determined during the follicular and luteal phases in 72 adult women with PE, and in 30 healthy controls. Hormonal data were correlated with seizure frequency, age, body weight, body mass index (BMI), disease onset and duration, and AED therapies.
Results: In patients, E2, fE2, and Pg levels were lower in both ovarian phases, whereas those of SHBG were higher than in controls. No significant changes in hormone levels and in prevalence of anovulatory cycles were observed between patients grouped according to their seizure frequency. However, when compared with those in healthy controls, luteal fE2 and Pg levels were chiefly impaired in women with more frequent seizures, mostly undergoing AED polytherapies, but not in those with absent or rarer seizures.
Conclusions: The actual changes in sex steroid levels and E2/Pg ratios did not explain an increased seizure frequency in adult women with AED-treated PE, but patients with more severe disease showed more relevant changes in their sex hormone profile and impaired Pg levels during the luteal phase.
Therefore, it has been hypothesized that spontaneous fluctuations in sex hormone levels and reproductive endocrine disorders, which are unusually common both in treated and untreated epilepsy, may conversely influence seizure occurrence (Herzog, 2008a,b). We have recently found that sex steroid levels, although globally decreased, did not correlate with seizure frequency in consecutive women with different types of epileptic syndrome, hormone changes being rather explained by antiepileptic drug (AED) treatments (Galimberti et al., 2009).
To avoid confounding factors related to different epilepsy types, in this cross-sectional observational study we have analyzed serum sex steroid levels and the ratios between serum estradiol (E2) and Pg levels throughout the ovarian cycle in women with partial epilepsy (PE), with an aim to assess: (1) differences in sex hormone levels versus healthy controls; and (2) the possible relationships between sex steroid set up, seizure frequency, and other clinical parameters. In fact, the elucidation of the role of hormone changes in influencing seizures and health in women with epilepsy might make possible an improved comprehensive management and treatment of these patients.
- Top of page
In this study, we have evaluated consecutive outpatients with AED-treated PE and attempted to correlate ovary function and sex steroid data with seizure frequency and other clinical parameters.
Interictal discharges and generalized or partial seizures may impact reproductive function at different levels by affecting gonadotropin pulsatile secretion, hypothalamic–pituitary control on the gonadal function, and sex hormone production (Drislane et al., 1994; Nappi et al., 1994; Cummings et al., 1995; Murialdo et al., 1997; Bauer et al., 1998; Bilo et al., 2001; Herzog, 2008a). In temporal lobe epilepsy (TLE), limbic discharge is held responsible for changes in the release of hypothalamic hormones, which regulate gonadotropin secretion and ovary function (Cummings et al., 1995; Herzog et al., 2003; Herzog, 2008a). However, it should be stressed that we have found no difference in sex hormone parameters between seizure-free cases (SFS 1) and those with occasional or more frequent crises. Therefore, the presence of epilepsy rather than seizure frequency seems to be relevant in inducing reproductive dysfunction and sex hormone changes in females with AED-treated PE.
Moreover, most of the AEDs considered in this study induce the activity of the cytochrome P450 (CYP) oxidative system and activate the synthesis of SHBG by the hepatocyte, thereby interfering with steroid metabolism and bioavailability (Perucca et al., 1984; Murialdo et al., 1998; Morrell et al., 2001; Isojärvi et al., 2005; Isojärvi, 2008). In addition, we have found an inverse correlation between body weight and SHBG levels. This aspect should be carefully pondered in patients with epilepsy because changes in SHBG levels and sex hormone free fractions must be considered as one of the markers of obesity and cardiovascular disease (Tchernof & Despres, 2000).
The most intriguing finding in this study concerned the relationship between epilepsy severity and sex hormone levels, which were different from those in healthy controls in patients with more frequent seizures but not when seizures were rarer or absent. An earlier onset and longer duration of the disease, in addition to a larger prevalence of AED polytherapies, characterized the groups with higher SFS. Younger age and longer AED treatment were also found to be predictors of reproductive endocrine disorders in women with epilepsy taking VPA (Isojärvi et al., 2005).
Neither total E2 nor Pg levels in our patients correlated significantly with clinical parameters, but an inverse relationship was found between fE2 levels and SFS during the follicular phase, and with disease duration in the luteal one. PE women with more frequent seizures showed higher mean SHBG levels than those in SFS 1–2 groups, but the differences were not statistically significant because of the large dispersion of values. The finding may be attributed to the longer duration of AED treatments and the more frequent use of enzyme-inducing AEDs in subjects with more severe seizures. Such changes in serum SHBG along with lowered total E2 levels implied a decrease in the biologically active fE2 titers in women with more frequent seizures and longer disease duration. Therefore, the impairment of E2 biologic activity due to the decrease of its free fraction, rather than changes in total E2, may be ascribed to the seizure frequency, disease severity, and the stimulating effect of AED on SHBG synthesis.
We did not analyze the effects of single AEDs on sex hormones in our patients because of the great variety of pharmacologic treatments and the small number of cases in some groups. Most of our PE women were treated with traditional enzyme-inducing AEDs, such as PB and CBZ, whereas VPA, which is more frequently associated with reproductive dysfunction (Isojärvi, 2008), was administered in only a negligible number of subjects.
Estrogens are produced from androgens through the activity of aromatases, and the inhibition of these enzymes by some AEDs may involve a decreased production of E2, mainly when polytherapies with multiple interfering drugs are administered. An inhibition of the aromatase complex (CYP19) by most of the AEDs used in our patients, but not by CBZ and PRM, has been demonstrated in vitro (Jacobsen et al., 2008), and possibly also occurs in vivo.
The majority of clinical studies used endometrial biopsy as the standard to diagnose luteal phase defects. However, in common practice, measurements of serum Pg may be considered as a reliable means of identifying anovulation or luteal dysfunction. Although the best correlation with endometrial histology is achieved by repeated Pg assays obtained during days 5–9 after ovulation, the current study was designed with only one sample in the mid-luteal phase. Pg levels obtained in this way are clearly an indirect parameter of the menstrual cycle investigation. With this limitation taken into account, anovulatory cycles were found in 23.6% and a luteal phase defect with reduced Pg levels (6.37–15.8 nmol/L) in a further 15.3% of our PE women, so that an overall cumulative frequency of 38.9% of patients had serum Pg levels lower than 15.8 nmol/L. This limit has been considered by other authors as indicative of anovulation when combined with basal temperature monitoring. Our prevalence findings with respect to anovulation and luteal phase defect in menstruating women with PE are similar to those reported in the literature (Cummings et al., 1995; Murialdo et al., 1997; Bauer et al., 1998; Herzog & Friedman, 2001; Morrell et al., 2002; Herzog et al., 2004; Herzog, 2006).
These prevalences appear to be higher than that currently quoted in the general population, ranging between 8.0 and 10.9% (Abdulla et al., 1983); but the actual frequency of the luteal phase failure in this form of epilepsy remains uncertain, being possibly influenced by case selection, investigational procedures, and methodologic approaches.
The seizure elicitation by the physiologic surge of sex steroids during the luteal phase and by the subsequent premenstrual Pg drop has been emphasized, so that catamenial seizures might be more frequent in women with Pg surge indicative of ovulatory cycles than in anovulatory ones (Bauer et al., 1998).
The impairment of Pg secretion during the luteal phase we observed may be only in part ascribed to AED regimens. Anovulation has been reported to be more frequent in women with TLE taking polytherapy than in those receiving a single AED, but also in this instance the difference was not statistically significant and appeared to be unaffected by seizure frequency (Cummings et al., 1995).
Lowered sex hormone levels in epilepsy may be, in part, accounted for by the AED therapy, which involves a decrease in estrogen production and conversion from androgens in liver and peripheral tissues. In addition, both seizures and AEDs may affect hormone secretions and gonadotropin ovulatory surge by directly targeting a number of anatomic substrates, including the limbic system, hypothalamus, pituitary, and gonads (Herzog, 2008a).
Neuroexcitatory effects of estrogens and anticonvulsant properties exerted by progestins have been quoted previously. However, in our PE patients with more frequent seizures, E2 and fE2 levels were rather decreased, and E2/Pg or fE2/Pg ratios did not explain higher seizure rates. Therefore, actual circulating E2 levels did not appear to be correlated with an increased seizure occurrence despite the well-known effects of estrogens on neuron excitability largely supported by in vitro evidence. But hormone changes potentially explaining an enhanced seizure frequency rather concerned luteal Pg levels, in view of the anti-epileptogenic activity attributed to progestins, in part due to their actions on γ-aminobutyric acid (GABA)A receptor–mediated chloride conductance (Bäckström, 1976; Beyenburg et al., 2001).
E2 and Pg exert favorable biologic and metabolic effects in females, and these aspects should be comprehensively assessed for an optimal evaluation of women with epilepsy. Menstrual disorders and sex hormone levels should be monitored closely in these patients to prevent adverse effects of reproductive dysfunction and ovarian failure, mainly in women with more frequent seizures often exposed to AED polytherapies.