Womens Center for Behavioral Endocrinology, McLean Hospital (H.J.), Belmont, Massachusetts 02478; and Perinatal and Reproductive Psychiatry Clinical Research Program, Department of Psychiatry (H.J.), and Reproductive Endocrine Unit, Department of Medicine (A.E.T., J.E.H.), Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114
Address correspondence and requests for reprints to: Hadine Joffe, M.D., McLean Hospital, 115 Mill Street, Belmont, Massachusetts 02478. E-mail: hjoffe{at}partners.org
Previous studies have described an association between epilepsy and features of the polycystic ovarian syndrome (PCOS) among women receiving treatment with antiepileptic drugs (AEDs), including valproic acid (VPA) (1, 2, 3, 4, 5). The association with PCOS has been attributed to epilepsy itself by some investigators (2), and to the use of VPA by others (1). VPA is an anticonvulsant used to treat epilepsy, migraine headaches, and bipolar disorder (6, 7, 8, 9). Data regarding the potential association of epilepsy and VPA with PCOS are sparse and suffer from substantial methodological and analytical problems. Despite the limitations of these data and the lack of causative evidence supporting an association between VPA use and PCOS, treatment of women with VPA-responsive disorders has been influenced by these reports. The study by Bilo et al. (10), in this issue of the journal, provides critical insight into the relationship between epilepsy, VPA use, and PCOS, and enriches our understanding of the role of the central nervous system (CNS) in PCOS.
The prevalence of PCOS in epilepsy is reported to be between 3.1% and
26% (Table 1) (1, 2, 3, 4, 10).
This information is derived from a total of five clinical case series
involving 20238 premenopausal-aged women receiving outpatient care in
epilepsy centers in the United States (2), Italy (3, 10), Germany (4), and Finland (1). One
study (1) did not report the prevalence of the clinical
syndrome of PCOS, as defined by the 1990 NIH consensus criteria
(11); however, we used the frequency of certain PCOS
features reported to estimate that the prevalence of PCOS in this
series was between 3.1% and 7.1%. In population-based studies, PCOS
has been estimated to occur in 4.06.8% of premenopausal women
(12, 13, 14). The prevalence of PCOS in epileptic populations
was elevated in three of five studies (Table 1
) (2, 3, 10)
and significantly so in two studies, including the study by Bilo
et al. (10).
|
An examination of the report of Bilo et al. (10) and other clinical series reveals significant disagreement among them in their conclusions about the association between epilepsy and PCOS (1, 2, 3, 4, 10). Several factors may explain the discrepancy in the findings. The studies are all relatively small, permitting chance observations and selection bias to exert substantial effects. The referral pattern for each of the epilepsy clinics in which studies were conducted may differ such that some may over-represent women with reproductive-endocrine disorders. In addition, the type of epilepsy (generalized, focal), location of seizure focus (temporal lobe, extra-temporal), and responsiveness to therapy of study subjects varies among the clinical reports. Associations between PCOS and generalized epilepsy (3), temporal-lobe epilepsy (2), and no specific epilepsy type or seizure focus location (1, 10) have been reported. Moreover, the frequency of use of specific AEDs and the proportion of women receiving no treatment for their epilepsy varies substantially among the clinical reports (1, 3, 4, 5, 15, 16). If an association of epilepsy with PCOS is mediated or caused by one or more AEDs, the number of subjects receiving that medication must be large enough to permit accurate analysis. Finally, personal characteristics of study subjects that may modulate the relationship between epilepsy and PCOSsuch as ethnicity and body weightvary markedly among the clinical reports.
There are two primary hypotheses to explain an association between epilepsy and PCOS (17). The first explanation is that the seizure disorder itself increases the risk of PCOS. Two studies, including the report by Bilo et al. (10), found that women with seizure disorders who were unmedicated had an elevated prevalence of PCOS, and that PCOS occurred as frequently in unmedicated epileptic women as it did in medicated women, regardless of the specific AED used (2, 10). The greater occurrence of PCOS in untreated epileptic women suggests that epilepsy itself is responsible for the association.
How could epilepsy increase the risk for PCOS? The pathogenesis of PCOS is probably multifactorial, and abnormalities in hypothalamic function, ovarian morphology, and insulin resistance have been described (18). Neuroendocrine abnormalities including increased LH amplitude pulsations, fast frequency LH secretion, and low to normal levels of FSH are present in most women with PCOS (19). Accelerated GnRH pulsatility may be an intrinsic abnormality in PCOS (20). In epileptics, ictal and interictal paroxysmal discharges may disrupt GnRH pulsatility, modulating CNS regulation of GnRH neurons by excitatory neurotransmitters (21). Receptors for the excitatory neurotransmitters glutamate and nitric oxide, including N-methyl-D-aspartate receptors, are located in hypothalamic nuclei known to be important for GnRH release (22, 23). The changes in excitatory neurotransmitter systems associated with epilepsy may potentially increase the risk of PCOS via modulation of GnRH pulsatility (24). The finding of increased LH pulse frequency in unmedicated epileptic women with regular menstrual cycles supports this hypothesis (16, 25).
An alternate explanation for the association between epilepsy and PCOS
is that PCOS is induced by use of VPA (17). An association
between VPA use and features of PCOS in epileptic women has been
reported in one study (1). In this study, ultrasonographic
evidence of polycystic ovarian morphology occurred in 14 of 31 (45.2%)
epileptic women treated with VPA, a significant increase over that seen
in epileptic women treated with other AEDs (1). The
prevalence of clinical PCOS among VPA users was not specifically
reported. However, we can use the information reported about the
frequency of specific PCOS features to estimate that 36 of 31
(9.719.4%) VPA-treated epileptic women met criteria for PCOS (Table 2) (1). A range of PCOS
prevalence is calculated because it is unclear whether the women with
hirsutism (n = 3) were the same as those with menstrual
irregularity and hyperandrogenemia (n = 3) in this series. Other
clinical reports have not found an association between VPA use and PCOS
in epileptic women, including that of Bilo et al. (4, 10). In the latter report, PCOS occurred in 23.1% of VPA users
and 23.8% of those receiving other AEDs (10).
|
The potential pathophysiology underlying a VPA-mediated increase in
PCOS prevalence is unknown. VPA may influence the risk of PCOS by its
CNS activity, but a mechanism that explains a specific effect of VPA on
PCOS has not been found (17). VPA increases
-aminobutyric acid synthesis and release and may block
N-methyl-D-aspartate-type glutamate
receptors, but other AEDs share these CNS properties (27, 28). VPA administration also alters the density of
-aminobutyric acid-ergic inputs to GnRH neurons in the medial
preoptic area of the hypothalamus in pubertal mice, but this effect
does not occur in postpubertal mice (29, 30, 31). An
alternative explanation for why VPA increases the risk of PCOS is
related to the association between VPA use and weight gain, which
increases insulin resistance and, potentially, the risk for PCOS
(26, 32). VPA use causes more weight gain than
carbamazepine (6, 33). However, obesity alone is unlikely
to fully explain the association of VPA use with PCOS as some
VPA-treated epileptic women are lean (10), and other
medications that cause significant weight gain (e.g.
clozapine) have not been associated with PCOS. A pharmacodynamic
property of VPA that might contribute to an association with PCOS is
that VPA lacks the effects on induction of hepatic cytochrome P450
(CYP) enzymes associated with other AEDs (17). Induction
of CYP isozymes facilitates clearance of gonadal steroids and reduces
circulating testosterone levels (34). Because VPA does not
induce CYP enzymes, it lacks these mitigating actions against
hyperandrogenemia, which may contribute selectively to the increased
risk of PCOS.
If VPA increases the risk of PCOS, it is likely that the pathophysiology is multifactorial and that several factors act simultaneously to cause this reproductive-endocrine disorder. Epilepsy may be required as a component factor in the association between VPA and PCOS. To date, the prevalence of PCOS is not known to be increased in women taking VPA for other indications, such as bipolar or migraine disorders. One small pilot study of PCOS in 22 women with bipolar disorder found no association of VPA with PCOS (35). It is plausible that the effect of VPA on PCOS may be specific to epileptic women if ictal and interictal discharges and altered GnRH pulsatility are required to increase risk for PCOS in the setting of VPA exposure. Neuroendocrine abnormalities that occur in epilepsy may make women with seizure disorders particularly susceptible to the effects of VPA on the hypothalamic-pituitary-ovarian axis (36). More data in nonepileptic VPA users are needed to determine whether VPA alone increases the risk of PCOS, or whether the association is found only in women with seizure disorders.
Despite the strength of the observations in the paper by Bilo et al. (10), there is no conclusive evidence that epilepsy or the use of VPA causes PCOS. All studies describing a relationship between epilepsy, VPA use, and PCOS suffer from several major methodological limitations. All studies are cross-sectional in nature, and none reports whether epilepsy preceded the onset of PCOS or whether PCOS began before or after the initiation of VPA for treatment of epilepsy. Epilepsy and VPA use must be established to precede the onset of PCOS symptoms to infer that epilepsy or VPA are risk factors for PCOS. Given the typically perimenarchal onset of PCOS in the general population (37), and the absence of data about the temporal sequence of epilepsy, VPA use, and PCOS, PCOS may precede the onset of epilepsy and its treatment.
Clinicians prescribing VPA for seizure, bipolar, or migraine disorders should be aware of the contradictory data describing the relationship between epilepsy, VPA use, and PCOS. There are currently two multicenter studies with longitudinal components underwayone in epileptic women and the other in bipolar womenthat will address whether VPA use increases the risk of PCOS. These large studies will provide information about the temporal relationship of the initiation of VPA and the onset of PCOS symptoms so that clinical decisions about the use of VPA can be informed by reliable data.
Received May 11, 2001.
Accepted May 13, 2001.
References