First morning voided urine samples were collected by each woman daily
for an entire menstrual cycle or to 50 days. Levels of estrone conjugates
(E1c), luteinizing hormone (LH), follicle-stimulating hormone (FSH), and creatinine
were measured in each sample. Hormone levels in each sample were normalized
to creatinine values. Levels are presented as mean (SEM) plotted against cycle
day, where day 0 is the day of the LH or estrogen peak for women with an estrogen
increase. Mean (SEM) levels are plotted against day of collection for women
with no estrogen increase.
Comparison of E1c levels (estrone conjugates) in women who had an LH
surge (group 1) (left panel) vs those who did not (group 2) (right panel).
E1c levels (mean [SEM]) for women with both estrogen increases and LH surges
are shown here, where day 0 is the day of maximum E1c.
Weiss G, Skurnick JH, Goldsmith LT, Santoro NF, Park SJ. Menopause and Hypothalamic-Pituitary Sensitivity to Estrogen. JAMA. 2004;292(24):2991-2996. doi:10.1001/jama.292.24.2991
Author Affiliations: Department of Obstetrics,
Gynecology and Women’s Health, New Jersey Medical School of UMDNJ, Newark,
NJ (Drs Weiss, Skurnick, Goldsmith, and Park) and Department of Obstetrics,
Gynecology and Women’s Health, Albert Einstein College of Medicine,
New York, NY (Dr Santoro).
Context The onset of human menopause is thought to be caused solely by ovarian
failure and oocyte depletion. However, clinical symptoms and certain recent
data in perimenopausal women suggest central nervous system involvement.
Objective To determine if modifications of hypothalamic-pituitary response to
estrogen feedback mechanisms occur in older reproductive-age women as a mechanism
of the onset of menopause.
Design, Setting, and Participants The Study of Women’s Health Across the Nation (SWAN) is a multiethnic
observational cohort study of the menopausal transition in 3302 women at 7
US sites. Of the subcohort of 840 women who participated in the Daily Hormone
Study between 1997 and 1999, 680 women had evidence of luteal activity. The
remaining 160 women (19%) did not have luteal activity and are the subject
of this report.
Main Outcome Measures Daily urinary hormone levels of estrogen and progesterone metabolites,
luteinizing hormone (LH), and follicle-stimulating hormone (FSH).
Results Three groups of women were studied: those with estrogen increases with
an LH surge, those with estrogen increases without an LH surge, and those
with neither. There were no differences in age or ethnicity among the 3 groups
of women. Women in the third group (no increases) experienced more menopausal
symptoms (hot flashes) than did women in the other groups with estrogen increases.
In older reproductive-age women, the frequent existence of anovulatory cycles
with estrogen peaks, equivalent to those that result in LH surges in younger
women, yet in which no LH surges occur, indicates failure of estrogen-positive
feedback on LH secretion. In other anovulatory cycles, follicular-phase estrogen
levels did not lower LH secretion as occurs in cycles of younger women, indicating
decreased estrogen-negative feedback on LH secretion.
Conclusion Our findings are compatible with hypothalamic-pituitary insensitivity
to estrogen in aging perimenopausal women.
By the year 2030, more than 1.2 billion women in the world will be at
least 50 years old.1 This increasing proportion
of the female population will be experiencing the menopausal transition with
its accompanying physiology and pathophysiology. Reproductive aging in women
and the hormonal changes that occur during the onset of human menopause have
been ascribed solely to ovarian failure and oocyte depletion.2 However,
in other species, the central nervous system is the major regulator of age-related
There are 4 events involving the hypothalamic-pituitary-ovarian axis
that control the human menstrual cycle: (1) The secretion of follicle-stimulating
hormone (FSH), responsible for the development of ovarian follicles and production
of estradiol.4 Throughout the cycle, estrogen
maintains low gonadotropin levels via its negative feedback effect on hypothalamic
gonadotropin-releasing hormone and consequently luteinizing hormone (LH) and
FSH secretion.5 (2) The FSH-induced increase
in ovarian estrogen secretion to levels of sufficient strength and duration
triggering an LH surge (positive feedback).6 (3)
The LH surge, a hypothalamic-pituitary response to the estrogen stimulus.
This positive feedback response of estrogen on LH secretion has been used
as a test of hypothalamic-pituitary function.7,8 (4)
Ovulation and luteinization of the follicle, triggered by the LH surge, forming
a corpus luteum. This is an ovarian response that results in progesterone
secretion necessary for the establishment of a pregnancy.9
How these events may be altered during menopausal transition has not
been well established.
The Study of Women’s Health Across the Nation (SWAN) is a multiethnic
observational cohort study of the menopausal transition in 3302 women at 7
sites across the United States designed to enhance understanding of the factors
that influence the health of women of diverse race and ethnicity.10 The details of enrollment have been previously reported.10 Race/ethnicity was self-determined by study participants
and was obtained by asking the following open-ended question: “How would
you describe your primary racial or ethnic group?” The responses then
were categorized as Caucasian (white), African American, Chinese, Japanese,
or Hispanic. This study was approved by all of the sites’ institutional
review board, and written informed consent was obtained from each participant.
A subcohort participated in the Daily Hormone Study (DHS) from 1997
to 1999. The women in the DHS included 257 Caucasian (white) women, 175 African
American women, 152 Chinese women, 170 Japanese women, and 86 women of Hispanic
origin. The cohort has been described previously.11,12 Inclusion
criteria were age 42 to 52 years; an intact uterus and at least one ovary;
at least one menstrual period in the prior 3 months; no use of sex-steroid
hormones in the previous 3 months; and not being pregnant. DHS enrollees completed
a daily collection of morning voided urine for an entire menstrual cycle ending
in bleeding or to 50 days, whichever came first. During the cycle that they
collected daily urine specimens, DHS enrollees also completed a daily diary,
a questionnaire in which they answered, once a day, whether they had experienced
within the preceding 24 hours any trouble sleeping and any hot flashes or
Urinary LH, FSH, the estradiol urinary metabolites estrone conjugates
(E1c), and the progesterone urinary metabolite pregnanediol glucuronide were
measured using chemiluminescent assays as described previously.11,12 Concentrations
were normalized for creatinine excretion. Previous studies have demonstrated
that urinary levels of these hormones, collected and measured by the methods
used herein, mirror serum hormone patterns during the menstrual cycle in eumenorrheic
controls so closely that patterns of serum and urinary gonadotropins and sex
steroids are superimposible.13
Of the 840 women who completed the DHS study, 680 women had evidence
of luteal activity based on a validated algorithm for pregnanediol glucuronide.12 The algorithm locates the 5 nadir days of pregnanediol
glucuronide in the follicular phase using moving averages throughout the cycle.
A 3-fold increase in pregnanediol glucuronide concentrations above this nadir
for at least 3 consecutive days was considered evidence of luteal activity.
The cycles of these women have been reported previously.12 The
remaining 160 women (19% of the total 840 women) did not have luteal activity.
One woman could not be subclassified due to missing data points. The 159 remaining
women are the subjects of this report.
All women in the SWAN DHS study had levels of FSH equal to or greater
than those previously demonstrated in younger women throughout the entire
cycle.12 This has been described previously
and is due to the decreased secretion of ovarian inhibin in older women.14,15 Decreased gonadotropin secretion,
as is found in some anovulatory cycles in premenopausal young women in their
teens to 30s, clearly did not occur in these women.
Conceptually, an estrogen increase is a high level compared with baseline,
in absolute terms and relative to observed variability, followed by substantial
decline. The specific criteria used here were adapted from previously established
definitions for mid-reproductive age women.13 An
estrogen increase was defined as an E1c level of at least (1) 50 pg/mg creatinine,
(2) twice the baseline level (the mean of 5 consecutive days starting 9 days
earlier), and (3) 3 standard deviations of the baseline levels above the baseline
mean. In addition, the estrogen peak was required to culminate in a drop to
no more than 1.5 times baseline at some time within the next 5 days. An LH
surge is considered present when a high level is observed relative to baseline
in absolute terms and in excess of day to day variability both before and
after the peak, established by a drop in levels. An LH surge was defined as
an LH level of at least (1) 6 mIU/mg creatinine, (2) 3 times the mean baseline
level (of 4 consecutive days starting 5 days earlier), (3) baseline mean plus
3 standard deviations of the 4-day baseline levels, (4) baseline
mean plus 2 standard deviations of levels on days 2 through 6 after the peak,
and (5) 0.8 times the maximum level in that cycle. In addition, the LH surge
was required to culminate in a drop to no more than 1.5 times baseline within
6 days following the peak.
Cycles were classified by these algorithms as falling into 1 of 3 distinct
patterns: (1) both estrogen increase and LH surge (coincident within 2 days),
(2) estrogen increase only, and (3) neither. Visual inspection of the cycle
data plots by 2 observers (G.W., J.H.S.) revealed that 20 cycles had apparent
estrogen increases that were missed by the defining algorithm due to an increase
too early in a short cycle to establish a baseline or a slow decline. Four
cycles with an algorithmically determined estrogen increase were reclassified
to “neither.” Thus, of 159 cycles, 29 were classified as “both,”
32 as “estrogen increase only,” and 98 as “neither.”
These cycle classifications were based solely on hormone levels and not age
or menopausal symptoms.
Analysis of variance was conducted to compare cycle classification groups
on women’s ages and body mass index; χ2 tests were conducted
to compare groups on ethnicity, reason for ending collection, and experience
of symptoms during the cycle. Rank-sum tests were conducted to compare E1c
levels of groups 1 and 2 by cycle day and to compare the 3 groups on women’s
percentage of cycle days with symptom occurrence. Reported P values are 2-sided, without adjustment for multiple comparisons.
Statistical analyses were performed with SAS version 8.2 (SAS Institute, Cary,
NC). Statistical significance was set at 2-sided P<.05.
Women in the 3 classification categories presented here were compared
by ethnicity, age, and body mass index. As shown in Table 1, there was no significant correlation between category of
cycle and any of these characteristics.
The results from the 29 individuals who had an estrogen increase followed
by an LH surge (group 1, “both”) are shown in Figure 1, left panels, in which hormone levels are synchronized
to the LH peak. As seen in ovulatory cycles, these LH surges are accompanied
by FSH surges. Hormone levels are similar to those of the previously reported
women in DHS who had luteal phases,12 indicating
an adequate hypothalamic-pituitary response. In women in group 1, the follicle
or follicles that secreted sufficient estrogen to elicit an LH and FSH surge
did not luteinize as documented by lack of an increase in pregnanediol glucuronide
levels. This is a defect at the ovarian level because hypothalamic-pituitary
responses were similar to those of women with apparently normal cycles.
The results from the 32 women who had clear estrogen increases but no
LH surges (group 2) are shown in Figure 1,
middle panel. Hormone levels are synchronized to the E1c peak because no LH
surges were seen in these women. The estrogen increases in these women were
equivalent to those seen in ovulatory women and to those in group 1. However,
in contrast to those of group 1 women, these estrogen increases did not produce
an LH surge. Comparisons of the E1c levels at each cycle day in group 1 women
with those in group 2 women were performed using rank-sum tests. At all cycle
days, E1c levels in group 2 women who had no LH surge were not lower than
E1c levels in group 1 women who had LH surges. In fact, no differences at
any cycle day were observed, with the exception of day –15, when E1c
levels in group 2 women were higher (P = .04)
than those in women in group 1.
That the secretion pattern of E1c in group 2 women was equivalent to
that of women in group 1 is not apparent in Figure
1, in which hormone levels are synchronized to the LH surge as is
convention. Therefore, the E1c levels synchronized to the E1c peak are presented
in Figure 2. For ease of illustration
of this similarity (between E1c levels in the women with gonadotropin surges
and those in women without surges), the corresponding data for group 2 from Figure 1 are also provided in Figure 2. These E1c patterns in women in groups 1 and 2 are also
similar to E1c secretion in the previously reported SWAN DHS study in women
with luteal activity.12 Thus, in group 2 women
there is adequate ovarian response, but the LH surge, a hypothalamic-pituitary
phenomenon, did not occur in the presence of an estrogen stimulus that is
adequate to elicit an LH surge in ovulating women and in younger women. This
clearly demonstrates unresponsiveness of the hypothalamic-pituitary axis to
an estrogen peak. Gonadotropin levels dropped in the latter part of the cycles,
likely due to negative feedback from the estrogen increase.
Hormone secretion in the 98 women who had no estrogen peaks or LH surges
(group 3) are shown in Figure 1, right
panels. Levels of LH are higher than those seen in either SWAN perimenopausal
women with luteal phases or in the other 2 groups presented here. Estrogen
levels are comparable to those in the early follicular phase of DHS luteal
women12 and group 1 or 2 women. Thus, group
3 women still have ovarian function but are unable to produce an estrogen
Twenty-eight of the 29 women with LH surges (group 1), 31 of the 32
women without LH surges (group 2), and 97 of the 98 women with neither estrogen
nor LH increases (group 3) participated in the daily diary component of the
study. For each symptom— trouble sleeping or hot flashes or night sweats—the
percentage of days a woman reported that she experienced the symptom was computed
as the total number of days she reported presence of the symptom divided by
the total number of days she reported either yes or no for that symptom ×
100. These percentages of days with positive reports for women in the 3 categories
were compared by rank-sum tests.
There were no differences in the percentages of days with trouble sleeping
among the 3 groups (Table 2). Comparison
of the percentages of days with hot flashes or night sweats among the 3 groups
revealed significant group differences (Table
2). The percentages of days with hot flashes or night sweats were
significantly higher for group 3 women than for either group 1 women (P = .01) or group 2 women (P = .02). The percentages of days with hot flashes or night
sweats in group 1 women and group 2 women did not differ (P = .73).
Hormone secretion patterns in older reproductive-age women demonstrate
significant alterations of hypothalamic-pituitary feedback mechanisms in addition
to decreased ovarian function. Cycles exist in which failure to mount an LH
surge occurs in the face of adequate estrogen stimulation. These findings
support the hypothesis that there is a relative hypothalamic-pituitary insensitivity
to estrogen in aging women that is manifested by both positive and negative
feedback mechanisms. Estrogen levels and patterns that produce LH surges in
younger women fail to do so in some older women. In addition, levels of estrogen
similar to those in younger women, which cause negative feedback of LH in
normal ovulatory women and in group 1 women, fail to do so in group 3 women,
who have elevated LH in the presence of early follicular-phase levels of E1c.
This situation may represent a later stage of the menopausal transition because
there is opening of the negative feedback loop between ovarian estrogen and
pituitary LH, as is seen in postmenopausal women. Levels of estrogen capable
of lowering LH in cycling women were not able to cause negative feedback of
estrogen on LH secretion. Because control of FSH secretion is more complex
than LH and includes major influences by inhibins and activins, FSH is not
a good marker for estrogen-negative feedback control of gonadotropin secretion.
Decreased LH pulse frequency has been observed in the presence of normal,
midreproductive sex steroid levels in both the follicular16 and
luteal17 phases of the menstrual cycles of
older premenopausal women, supporting this hypothesis as well.
A predominant hypothesis to explain the onset of puberty is the occurrence
of a gradual decrease in estrogen sensitivity of the hypothalamic-pituitary
axis, such that small levels of circulating estrogen, which suppress gonadotropin
secretion prepuberty, are unable to do so in the pubertal transition.18 Decreased sensitivity in later life may simply be
a continuation of the same pattern of progressive age-related estrogen insensitivity.
Levels of LH are higher in perimenopausal women than in younger women, even
in the presence of estrogen concentrations that result in lower LH levels
in younger women.14 Symptoms such as hot flashes
and sleep disturbances occur more commonly in perimenopausal women than in
postmenopausal women.19 Yet, the perimenopausal
transition is a time when circulating estrogen levels are equivalent or higher
than levels observed in younger women.14 Additionally,
exogenous estrogen is therapeutic in perimenopausal women.20 These
observations are consistent with the hypothesis that a decrease in estrogen
sensitivity occurs as women age through the menopausal transition.
We found no differences in reported symptoms of sleep disturbances or
vasomotor changes in women with different LH-positive feedback responses to
equivalent circulating estrogen peaks (categories 1 and 2). Although group
3 women did not differ in the prevalence of sleep disturbances, they had a
significantly higher prevalence of hot flashes or night sweats. While all
women in this study had similar baseline estrogen levels, the group 3 women
did not have mid-cycle estrogen peaks. Since younger women with estrogen levels
similar to these women12 do not get hot flashes,
the flashes may be due to opening of the negative feedback loop of estrogen
on gonadotropin secretion.
All women have hormonal changes during their menopausal transition,
but not all women experience symptoms. It is likely that the gonadotropin
control centers in the brain differ from the areas involved in the symptoms
assessed in our study. These regions may have different steroid sensitivities
and control mechanisms, accounting for our findings. However, certain changes
in gonadotropin levels may be permissive to alterations in central nervous
system function, which result in symptoms. Physicians should be aware of the
central nervous system changes involved in the menopausal transition because
these changes best explain their patients’ symptoms. An appreciation
of these changes may assist patients in understanding and dealing with their
own menopause. Other symptoms, such as mood and changes in affect, may also
have similar explanations.
Demonstration of hypothalamic-pituitary insensitivity to estrogen in
perimenopausal women shows that certain human menopausal responses are similar
to those of other species, including rats. Age-related decreased expression
of estrogen receptor β in some hypothalamic areas in the rat has been
described.21 These results suggest that the
rat may be a useful model for the study of human central nervous system aging
and that the mechanisms of reproductive aging may be more similar in diverse
mammalian species than previously thought.
Corresponding Author: Laura T. Goldsmith,
PhD, Department of Obstetrics, Gynecology and Women’s Health, New Jersey
Medical School of UMDNJ, 185 S Orange Ave, Newark, NJ 07103 (firstname.lastname@example.org).
Author Contributions: Dr Weiss had full access
to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analysis.
Study concept and design: Weiss, Skurnick,
Goldsmith, Santoro, Park.
Acquisition of data: Weiss, Santoro.
Analysis and interpretation of data: Weiss,
Skurnick, Goldsmith, Santoro, Park.
Drafting of the manuscript: Weiss, Skurnick,
Critical revision of the manuscript for important
intellectual content: Weiss, Skurnick, Goldsmith, Santoro, Park.
Statistical analysis: Skurnick.
Obtained funding: Weiss, Skurnick, Santoro.
Administrative, technical, or material support:
Weiss, Goldsmith, Santoro, Park.
Study supervision: Weiss, Santoro.
SWAN Clinical Centers: University of Michigan,
Ann Arbor: MaryFran Sowers, PI (U01 NR04061); Massachusetts General Hospital,
Boston: Robert Neer, PI 1995-1999, Joel Finkelstein, PI 1999- present (U01
AG012531); Rush University, Rush-Presbyterian-St Luke’s Medical Center,
Chicago, Ill: Lynda Powell, PI (U01 AG012505); University of California, Davis/Kaiser:
Ellen Gold, PI (U01 AG012554); University of California, Los Angeles: Gail
Greendale, PI (U01 AG012539); University of Medicine and Dentistry/New Jersey
Medical School, Newark: Gerson Weiss, PI 1995-2004, Nanette Santoro, PI, 2004-present
(U01 AG012535); and the University of Pittsburgh, Pittsburgh, Pa: Karen Matthews,
PI (U01 AG012546).
NIH Program Office: National Institute on Aging,
Bethesda, Md: Sherry Sherman, 1994-present, Marcia Ory, 1994-2001; National
Institute of Nursing Research, Bethesda, Md: Janice Phillips, 2002-present,
Carole Hudgings, 1997-2002.
Central Laboratory: University of Michigan,
Ann Arbor: Rees Midgley, PI 1995-2000, Daniel McConnell, 2000-present (U01
AG012495, Central Ligand Assay Satellite Services).
Coordinating Centers: University of Pittsburgh,
Pittsburgh, Pa: Kim Sutton-Tyrrell, PI (U01 AG012546) 2001-present and New
England Research Institutes, Watertown, Mass: Sonja McKinlay, PI (U01 AG012553),
Steering Committee: Chris Gallagher, Chair,
1995-1997; Jenny Kelsey, Chair, 1997-2002; Susan Johnson, Chair, 2002- present.
Funding/Support: The Study of Women’s
Health Across the Nation (SWAN) was funded by the National Institute on Aging.
Role of the Sponsor: The National Institute
on Aging funded the SWAN study. The current study used data generated by the
SWAN Daily Hormone Study (DHS). The DHS was designed and approved by the SWAN
Steering Committee. The design and conduct of the current study were a product
of the authors and the samples were collected according to the SWAN DHS protocol
and occurred at all 7 clinical sites. Data management was performed by the
SWAN Coordinating Center. Analysis and interpretation of the data were accomplished
by the authors. Preparation of the manuscript was performed by the authors.
The SWAN Presentation and Publication Committee reviewed and approved the
concept, study plan, and manuscript draft for consistency with the SWAN study
and appropriateness of the current work in accordance with SWAN policy.
Acknowledgment: We thank the study staff at
each site and all the women who participated in SWAN.