0013-7227/02/$15.
00/0
Printed in U.S.A.
Daily Low-Dose Mifepristone Has Contraceptive
Potential by Suppressing Ovulation and Menstruation:
A Double-Blind Randomized Control Trial of 2 and 5 mg
per Day for 120 Days
AUDREY BROWN, LINAN CHENG, SUIQING LIN,
Contraceptive Development Network (A.B., D.T.B.), Centre for Reproductive Biology, University of Edinburgh, Edinburgh,
EH3 9ET United Kingdom; and Shanghai Institute of Family Planning Technical Instruction (L.C., S.L.), International
Peace Maternity and Child Health Hospital, China Welfare Institute, Shanghai 200030, People’s Republic of China
Daily administration of progesterone (P) antagonists to
women inhibits ovulation and disrupts endometrial function.
In this double-blind randomized trial, we have explored the
contraceptive potential of two doses of the P antagonist mife-
pristone in healthy volunteers in Edinburgh and Shanghai.
Ninety-eight women (58 in Edinburgh and 40 in Shanghai)
were randomized to receive either 2 or 5 mg mifepristone daily
for 120 d. Ovarian activity was monitored by the weekly mea-
surement of steroid metabolites in urine and of E2 and P in
plasma every month. Endometrial function was assessed by
menstrual records, and ultrasound measurement of endome-
trial thickness was assessed every month. Endometrial biopsy
was collected on d 12 of the control cycle and after 60 and 120 d
of treatment.
Ninety women (50 in Edinburgh and 40 in Shanghai) com-
pleted the study. Follicular activity continued during treat-
ment with both doses in Edinburgh women, although ovula-
tion was suppressed in the majority of cycles (90 and 95% of
A LTHOUGH THE COMBINED oral contraceptive pill
containing an estrogen and a progestogen is a highly
effective and popular form of contraception, concern about
risks to health associated with estrogen, e.g. venous throm-
boembolism, have pointed to the need to develop alternative
hormonal methods that contain no estrogen (1). The proges-
togen-only pill has been available for over 30 yr but remains
relatively unpopular, accounting for only 8% of the market
in the United Kingdom (2). Problems, including a relatively
high failure rate, unpredictable effect on the pattern of men-
strual bleeding, and the development of functional cysts,
result in a high rate of discontinuation.
Progesterone (P) is necessary for the establishment as well
as the maintenance of pregnancy. Mifepristone and other
antigestogens are now licensed for the termination of preg-
nancy in a number of countries worldwide (3, 4). However,
the contraceptive potential of these compounds is much less
explored (5, 6). It has been demonstrated previously that
administration of mifepristone in daily doses as low as 2 mg
will inhibit ovulation and prevent the formation of a secre-
tory endometrium (7–9). Menstrual bleeding was delayed
until the mifepristone was stopped after 30 d. Others have
Abbreviations: COC, Combined oral contraceptive; Cr, creatinine;
E1G, estrone glucuronide; P, progesterone.
63
Downloaded from jcem.endojournals.org by on December 2, 2009
The Journal of Clinical Endocrinology & Metabolism 87(1):63–70
Copyright © 2002 by The Endocrine Society
AND DAVID T. BAIRD
cycles in 2- and 5-mg groups, respectively). The women in
Shanghai showed evidence of ovulation in only 3 of 160
months of treatment (2 in 2-mg group and 1 in 5-mg group). The
majority of women in both centers were amenorrheic (65% in
2-mg group and 88% in 5-mg group in Edinburgh, and 90% in
both dose groups in Shanghai). The endometrial thickness
increased significantly in women in Edinburgh and decreased
in Shanghai; histology showed either atrophic or cystic
changes without evidence of hyperplasia. There was no preg-
nancy reported in the 200 months of exposure in 50 sexually
active women who had used no other method of contraception
during the study.
We conclude that mifepristone in low daily doses inhibits
ovulation and induces amenorrhea in the majority of women
and has the potential to be developed as a novel estrogen-
free oral contraceptive pill. (J Clin Endocrinol Metab 87:
63–70, 2002)
demonstrated that ovulation is delayed or inhibited when
lower doses, e.g. 1 mg, are given for up to 5 months (10, 11).
However, it is uncertain whether the minor effects that are
observed on the endometrium would be incompatible with
the establishment of pregnancy (12).
We have chosen to investigate the effect of the daily ad-
ministration of mifepristone in a dose that would inhibit
ovulation reproducibly. In this study, we report the effects of
a double-blind randomized control trial comparing the ef-
fects of administration of 2 and 5 mg mifepristone per day
for 120 d on ovulation and pattern of menstrual bleeding. We
tested contraceptive efficacy in a subset of women who were
exposed to the risk of pregnancy.
Materials and Methods
Ninety-eight healthy volunteers with regular menstrual cycles (25–35
d) aged 18 – 40 yr were recruited for the study (58 women in Edinburgh
and 40 women in Shanghai). Studies were approved by local ethical
committees (Institutional Review Board) at both centers. All women
gave written informed consent before being enrolled and were screened
before entering the study. Screening included routine physical and gy-
necological examination and measurement of height, weight, blood
pressure, and pulse. Blood samples were collected for measurement of
full blood count, urea and electrolytes, liver function tests, random
glucose, PRL, lipids, and cortisol. Human CG was also measured to
exclude pregnancy before entering the trial. In the Edinburgh center, the
day on which screening blood samples were collected varied throughout
64 J Clin Endocrinol Metab, January 2002, 87(1):63–70
the cycle before the treatment cycle, whereas in Shanghai, samples were
all collected on d 1 of the pretreatment cycle.
Subjects were studied for one pretreatment cycle, for four cycles of
treatment (120 d), and for one posttreatment cycle. Subjects were ran-
domly allocated to receive 2 or 5 mg mifepristone daily for the 120
treatment days. The randomization was achieved from a table of random
numbers and stratified by dose in groups of 10. The daily doses were
issued in prepackaged identical bottles containing either 2⫻ 1 mg mife-
pristone plus placebo or 1⫻ 5 mg plus two placebos. The investigators
and subjects were unaware of the dose that each subject was taking.
All subjects in Shanghai used this method as the sole method of
contraception for the 4 months of treatment. In the Edinburgh center, 10
women used the method as the sole contraceptive, the remainder being
surgically sterilized, not sexually active, or using barrier contraceptives.
Assessment of ovarian function
Ovarian function was monitored by measurement of ovarian steroids
in urine and plasma and by transvaginal sonography. All subjects col-
lected a weekly sample of early morning urine during the pretreatment
cycle, starting in the early follicular phase (d 1–5), during the four cycles
of treatment, and during the posttreatment cycle. Aliquots were frozen
and stored at ⫺20 C until assayed for estrone glucuronide (E1G), preg-
nanediol glucuronide, and creatinine (Cr). Blood samples were collected
before treatment, after 30, 60, 90, and 120 d of treatment, and in the
posttreatment cycle, and were assayed for E2 and P.
Ovarian follicular activity during treatment was compared with that
in the follicular phase of the pretreatment cycle. The baseline ovarian
function was taken to be the mean of the two samples collected in the
follicular phase of the pretreatment cycle, and the activity during treat-
ment was scored according to the following criteria: 1) totally sup-
pressed, E1G throughout treatment is less than 50% above the mean
baseline; 2) partially suppressed, E1G is raised at least 50% above the
mean baseline on one or two consecutive weeks (i.e. one episode of
follicular activity); 3) continued follicular activity, E1G raised at least 50%
above the mean baseline on at least two separate occasions (i.e. separated
by at least 2 wk). Ovulation was deemed to have occurred if the excretion
of pregnanediol glucuronide exceeded 0.5 mmol/mol Cr and was at
least 3-fold higher than that in the preceding week (13).
Each subject attended for transvaginal sonography on d 12 of the
pretreatment cycle, after 30, 60, 90, and 120 d of treatment, and in the
posttreatment cycle, when ovarian dimensions, follicular number and
diameter, and any ovarian cysts were measured.
Assessment of endometrial development
Endometrial thickness was measured by transvaginal sonography on
d 12 of the control cycle, after 30, 60, 90, and 120 d treatment, and in the
posttreatment cycle. Measurement was performed in the sagittal plane,
from one basal layer to the other; the diameter of any luminal fluid
present was subtracted from the measurement. Endometrial biopsies
were collected by Pipelle biopsy, fixed in normal buffered formalin, and
embedded in paraffin. Samples were collected on d 12 of the pretreat-
ment cycle and after 60 and 120 d of treatment. Sections were stained
with hematoxylin-eosin, and endometrial histology was assessed by two
observers who were blind to sample timing and dose group.
Assessment of menstrual bleeding pattern
Each subject was required to keep a menstrual bleeding diary for the
duration of the study. Each day was classified as no bleeding, spotting,
normal bleeding, or heavy bleeding.
Assay methods
All assays were performed in the laboratory of the center using
similar methods. Pregnanediol glucuronide was measured using a direct
enzyme immunoassay (working range, 0.25–32 ␮m/liter), whereas di-
rect immunoassay was used to measure E1G (working range, 8.4 –2140
nm/liter). In Edinburgh, the primary antibodies were rabbit E-3-G BSA
and anti P-3-G BSA for E1G and pregnanediol glucuronide, respectively,
obtained from the MRC/AFRC Comparative Physiology Research
Group (London, UK) and used as described previously (14, 15) from our
Downloaded from jcem.endojournals.org by on December 2, 2009
Brown et al. • Daily Mifepristone Contraception
laboratory. In Shanghai, kits using similar reagents were obtained from
Immunometrics Ltd. (London, UK). Intra- and interassay coefficients of
variation were 6 and 9% for E1G, and 10 and 13% for pregnanediol
glucuronide in Edinburgh; and 7.5 and 9.1% for E1G, and 8.9 and 10.3%
for pregnanediol glucuronide in Shanghai. Geometric means of daily
replicates were divided by the respective daily Cr concentration to
correct for variations in the dilution of the urine specimen.
E2 and P in plasma were measured by RIA. In Edinburgh, E2 was
assayed after ether extraction using I125 E2 3-carboxymethyl as label and
sheep antiserum (BW/26/9/90). P was measured by direct immuno-
assay using I125-11␣ progesterone glucuronide (Amersham Interna-
tional, Bucks, UK) as label, and rabbit antiprogesterone 11-hemisucci-
nate BSA antiserum. In Shanghai, E2 and P were assayed directly using
commercially available ELIZA kits (Biosource Technologies, Inc., Niv-
elles, Belgium). In Edinburgh, intra- and interassay coefficients of vari-
ation were 8 and 11% for E2, and 8 and 10% for P (9), whereas in Shanghai
the corresponding values were 4.4 and 7.6% for E2 and 6.8 and 7.4% for P.
Fifty samples that were assayed for E2 in both Shanghai and Edin-
burgh gave comparable values over the range measured (correlation
coefficient, 0.89)
Safety parameters
Each subject was reviewed in the pretreatment cycle, after 30, 60, 90,
and 120 d of treatment, and in the posttreatment cycle. At each visit,
blood pressure and pulse were measured, and blood was taken for full
blood count, urea and electrolytes, liver function tests, glucose, lipids,
PRL, and cortisol. In addition, each subject was asked to report any
health problems or adverse events that had occurred since the last visit.
Statistical methods
The power calculation was based on the numbers required to show
a significant difference in ovulation rates, which was assumed to be 25%
for 2 mg and 5% for 5 mg (7–9). It was calculated that 92 women would
be required to demonstrate with 80% power a significant difference (P ⬍
0.05) between doses.
Statistical analysis was carried out using SPSS (SPSS, Inc., Chicago,
IL). Wilcox signed rank test and paired t test were used to compare
plasma E2, urinary estrone, and endometrial thickness before and at
various time points throughout treatment. Analysis of covariance was
used to compare the dose effect on plasma E2 levels and on endometrial
thickness.
Results
Women in the two centers differed in their response to
treatment, and hence the results are considered separately
(Table 1).
In Edinburgh, a total of 58 women were recruited for the
study. Eight women withdrew prematurely from the study.
Two women withdrew due to gastric upset, one woman due
to irregular bleeding, and one woman was withdrawn by the
investigator because she developed hypertension. (Her
blood pressure rose from 135/85 mm Hg pretreatment to a
value of 150/100 mm Hg after 2 wk when she was discon-
TABLE 1. Characteristics of subjects receiving daily mifepristone
(mean ⫾ SE)
Edinburgh Shanghai
2 mg 5 mg 2 mg 5 mg
n 26 24 20 20
Age (yr) 30.1 ⫾ 1.1 30.9 ⫾ 0.9 33.3 ⫾ 0.6 32.5 ⫾ 0.8
Weight (kg) 62.8 ⫾ 1.6 68.0 ⫾ 2.7 62.0 ⫾ 0.6 60.8 ⫾ 0.8
Height (cm) 163 ⫾ 0.01 164 ⫾ 0.02 166 ⫾ 0.01 163 ⫾ 0.01
Parous 9 (35) 12 (50) 20 (100) 20 (100)
Previous abortion 6 (23) 6 (25) 18 (90) 20 (100)
Percentage is shown in parentheses.
Brown et al. • Daily Mifepristone Contraception J Clin Endocrinol Metab, January 2002, 87(1):63–70 65

tinued.) Four women withdrew for personal reasons. Thus,
50 women in Edinburgh (26 receiving 2 mg daily, and 24
receiving 5 mg daily) completed the study, and their results
are available for analysis of endocrine function.
Ovarian function
There was considerable variation in the pattern of excre-
tion of E1G, with some women showing cyclical fluctuations
representing growth and regression of ovarian follicles (Fig.
1A). In others, the level remained consistently low (Fig. 1C).
Overall, the mean urinary E1G level during treatment was
significantly lower than that in the pretreatment cycle in the
women treated with 5 mg mifepristone daily (12.9 ⫾ 0.4 vs.
15.0 ⫾ 0.9 ␮mol/mol Cr; P ⬍ 0.04) (Fig. 2). There was a small
but statistically insignificant fall in E1G excretion in the 2-mg
group (P ⫽ 0.06). The mean E1G level during the 4 months
of treatment was similar to that found during the follicular
phase of the pretreatment cycle in both groups (P ⬎ 0.1).
There was no significant change in the mean level of E2 in
samples collected monthly in either dose (Fig. 3) (mean ⫾ se;
2 mg, 478.1 ⫾ 60.7 vs. 408.5 ⫾ 28.8 pmol/liter, P ⫽ 0.26; 5 mg,
452.1 ⫾ 42.6 vs. 435.7 ⫾ 50.9 pmol/liter, P ⫽ 0.83).
The majority of women in Edinburgh showed evidence of
some follicular activity with both doses of mifepristone (Ta-
ble 2). Fewer women in the higher dose group (5 mg) than
in the 2-mg group (19 vs. 38%) ovulated once during treat-
ment, and there were fewer ovulatory cycles (10 of 104 vs. 5
of 96, respectively). Transvaginal sonography identified
eight fluid-filled ovarian cysts (30 – 63 mm) in the Edinburgh
group, four in each dose group. All were asymptomatic and
resolved spontaneously during treatment.
All 40 women recruited for the study in Shanghai com-
pleted the study. The mean E1G excretion during the pre-
treatment cycle was significantly lower in Chinese than in
Edinburgh women (9.2 ⫾ 0.6 vs. 16.4 ⫾ 0.7 ␮mol/mol Cr; P ⬍
0.001), as was the plasma E2 concentration (216 ⫾ 8.7 vs.
466 ⫾ 37.3 pmol/liter; P ⬍ 0.001). In contrast to the Edin-
burgh women, during treatment in Chinese women there
was a significant suppression of mean excretion of E1G in
both 2-mg (9.6 ⫾ 0.89 vs. 4.0 ⫾ 0.12 ␮mol/mol Cr; P ⬍ 0.001)
and 5-mg (8.9 ⫾ 0.71 vs. 3.8 ⫾ 0.10 ␮mol/mol Cr; P ⬍ 0.001)
doses (Fig. 4). The concentration of E2 was significantly (P ⬍
0.001) lower in both treatment groups by the end of the first
month (Fig. 5), and this depression was sustained throughout
treatment (mean ⫾ se, 2 mg, 210.2 ⫾ 9.4 vs. 124.0 ⫾ 3.4
pmol/liter; 5 mg, 221.4 ⫾ 14.8 vs. 124.4 ⫾ 2.3 pmol/liter; P ⬍
0.001).
During treatment, 60 –70% of women in both groups in
Shanghai showed complete suppression of ovarian activity,
and there was a total of only three ovulatory episodes (Table
2). Consistent with a profound degree of ovarian suppression

FIG. 1. Excretion of metabolites of ovarian steroids in women taking either 2 or 5 mg mifepristone per day for 120 d. E1G and pregnanediol
glucuronide are expressed per mol Cr. f, Menses. A, Persistent follicular activity but no ovulation in a woman in Edinburgh on 2 mg/d. B, Single
ovulatory episode in Edinburgh woman on 2 mg/d. C, Complete suppression of ovarian activity in Edinburgh women on 5 mg/d. D, single
ovulatory cycle in woman in Shanghai on 5 mg/d.

Downloaded from jcem.endojournals.org by on December 2, 2009

66 J Clin Endocrinol Metab, January 2002, 87(1):63–70 Brown et al. • Daily Mifepristone Contraception

FIG. 2. Excretion of E1G and pregnanediol glucuronide in women in

Edinburgh before, during, and after stopping treatment with mife-
pristone for 120 d (mean ⫾ SE). Overall, there was a significant fall
in E1G level in the women who received 5 mg (P ⬍ 0.04) but not 2 mg
mifepristone. A, 2 mg/d (n ⫽ 26); B, 5 mg/d (n ⫽ 24). FIG. 3. Concentration of E2 (top) and endometrial thickness (bottom)
in women in Edinburgh before, during, and after treatment with
mifepristone for 120 d. There was no significant change in the con-
was the fact that no ovarian cysts were detected during the centration of E2 of either dose (mean ⫾ SE; 2 mg, n ⫽ 26; 5 mg, n ⫽ 24).
monthly examination by transvaginal ultrasound.
TABLE 2. Degree of follicular activity and ovulatory episodes
Within 30 d of stopping treatment, ovulatory cycles re- during treatment with daily mifepristone
turned in all women in both Edinburgh and Shanghai.
Follicular activity
Ovulatory
Menstrual bleeding pattern Group n Total Partial Continued episodes
suppression suppression
In Edinburgh, 17 of the 26 women (65%) treated with 2 mg
mifepristone and 21 of 24 women (88%) given 5 mg were Edinburgh
2 mg 26 5 (19) 9 (35) 12 (46) 10
amenorrheic throughout the treatment phase. In Shanghai, 5 mg 24 9 (37) 10 (42) 5 (21) 5
18 of the 20 women (90%) in each dose group were amen- Shanghai
orrheic during treatment. 2 mg 20 12 (60) 7 (35) 1 (5) 2
In Edinburgh, the mean number of days of bleeding dur- 5 mg 20 14 (70) 3 (15) 3 (15) 1
ing treatment was 4.4 d for the 2-mg group and 0.6 d for the Percentage is shown in parentheses.
5-mg group. In Shanghai the mean number of days of bleed-
ing was 0.4 and 0.7 d for the 2- and 5-mg groups, respectively.
These differences in mean number of days of bleeding are who showed biochemical evidence of ovulation, the two
largely due to two women in the Edinburgh 2-mg group who women in the 2-mg group reported a subsequent menstrual
reported 24 and 53 d of bleeding during treatment. Analysis bleed, but the woman who ovulated on 5 mg daily did not
by Mann-Whitney U test did not show any statistically sig- (Fig. 1D).
nificant treatment difference in mean number of days of All subjects reported a menstrual bleed within 3 wk of
bleeding during treatment. stopping mifepristone treatment. There was a trend in both
Of those 15 women in Edinburgh (10 in the 2-mg group Edinburgh and Shanghai for this posttreatment bleeding to
and 5 in the 5-mg group) who showed biochemical evidence be longer than the normal pretreatment menses (Table 3).
of ovulation, only 4 (3 in the 2-mg group and 1 in the 5-mg However, the only difference that reached significance was
group) reported a menstrual bleed subsequent to the rise in the increase in the number of days of bleeding post treatment
pregnanediol excretion, suggesting asynchrony between the in the Edinburgh group treated with 2 mg (5.2 vs. 10 d; P ⬍
ovarian and endometrial cycles. Of the Shanghai women 0.05; Mann-Whitney U test).

Downloaded from jcem.endojournals.org by on December 2, 2009

Brown et al. • Daily Mifepristone Contraception J Clin Endocrinol Metab, January 2002, 87(1):63–70 67

FIG. 5. Concentration of E2 (top) and endometrial thickness (bottom)

FIG. 4. Excretion of E1G and pregnanediol glucuronide in women in in women in Shanghai before, during, and after treatment with mife-
Shanghai before, during and after treatment with mifepristone for pristone for 120 d. There was a significant (P ⬍ 0.001) suppression of
120 d. The excretion of E1G was significantly lower during treatment the mean concentration of E2 and endometrial thickness for both
in both doses than in the control cycle (P ⬍ 0.001). A, 2 mg/d (n ⫽ 20); doses (2 mg, n ⫽ 20; 5 mg, n ⫽ 20).
B, 5 mg/d (n ⫽ 20).
TABLE 3. Length of menses before and after mifepristone
Endometrial development
Days median (range)
Transvaginal sonography Before After
In the Edinburgh women, there was a trend during treat- Edinburgh
ment for the endometrial thickness to increase relative to the 2 mg 5.2 (4 –7) 10 (4 –37)a
baseline measurement on d 12 of the pretreatment cycle (Fig. 5 mg 5.7 (4 –9) 6.7 (3–15)
3). This increase in thickness was significant after 3 months Shanghai
2 mg 5.1 (4 –7) 5.7 (5–9)
of treatment in the 2-mg group (P ⫽ 0.020) and after 2 months 5 mg 4.8 (4 –7) 6.0 (4 – 8)
of treatment in the 5-mg group (P ⫽ 0.015). Analysis of
covariance did not show any treatment difference at any of
a
P ⬍ 0.05, Mann-Whitney U test.
the time points in the Edinburgh women.
In contrast, the Shanghai women showed a significant the majority of subjects showing either simple proliferative
reduction in endometrial thickness in both dose groups (2 or inactive endometrium with some cystic dilatation (Fig. 6).
mg, P ⫽ 0.003; 5 mg, P ⬍ 0.001) by the end of the first month One subject in the 5-mg group showed secretory changes in
of treatment (Fig. 5). This significant thinning of endome- the endometrium collected after 120 d of treatment. This
trium was maintained in the Shanghai women throughout correlated with biochemical evidence of preceding ovula-
the 4 months of treatment. Analysis of covariance showed a tion. Two subjects in the 2-mg group were in the luteal phase
treatment difference in the Shanghai women after 1 month following an episode of ovulation when the endometrium
of treatment (P ⫽ 0.034), but not at any other point. was collected. Neither of these samples showed secretory
changes. One Edinburgh subject in the 5-mg group showed
Endometrial histology
features of complex hyperplasia after 60 d of treatment, but
Normal endometrium was confirmed on all pretreatment after 120 d this subject’s endometrium had become inactive.
biopsies. In the Edinburgh women, the endometrial appear- No cases showed any evidence of cytological atypia.
ances were similar after both 60 and 120 d of treatment, with The endometrium from women in Shanghai showed sim-

Downloaded from jcem.endojournals.org by on December 2, 2009

68 J Clin Endocrinol Metab, January 2002, 87(1):63–70 Brown et al. • Daily Mifepristone Contraception

FIG. 6. Histology of the endometrium in six Edinburgh women who took mifepristone 2 or 5 mg/d for 120 d. The control endometrium on d 12
shows normal proliferative changes with straight glands. On d 60, both specimens show dilated glands lined with a single layer of epithelium
but no evidence of secretory changes. By d 120, the dense stroma persists. Some specimens show atrophic looking glands (top), whereas in others
there are dilated glands lined with inactive epithelium (hematoxylin and eosin, ⫻40).

ilar appearances to the Edinburgh tissue, with the majority
of samples collected during treatment showing simple pro-
liferative changes. There was no consistent difference be-
tween the appearances at 60 d and those at 120 d. Four
samples (two in each dose group) showed some secretory
features, although endocrinology did not support preceding
ovulation. There were no cases of endometrial hyperplasia in
the Shanghai women, and no cytological atypia was seen.
Safety parameters and adverse events
There were no changes in mean blood pressure and pulse
during treatment with either dose of mifepristone. Likewise,
monitoring of full blood count, serum biochemistry, and
lipids did not show any significant changes during the 4
months of treatment. There were no changes in the levels of
PRL and cortisol.
In the Edinburgh group, three subjects showed an eleva-
tion of alanine transaminase during treatment. The other
markers of liver function (␥ glutamyl transferase, bilirubin,
and alkaline phosphatase) remained within the normal
ranges. In two subjects (2 and 5 mg), the alanine transaminase
single elevated value was associated with a flu-like illness
and returned to within the normal range while treatment
continued. In the third subject (5 mg), who was asymptom-
atic, the level rose from 34 U/liter before treatment to 111
U/liter at the end of 120 d and fell to 39 U/liter 2 months after
stopping treatment.
In the Edinburgh group, six subjects reported hot flushes
during treatment. In all subjects, the plasma E2 level was
within the normal range, and the flushes resolved once the
study medication stopped. No hot flushes were reported in
the Shanghai group.
Contraceptive efficacy
All 40 women in Shanghai and 10 of the Edinburgh women
used no other method of contraception during the 4 months
of treatment, giving a total of 200 months exposure to preg-
nancy risk. No pregnancies occurred.
Discussion
This study extends our previous reports that mifepristone
in daily doses of 2 mg or more suppresses ovulation in the
majority of women (7, 9). In Caucasian women in Edinburgh,
a degree of follicular activity continued during treatment, as
indicated by the fluctuating levels of E1G excretion. Mife-
pristone inhibits the ability of estrogen to provoke an LH
surge (14), an effect that can be overcome by administration
of excess P (10). The incidence of ovulatory cycles (as de-
tected by a rise in pregnanediol excretion) was more with 2
mg (5%) than with 5 mg (2.5%) but was lower than had been
anticipated. In previous studies on which our power calcu-
lation was based, the mifepristone was only given for 1
month when many of the ovulations observed in the present
study occurred. However, occasional apparently ovulatory
cycles occurred as late as the last month of treatment and are
a potential source of concern as to whether pregnancy could
occur. However, the histological appearance of the endome-

Downloaded from jcem.endojournals.org by on December 2, 2009

Brown et al. • Daily Mifepristone Contraception
trium during mifepristone suggests that it is unlikely that
successful implantation could occur.
In Shanghai, ovarian activity was more profoundly sup-
pressed with both doses of mifepristone than in Edinburgh.
Levels of urinary E1G and plasma E2 were suppressed dur-
ing treatment to those found during the follicular phase of
the cycle. Moreover, fewer women ovulated, and the endo-
metrium was thinner than in Edinburgh women. The pos-
sible reasons for these differences include ethnic, dietary, and
body composition. The Chinese women were significantly
lighter than the women from Edinburgh and had a lower
body mass index. It is possible that the difference in diet may
alter the enterohepatic circulation, which is known to influ-
ence metabolism of steroids (16). The concentration of E2 in
the control cycle was significantly lower in women in Shang-
hai than those in Edinburgh. However, in Shanghai the sam-
ples were collected in the early follicular phase of the cycle,
whereas those in Edinburgh included women at mid-cycle
and the luteal phase. The values from women in the follicular
phase in both centers are similar, so we think it is unlikely
that differences are due to the differences in assays, which
correlated very well. Previous investigations have reported
lower levels of E2 and estrone sulfate in Chinese compared
with Caucasian women (17) as well as differences in urinary
metabolites (18).
There were differences in other parameters of ovarian
activity between the two groups of women. No ovarian cysts
were detected in the Chinese women, in keeping with con-
sistent suppression of ovarian activity. In contrast to the
women in Edinburgh, there was a significant decline in the
thickness of the endometrium in Chinese women in both
dose groups. Taken together with the suppression of ovarian
steroids, these observations all suggest that ovarian activity
is more easily suppressed by mifepristone in women in
Shanghai than those in Edinburgh.
Concern has been expressed previously that long-term
exposure to antigestagen compounds will carry a risk of
endometrial hyperplasia and possible malignancy by allow-
ing the endometrium continuous exposure to estrogen un-
opposed by P (19). A previous study reporting on daily
treatment with 1 mg mifepristone over five cycles did not
find any cases of endometrial hyperplasia, although a num-
ber of subjects were found to have a thickened endometrium
on ultrasonography (11). Our study has confirmed that there
were no cases of endometrial hyperplasia in 90 women after
4 months of exposure to daily low-dose mifepristone. One
subject showed hyperplastic change after 2 months of expo-
sure, but this had resolved to inactive endometrium by the
end of 4 months of treatment.
In the group of women treated in Edinburgh, endometrial
thickness was found to increase throughout the treatment
period, although it remained within the range found in the
luteal phase of the cycle. However, this thickening did not
appear to reflect exposure to unopposed estrogen because
even the cases with thicker endometrium did not show ev-
idence of hyperplasia. In monkeys, antiprogestogen induces
a dose-dependent decrease in endometrial thickness and
proliferation despite up-regulation of estrogen receptors (20,
21). A number of women showed inactive but dilated glands,
and it may be that a thickening apparent on ultrasonography
Downloaded from jcem.endojournals.org by on December 2, 2009
J Clin Endocrinol Metab, January 2002, 87(1):63–70 69
reflects the collection of fluid within the glandular lumen
rather than true thickening. In contrast, the women treated
in Shanghai did not follow this pattern. Indeed, there was
progressive reduction in endometrial thickness in the Chi-
nese women during the 4 months of treatment, and the
endometrium of the Chinese women did not show the cystic
dilatation that was frequently observed in the Edinburgh
tissue. The mechanism by which mifepristone prevents hy-
perplasia of the endometrium in the presence of unopposed
estrogen is not clear but has been observed with other an-
tigestogens (21, 22).
Edinburgh women showed more ovarian activity and had
higher plasma E2 levels during treatment than those in
Shanghai, and it may be that these differences result in vary-
ing endometrial exposure to circulating E2. Conversely, al-
though there was less ovarian activity in the Shanghai
women, there were more cases showing secretory change.
Previous work using 1 mg mifepristone daily also identified
cases of endometrial secretory change without a preceding
ovulation and luteal phase (11). It has been postulated that
mifepristone can have agonist effects on the endometrium,
particularly when endogenous P levels are very low (21). It
may be that the endometrial appearances in the Shanghai
women reflect this agonist activity that becomes manifest as
a result of the ovarian suppression and lack of endogenous P.
The major advantage of the combined oral contraceptive
(COC) is high efficacy combined with predictable pattern of
bleeding. Many of the more serious side effects associated
with COCs, e.g. venous thromboembolism, are largely due to
the estrogen component (1). Progestogen-only contraceptive
pills are less effective than COCs and are associated with
considerable upset in the pattern of menstrual bleeding. A
published study using 1 mg mifepristone per day demon-
strated that more than half the women showed irregular
bleeding or amenorrhea (11). Regular menses are maintained
when the dose is reduced to 0.5 mg/d, but unfortunately the
contraceptive efficacy is insufficient (12). The fact that the
majority of women in our study have amenorrhea or a mark-
edly reduced number of days of bleeding suggests that it
would be a popular option at least in some European coun-
tries (23). Suppression of ovulation combined with disrup-
tion in the formation of a secretory endometrium makes it
very likely that it would be a particularly effective contra-
ceptive, as suggested by our preliminary findings.
The results of this study confirm that a daily dose of 2 or
5 mg mifepristone has antifertility effects. There were no
pregnancies in the 50 sexually active women who were ex-
posed to the risk of pregnancy for a total of 200 months.
However, it is not possible to assess the contraceptive effi-
cacy accurately from these preliminary data, particularly
because the majority of women exposed to the risk of preg-
nancy (40) were Chinese in whom ovarian activity was sup-
pressed more effectively than in women in Edinburgh. The
2-mg dose allows more ovulatory episodes and has a less
predictable bleeding pattern. The 5-mg daily dose has a more
predictable pattern of amenorrhea and may therefore have
greater acceptability in practice. Longer term studies are
required to further assess the endometrial safety of daily
low-dose mifepristone and to confirm the contraceptive
efficacy.
70 J Clin Endocrinol Metab, January 2002, 87(1):63–70
Acknowledgments
We are grateful to Mrs. Ann Mayo and Dr. Karen Smith in Edinburgh
for help in running this study; Dr. Rob Elton for statistical advice; and
the Department for International Development and the Medical Re-
search Council, United Kingdom, for support for the Contraceptive
Development Network. The mifepristone was supplied by WHO Special
Programme of Research, Development and Research Training in Human
Reproduction (Project no. 96503).
Received July 18, 2001. Accepted September 19, 2001.
Address all correspondence and requests for reprints to: Prof. D. T.
Baird, Contraceptive Development Network, Centre for Reproductive
Biology, University of Edinburgh, 37 Chalmers Street, Edinburgh, EH3
9ET United Kingdom. E-mail: cdn@ed.ac.uk.
This work was supported by Grant G9523250 to the Contraceptive
Development Network.
References
1. Baird DT, Glasier AF 1999 Science, medicine and the future. Contraception.
BMJ 319:969 –972
2. Fraser IS 2000 Progestogen only contraception. Glasier A, Gebbie A, eds.
Handbook of family planning and reproductive healthcare. Edinburgh:
Churchill Livingstone; 77–103
3. Sitruk-Ware R 2000 Approval of mifepristone (RU486) in Europe. Zentrabl
Gynakol 122:241–247
4. Ullmann A 2000 The development of mifepristone: a pharmaceutical drama
in three acts. J Am Med Womens Assoc 55:117–120
5. Baird DT 1993 Potential contraceptive effects of antigestogens. In: Donaldson
MS, Dorflinger L, Brown SS, Benet LZ, eds. Clinical application of mifepristone
RU486 and other antiprogestogens. Washington, DC: National Academy Press;
143–163
6. Spitz IM, Van Look PF, Coelingh Bennink HJ 2000 The use of progesterone
antagonists and progesterone receptor modulators in contraception. Steroids
65:817– 823
7. Ledger WL, Sweeting VM, Hillier H, Baird DT 1992 Inhibition of ovulation
by low dose mifepristone (RU486). Hum Reprod 7:945–950
8. Croxatto HB, Salvatierra AM, Croxatto HD, Fuentealba B 1993 Effects of
continuous treatment with low dose mifepristone throughout one menstrual
cycle. Hum Reprod 8:201–207
9. Cameron ST, Thong KJ, Baird DT 1995 Effect of daily low dose mifepristone
on the ovarian cycle and on the dynamics of follicle growth. Clin Endocrinol
(Oxf) 43:407– 414
Downloaded from jcem.endojournals.org by on December 2, 2009
Brown et al. • Daily Mifepristone Contraception
10. Batista MC, Cartledge TP, Zellmer W, Nieman LK, Merriam GR, Loriaux DL
1992 Evidence for a critical role of progesterone in the regulation of the
midcycle gonadotropin surge and ovulation. J Clin Endocrinol Metabol 74:
565–570
11. Croxatto HB, Kovacs L, Massai R, Resch BA, Fuentealba B, Salvatierra AM,
Croxatto HD, Zalanyi S, Viski S, Krenacs L 1998 Effects of long-term low-dose
mifepristone on reproductive function in women. Hum Reprod 13:793–798
12. Marions L, Viski S, Gemzell-Daniellson K, et al. 1999 Contraceptive efficacy
of daily administration of 0.5 mg mifepristone. Hum Reprod 14:2788 –2790
13. Glasier AF, Irvine DS, Wickings EJ, Hillier SG, Baird DT 1989 A comparison
of the effects on follicular development between clomiphene citrate, its two
separate isomers and spontaneous cycles. Hum Reprod 4:252–256
14. Baird DT, Thong KJ, Hall C, Cameron ST 1995 Failure of estrogen induced
LH surge in women treated with mifepristone (RU486) every day for 30 days.
Hum Reprod 10:2270 –2276
15. Yong EL, Glasier AF, Ledger W, Caird L, Beattie G, Thong J, Baird DT 1992
Effect of cyclofenil on hormonal dynamics, follicular development and cervical
mucus in normal and oligomenorrhoeic women. Hum Reprod 7:39 – 43
16. Humpel M 1989 Comparative pharmacokinetics of selected contraceptive
steroids in animal species and man. Michael F, ed. In: Safety requirements for
contraceptive steroids. Cambridge, UK: Cambridge University Press; 193–210
17. Adlercreutz H, Gorbach SL, Goldin BR, Woods MN, Dwyer JT, Hamalainen
E 1994 Estrogen metabolism and excretion in Oriental and Caucasian women.
J Natl Cancer Inst 86:1076 –1982
18. Ursin G, Wilson M, Henderson BE, Kolonel LN, Monroe K, Lee HP, Seow
A, Yu MC, Stanczyk FZ, Gentzschein E 2001 Do urinary metabolites reflect
differences in the breast cancer risk between Singapore Chinese and United
States African-American and white women? Cancer Res 61:3326 –3329
19. Murphy AA, Kettel LM, Morales AJ, Roberts V, Parmley T, Yen SS 1995
Endometrial effects of long-term low-dose administration of RU486. Fertil
Steril 63:761–766
20. Neulen J, Williams RF, Breckwoldt M, Chwalisz K, Baulieu EE, Hodgen GD
1995 Non-competitive anti-estrogenic actions of progesterone antagonists in
primate endometrium; enhancement of estrogen and progesterone receptors
with blockade with post-receptor proliferative mechanisms. Hum Reprod
11:1533–1537
21. Chwalisz K, Brenner RM, Fuhrmann UU, Hess-Stumpp H, Elger W 2000
Antiproliferative effects of progesterone antagonists and progesterone recep-
tor modulators on the endometrium. Steroids 65:741–751
22. Cameron ST, Critchley HO, Thong KJ, Buckley CH, Williams AR, Baird DT
1996 Effects of daily low dose mifepristone on endometrial maturation and
proliferation. Hum Reprod 11:2518 –2526
23. Den Tonkelaar I, Oddens BJ 1995 Preferred frequency and characteristics of
menstrual bleeding in relation to reproductive status, oral contraceptive use
and hormone replacement therapy use. Contraception 59:357–362