Epilepsia, 47(9):1447–1451, 2006
Blackwell Publishing, Inc.
C ?2006 International League Against Epilepsy
Hormone Replacement Therapy in Women with Epilepsy: A
Randomized, Double-Blind, Placebo-Controlled Study
∗Cynthia L. Harden, †Andrew G. Herzog,∗Blagovest G. Nikolov, ‡Barbara S. Koppel, §Paul J.
Christos, †Kristen Fowler,∗Douglas R. Labar, and ?W. Allen Hauser
∗Comprehensive Epilepsy Center, Department of Neurology and Neuroscience, Weill Medical College of Cornell University, New
York, New York; †Beth Israel Deaconess Medical Center, Harvard Neuroendocrine Unit, Boston, Massachusetts; ‡Department of
Neurology, New York Medical College, Metropolitan Hospital Center, §Division of Biostatistics and Epidemiology, Department of
Public Health, Weill Medical College of Cornell University, and ?G.H. Sergievsky Center, Department of Neurology and Department
of Epidemiology, College of Physicians and Surgeons of Columbia University, New York, New York, U.S.A.
Summary: Purpose: Previous reports have suggested that hor-
HRT to the medication regimen of postmenopausal women with
epilepsy was associated with an increase in seizure frequency.
Methods: This was a randomized, double-blind, placebo-
controlled trial of the effect of HRT on seizure frequency
in postmenopausal women with epilepsy, taking stable doses
of antiepileptic drugs (AEDs), and within 10 years of their
last menses. After a 3-month prospective baseline, subjects
were randomized to placebo, Prempro (0.625 mg of conjugated
CEE/MPA) daily, or double-dose CEE/MPA daily for a 3-month
ing baseline. The subjects’ ages ranged from 45 to 62 years
(mean, 53 years; SD, ±5), and the number of AEDs used ranged
from none to three (median, one). Five (71%) of seven sub-
quency of at least one seizure type, compared with four (50%)
of eight taking single-dose CEE/MPA and one (17%) of six tak-
ing placebo (p = 0.05). An increase in seizure frequency of the
subject’s most severe seizure type was associated with increas-
ing CEE/MPA dose (p = 0.008). An increase in complex partial
dose (p = 0.05). Two subjects taking lamotrigine had a decrease
in lamotrigine levels of 25–30% while taking CEE/MPA.
Conclusions: CEE/MPA is associated with a dose-related
increase in seizure frequency in postmenopausal women
with epilepsy. CEE/MPA may decrease lamotrigine lev-
Key Words: Seizures—Epilepsy—Postmenopausal—
Menopause—Hormone replacement therapy.
menopausal women have stable reproductive hormone
levels. When menstrual cycles no longer occur, the circu-
terone levels and elevated folliculin (follicle-stimulating
(1). Women with epilepsy of reproductive potential are at
risk for anovulatory menstrual cycles (2) and catamenial
seizure exacerbations (3). These risks, particularly cata-
menial seizure exacerbations, are likely caused by inter-
actions between the epileptic brain and reproductive hor-
mones in the brain. These risks are no longer relevant at
menopause, and further, it has been reported that women
Accepted November 19, 2005.
Address correspondence and reprint requests to Dr. C.L. Harden at
Comprehensive Epilepsy Center, Room K-615, 525 East 68th Street,
Weill Cornell Medical Center, New York, NY 10021, U.S.A. E-mail:
with epilepsy may experience a decrease in seizure fre-
quency at menopause, particularly if a catamenial exacer-
bation occurred in the reproductive years (4).
For postmenopausal women with epilepsy, hormone
replacement therapy (HRT) presents another potential
source of hormonal influence on seizure activity. The ef-
fect of HRT in women with epilepsy has not been sys-
tematically investigated. One survey suggests that it may
exacerbate seizures (4). This would seem plausible, given
that estrogen is proconvulsant in several animal models of
epilepsy, including amygdalal kindling and pentylenete-
trazol administration in ovariectomized rats (5). Estro-
gen induces the formation of new excitatory synapses
in the CA1 region of the hippocampus, and further, this
estrogenic induction involves activation of N-methyl-D-
aspartate (NMDA) receptors (6). Increasing the complex-
for the proconvulsant activity of estrogen. Standard hor-
mone replacement, which includes estrogen and a pro-
gestin, in postmenopausal women with epilepsy can be
1448 C. L. HARDEN ET AL.
postulated to have an effect on seizures that is more ev-
ident than that of oral contraceptives in cycling women
with epilepsy, because reproductive hormone levels dur-
ing menopause are low and unchanging. Therefore the
brain hormonal milieu in which exogenous hormones are
introduced is markedly different in menopause from that
in menstruating women.
Therefore the specific objective of this study was to
determine in a prospective manner if standard HRT in-
teria for use of HRT prior to July 2002, which included no
thromboembolic history or history of breast cancer, and
were candidates for the routine use of HRT at the time
for relief of menopausal symptoms and prevention of os-
teoporosis and cardiovascular disease. In July 2002, the
results of the Women’s Health Initiative (WHI) study al-
tered the previously accepted standard for HRT use (7).
The main findings were that HRT is not recommended
the risks of HRT outweigh the benefits, because of in-
ease, stroke, and pulmonary embolism.
PATIENTS AND METHODS
This study was a randomized, placebo-controlled,
double-blind clinical trial of the effect of HRT compared
with placebo on seizure frequency in postmenopausal
women with epilepsy. This study was approved by the
Committees on Human Studies at the Weill Cornell Med-
ical Center and at Beth Israel Deaconess Medical Cen-
ter. The HRT chosen for this study was Prempro, which
is 0.625 mg of conjugated equine estrogens plus 2.5
mg of medroxyprogesterone acetate, designated here as
CEE/MPA. CEE/MPA was used because it was the most
widely prescribed form of HRT at the time of this study
inception and therefore would give the broadest applica-
bility to the results. A dose effect of CEE/MPA on seizure
frequency was assessed by using two doses: a single and
a double dose. Double-dose CEE/MPA has been used
in clinical practice for further reduction in menopausal
symptoms that persist with a single dose. Therefore the
three treatment arms consisted of single-dose CEE/MPA,
daily dose capsules for each arm.
(b) being postmenopausal (≥1 year without menses un-
and FSH levels, within 10 years of their last menstrual pe-
riod. Subjects enrolled in this study met standard medical
criteria for use of HRT prior to July 2002.
Subjects were recruited from the epilepsy practices at
the Weill Cornell Medical Center and at Beth Israel Dea-
coness Medical Center. Data were collected at both sites
and entered into a database at the Weill Cornell Medical
Center. Subjects underwent a prospective 84-day baseline
period on their stable optimal AED dosages. During the
baseline phase, seizure frequency was documented, and
hormone and AED levels were measured. Subjects were
prospectively for the next 84 days, without alteration in
AED dose. Study visits occurred every 4 weeks during
the study. Safety of HRT use was monitored according to
medically accepted gynecologic practice, which includ-
ing clinically monitoring for thromboembolic events and
through vaginal bleeding. The exit criterion for increased
line, exit criteria were two seizures during the treatment
phase or one seizure if, in the opinion of the investigator,
the subject should discontinue the study drug because of
Sample size was initially calculated with an estimate of
seizure increase in 15% of the placebo-treated subjects,
30% of single-dose HRT-treated subjects, and 45% of the
double-dose HRT–treated subjects, with a power of 80%
and two-tailed significance level of 0.05. The estimated
sample size was 40 subjects in each arm, for a total of 120
randomized subjects. However, recruitment was stopped
after consideration of the WHI study results (7), and this
report is on the recruited randomized subjects.
of each of the three treatment arms per block (placebo,
single-dose CEE/MPA, and double-dose CEE/MPA) gen-
erated in a random order. The purpose of blocked ran-
domization was to ensure an equal number of patients
in the three study arms after every six patients, in the
event that the study ended before the targeted accrual was
complete. Furthermore, because it was unknown whether
subjects with low versus high seizure frequency would
respond differently to any effect of HRT on seizures, the
randomization procedure was stratified by seizure sever-
one seizure/month, more than one to four seizures/month,
and more than four seizures/month) were defined, and
blocked randomization (as defined earlier with blocks of
six) was performed within each seizure-frequency stra-
in the three study arms, within each seizure-frequency
category. The study pharmacist encapsulated study drugs
according to the randomized sequences in blocks of six
and created study drug containers numbered sequentially
with randomization numbers. Investigators and study co-
ordinators enrolled subjects and assigned them to their
seizure-frequency strata at the time of randomization (af-
ter the prospective baseline phase) and then randomized
the subjects to study drugs/placebo sequentially within
Epilepsia, Vol. 47, No. 9, 2006
HRT IN WOMEN WITH EPILEPSY 1449
the blocks of six assigned to that seizure-frequency stra-
tum. All investigators and study coordinators remained
blinded as to the treatment arm until the study database
was completed and closed. Subjects also were blinded as
to treatment arm. The biostatistician was aware of which
subjects were in each treatment arm but remained blinded
as to the exact treatment regimen of each arm. Only the
study safety monitor and the study pharmacist, who en-
capsulated the study drug, had access to the unblinding
code. Neither of these study personnel enrolled subjects
monitor was to be performed only in situations of medical
Baseline characteristics between treatment arms were
compared by using univariate analysis of variance
(ANOVA) or the nonparametric Kruskal–Wallis test, as
appropriate. The primary end point of this study was to
ing HRT treatment, and this was assessed by compar-
ing the proportion of subjects with increased seizure fre-
quency between the three treatment arms. The χ2test for
trend was used to examine the potential dose relation be-
tween treatment status (i.e., placebo, single dose, double
dose) and increased seizure frequency (yes/no). Change
in seizure frequency was determined by comparing the
daily seizure rate during baseline with the daily seizure
rate while the study drug was taken. The clinical impact
of seizure increase was assessed by determining whether
a subject had an increase in her most severe seizure type.
cantly associated with decreased health-related quality of
increased in frequency, but the complex partial seizures
did not, the subject was determined not to have a worsen-
ized tonic–clonic > complex partial > simple partial. All
p values are two-sided, with statistical significance evalu-
ated at the 0.05 alpha level. All analyses were performed
in SPSS v13.0 for Windows.
Subjects were recruited from September 2000 to June
2003, including 6 months of suspension of enrollment
in 2002 because of the WHI results (7). The study was
stopped in 2003 because of safety concerns raised by the
WHI results (7). Eight subjects were in baseline at the
During the study, three subjects withdrew consent during
baseline because of the WHI results (7), and 11 subjects
they did not meeting inclusion criteria, mostly because of
not meeting criteria to take HRT safely. These reasons
included history discovered during baseline of uterine fi-
broids, history of any thromboembolic disease or stroke,
and virginity (making it difficulty to do an endometrial
biopsy should abnormal bleeding occur). This remainder
of this report is confined to the subjects who were ran-
baseline. The subjects’ ages ranged from 45 to 62 years
a median of one AED. All subjects had a diagnosis of par-
tial epilepsy. During the study period, eight subjects had
simple partial seizures, nine subjects had complex partial
seizures, and three had secondarily generalized seizures.
drug during the trial. Six subjects were randomized to the
placebo group; eight, to the single-dose CEE/MPA group;
line characteristics are presented in Table 1. The number
of seizures presented in Table 1 is the median, minimum,
and maximum number for the entire baseline period in
each treatment arm. No differences between treatment
arms were present at baseline for age, age at menopause,
number of AEDs, and all but one of the seizure-frequency
nificance (p = 0.04), although the median number of sim-
ple partial seizures was near identical for the three groups
(Table 1). The AEDs taken by subjects in each treatment
arm are listed in Table 2.
Five subjects in the double-dose CEE/MPA arm dis-
continued before completion of the treatment period: one
because of seizure increase after 39 days (met exit cri-
teria); three because of HRT-related adverse effects after
30, 39, and 60 days, and all three also had increases in
seizure frequency of either complex partial or secondarily
generalized seizures; and one was lost to follow-up after
35 days of treatment (visit 1 of treatment phase). The sub-
ject lost to follow-up had no seizures during baseline or
treatment. The subject who met exit criteria because of
seizure increase did not have HRT-related side effects.
Seven subjects had no seizures of any type during base-
line; one of these had a seizure during treatment. In no
subject did a new or worsened seizure type develop than
she had experienced before the study while taking the
study drug. Seizure-frequency increases with HRT treat-
ment were mild, and only one subject met the exit criteria
of seizure-frequency doubling.
One subject in the single-dose CEE/MPA arm dis-
continued early after 32 days, because of HRT-related
adverse effects, and she did not have seizure increase dur-
ing treatment. Only one subject in the placebo arm dis-
continued early after 51 days of treatment, because of
the WHI study results and stopping of this study; she
had no seizures during baseline or treatment periods. The
Epilepsia, Vol. 47, No. 9, 2006
1450 C. L. HARDEN ET AL.
TABLE 1. Characteristics of subjects in each treatment arm (N = 21) with number of seizures during 84-day baseline period
age in age inseizures
Treatment group(yr) ± SDa
Placebo (n = 6)
Single CEE/MPA (n = 8)
Double CEE/MPA (n = 7)
ap = 0.85 by ANOVA test.
bp = 0.80 by ANOVA test.
cp = 0.04 by Kruskal–Wallis test.
dp = 0.41 by Kruskal–Wallis test.
ep = 0.20 by Kruskal–Wallis test.
fp = 0.43 by Kruskal–Wallis test.
gp = 0.30 by Kruskal–Wallis test.
(yr) ± SDb
(median; min, max)c
(median; min, max)d
0; 0, 0
0; 0, 16
1; 0, 189
0; 0, 5
1; 0, 19
0; 0, 7
0; 0, 3
0; 0, 0
0; 0, 1
1; 0, 5
1; 0, 19
2; 0, 189
2; 0, 2
1; 1, 2
1; 1, 3
HRT-related adverse effects resulting in discontinuation
included headache, vaginal spotting, and pelvic discom-
The association of increased seizure frequency with in-
creasing CEE/MPA dose was significant for an increase
in seizure frequency of any seizure type after HRT treat-
ment (p = 0.05 by χ2test for trend) (see Table 3). Fur-
ther, the association of increased complex partial seizure
frequency with increased CEE/MPA dose also was signif-
icant (p = 0.05) (Table 3). The association of increased
most-severe seizure type with increased CEE/MPA dose
was significant (p = 0.008). Because of curtailment of
enrollment, comparison of seizure-frequency change by
seizure-frequency strata could not be performed.
Little change in AED levels was associated with treat-
ment arms, although the numbers of subjects taking spe-
cific AEDs was small. In the two subjects randomized
to CEE/MPA who also were taking lamotrigine (LTG), a
ceiving polytherapy, LTG levels declined from a mean of
9.6 μg/ml (range, 10.1–9.2) during baseline, to a mean of
in seizures. In another subject receiving monotherapy, the
TABLE 2. AEDs in each treatment arm
AEDsPRM, PB PRM TPM, CBZ,
AED, antiepileptic drug; CBZ, carbamazepine; CZP, clonazepam;
GBP, gabapentin; LTG, lamotrigine; LEV, levetiracetam; PB, pheno-
barbital; PHT, phenytoin; PRM, primidone; TPM, topiramate; TGB,
tiagabine; VPA, valproic acid.
LTG levels declined from a mean of 2.1 μg/ml (range,
2.6–1.5) during baseline to a mean of 1.4 μg/ml (range,
1.3–1.5) during treatment, and this subject did have in-
creased seizures during treatment. Therefore LTG levels
declined by 25–30% in these subjects. Both subjects were
in the single-dose CEE/MPA treatment arm.
Because in one subject, the seizure increase could be
explained by a decrease in LTG levels, statistical analyses
were repeated, eliminating this subject from the analysis,
and the results showed no material change.
ciated with dose-related increases in seizure frequency in
postmenopausal women with epilepsy. Although recruit-
study and the number of subjects necessary to prove that
seizure frequency was increased with each seizure type
was not obtained, the available data show that seizure oc-
currence increased with CEE/MPA use in a dose-related
Seizure increase is thought to be due to the estro-
genic component of HRT; however, this protocol can-
not distinguish between the effect of estrogen or medrox-
yprogesterone acetate on seizure activity. More scientific
evidence exists for the seizure-promoting potential of es-
to increased synthetic progestin may have been related to
The decrease in LTG levels with HRT use is consistent
with previous reports of decreased LTG levels during oral
contraceptive use and with pregnancy (10–13). Increased
LTG clearance in the setting of elevated circulating repro-
ductive hormones is thought to be due to effects on the
glucoronidation pathway of LTG metabolism, leading to
a mean decrease in levels of >50% during oral contra-
ceptive use (10) and an increase in clearance of 40–330%
during pregnancy (11–13).
Epilepsia, Vol. 47, No. 9, 2006
HRT IN WOMEN WITH EPILEPSY 1451 Download full-text
TABLE 3. Number of subjects with increase in seizure frequency by treatment arm (N = 21)
groupn (%)n (%)n (%)n (%)n (%)n (%)
Placebo (n = 6)
Single CEE/MPA (n = 8)
Double CEE/MPA (n = 7)
ap = 0.88 by χ2test for trend.
bp = 0.05 by χ2test for trend.
cp = 0.10 by χ2test for trend.
dp = 0.05 by χ2test for trend.
ep = 0.008 by χ2test for trend.
This study, although limited in its ability to be gener-
alized because of both the small number of subjects in
this study and the precipitous decline of HRT use, pro-
vides information about the potential detrimental effect
of CEE/MPA on seizures for women with epilepsy and
supports the previously reported survey results (4).
Certainly women with epilepsy have taken HRT in the
past without seizure exacerbations and will need to do so
in the future for short-term management of menopausal
symptoms, specifically hot flashes, insomnia, and vagi-
nal atrophy. This study does not imply a contraindica-
tion to HRT in women with epilepsy but does suggest
that CEE/MPA may not be the optimal HRT regimen for
women with epilepsy. Possibly a safer HRT regimen in
women with epilepsy would be an estrogen along with
propriate for women with epilepsy because natural pro-
gesterone has neuroactive metabolites such as allopreg-
nanolone, which has been shown to be a potent positive
allosteric modulator at the γ-aminobutyric acid (GABA)
plement may reduce seizure frequency in women with
catamenial epilepsy (19,20). No reports of effect of the
synthetic progestin used in this study, medroxyproges-
terone acetate, on seizure threshold are available.
These results also raise the possibility that HRT de-
creases LTG levels, which may have clinical significance
for some patients. This is likely due to the same mecha-
nism as is present when LTG levels decline in the setting
of pregnancy and oral contraceptive use.
NS38473 and the Weill Cornell Medical Center GCRC.
This study was supported by RO1-
1. Burger HG. The endocrinology of the menopause. Maturitas
2. Morrell MJ, Giudice L, Flynn KL, et al. Predictors of ovulatory
failure in women with epilepsy. Ann Neurol 2002;52:704–711.
3. Herzog AG, Harden CL, Liporace J, et al. Frequency of catamenial
seizure exacerbation in women with localization-related epilepsy.
Ann Neurol 2004;56:431–434.
4. Harden CL, Pulver MC, Jacobs AR. The effect of menopause and
perimenopause on the course of epilepsy. Epilepsia 1999;40:1402–
kindled by repeated amygdala stimulation or pentylenetetrazol ad-
ministration in ovariectomized female rats. Epilepsia 1986;27:103–
6. McEwen B. Estrogen actions throughout the brain. Recent Prog
Horm Res 2002;57:357–384.
7. The Writing Group for the Women’s Health Initiative Investiga-
tors. Risks and benefits of estrogen plus progestin in healthy post-
postmenopausal women: principal results of the Women’s Health
Initiative randomized, controlled trial. JAMA 2002;288:321–333.
8. Johnson EK, Jones JE, Seidenberg M, et al. The relative impact of
anxiety, depression and clinical seizures features on health-related
quality of life in epilepsy. Epilepsia 2004;45:544–550.
9. Vickrey BG, Berg AT, Sperling MR, et al. Relationships between
seizure severity and health-related quality of life in refractory
localization-related epilepsy. Epilepsia 2000;41:760–764.
10. Sabers A, Ohman I, Christensen J, et al. Oral contraceptives reduce
lamotrigine plasma levels. Neurology 2003;61:570–571.
in lamotrigine pharmacokinetics: a monotherapy study. Neurology
12. Tran TA, Leppik IE, Blesi K, et al. Lamotrigine clearance during
pregnancy. Neurology 2002;59:251–255.
13. Pennell PB, Newport DJ, Stowe ZN, et al. The impact of preg-
nancy and childbirth on the metabolism of lamotrigine. Neurology
14. Spritzer PM, Vitola D, Vilodre LC, et al. One year follow-up of hor-
tension. Exp Clin Endocrinol Diabetes (Germany) 2003;111:267–
with a transdermal estradiol gel and oral micronized progesterone:
monersatztherapie mit einem perkutan wirksamen Oestradiol-Gel
klimakterische Symptome und den Lipidstoffwechsel.] Wien Klin
Wochenschr (Austria) 2000;112:629–633.
16. Frye CA, Rhodes ME, Walf A, et al. Progesterone reduces
pentylenetetrazol-induced ictal activity of wild-type mice but not
those deficient in type I 5alpha-reductase. Epilepsia 2002;43(suppl
17. Lonsdale D, Burnham WM. The anticonvulsant effects of pro-
gesterone and 5alpha-dihydroprogesterone on amygdala-kindled
seizures in rats. Epilepsia 2003;44:1494–1499.
18. Reddy DS, Kim HY, Rogawski MA. Neurosteroid with-
drawal model of perimenstrual catamenial epilepsy. Epilepsia
19. Herzog AG. Progesterone therapy in women with complex par-
tial and secondary generalized seizures. Neurology 1995;45:1660–
20. Herzog AG. Progesterone therapy in women with epilepsy: a 3-year
follow-up. Neurology 1999;52:1917–1918.
Epilepsia, Vol. 47, No. 9, 2006