Evaluation of Four Proposed Bleeding Criteria for the Onset of Late Menopausal Transition

Abstract

Context:

The current criterion for onset of late menopausal transition is amenorrhea of 90 d or more. The Stages of Reproductive Aging Workshop proposed alternative criteria based on a shorter period of amenorrhea. Empirical data comparing proposed criteria are not available.

Objective:

This paper evaluates the several bleeding criteria that served as the basis of these recommendations. The goal was to provide empirically based guidance regarding which bleeding criterion may be optimal for widespread application in clinical and research settings.

Design/Setting:

The study used prospective menstrual calendar data from four community and population-based cohort studies: TREMIN, Melbourne Women's Midlife Health Project, Seattle Midlife Women's Health Study, and Study of Women's Health Across the Nation.

Participants:

The study included 735 TREMIN, 279 Seattle Midlife Women's Health Study, 216 Melbourne Women's Midlife Health Project, and 2270 Study of Women's Health Across the Nation women aged 35–57 yr at baseline who contributed 10 menstrual cycles or more.

Main Outcome Measure(s):

The main measures were the frequency of and median age at occurrence and time from occurrence to final menstrual period (FMP) for four criteria: skipped segment, 10-segment running range, 60- and 90-d amenorrhea.

Results:

A skipped segment, 10-segment running range greater than 42 d and 60-d amenorrhea identify a similar time in women's reproductive lives. The latter two identify the exact same date in two thirds of women. All three criteria occur in a greater proportion of women than the 90-d criterion and are equally predictive of the FMP, although they occur 1–2 yr earlier.

Conclusions:

These findings support the recommendation of the Stages of Reproductive Aging Workshop that 60 d of amenorrhea be used to define onset of the late menopausal transition.

Althought the menopause marks a period of critical change in women's physiology and reproductive status (1), a staging system for reproductive aging has not been definitively established. In clinical practice, a staging system would provide a basis for assessment of the transition from active reproduction to postmenopause, enabling women and clinicians to better predict the timing of the transition experience. Standardized definitions would improve comparability of studies and facilitate efforts to distinguish effects of ovarian vs. chronological aging on health.

In 2001 the Stages of Reproductive Aging Workshop (STRAW) proposed that reproductive life could be characterized by seven stages. Before menopause, reproductive life is divided into the reproductive years (three stages) and the transition years (two stages, early and late transition). Post-menopause (two stages) follows the final menstrual period (FMP). An initial consensus was reached regarding menstrual and hormonal changes to be used as criteria for, or markers of, each stage (2). Entry into early transition is characterized by increasing levels of FSH and increasing variability in menstrual cycle length. Entry into the late transition is characterized by continued elevation of FSH and the occurrence of skipped cycles or amenorrhea.

STRAW's recommended bleeding criterion for onset of the late menopausal transition (more than two skipped cycles and 60 d of amenorrhea) was based on results from three studies (3-6). Taffe and Dennerstein (5), using data from the Melbourne Women's Midlife Health Project (MWMHP), proposed the running range (the number of days difference between the longest and shortest cycle over a defined period); once it reached 42 d, fewer than 20 cycles remained until the FMP. Mitchell et al. (3), using data from the Seattle Midlife Women's Health Study (SMWHS), proposed the skipped cycle. Analyses of TREMIN data (6) suggest amenorrhea of 60 d. Notably, each study suggested a shorter interval of amenorrhea than the current definition, amenorrhea of 3 months or 90 d (7, 8). STRAW's recommendations were based on a considered, but not empirical, assessment of the likely concordance within women of these three measures. An underlying assumption was that these newer criteria would be equally predictive of the FMP and less likely to misclassify women who do not experience extended amenorrhea before menopause.

Investigators from four cohort studies formed the ReSTAGE collaboration to empirically validate STRAW's recommendations. Of key interest was whether the various criteria define a similar time in women's reproductive lives and whether criteria developed within one cohort were replicable in others. We assess the empirical evidence that proposed bleeding criteria are valid and robust markers of onset of late menopausal transition across diverse populations and ages at menopause to evaluate whether this evidence supports adoption of one proposed marker over others. We address these questions using menstrual calendar data from four cohort studies of the menopausal transition: TREMIN (9); MWMHP (10); SMWHS (3); and the multisite, multiethnic Study of Women's Health Across the Nation (SWAN) (8). Our aim was to provide empirically based guidance regarding which bleeding criteria may be optimal for widespread application in clinical and research settings.

Subjects and Methods

Experimental subjects

We conducted secondary analyses using menstrual calendar information accumulated in four longitudinal studies. This analysis has been approved by Institutional Review Boards at the University of Michigan, University of Washington, University of Massachusetts, and University of Melbourne.

TREMIN (9) has prospectively recorded menstrual calendars for women across the entire reproductive life span. Numerous publications from this study, initiated by Treloar and colleagues (11, 12), define our understanding of the menopausal transition. From 1935 to 1939, 1997 women students at the University of Minnesota were enrolled. This analysis includes records from 735 women who were still participating and not using hormones at age 35 yr, the baseline age for these analyses. (Data tape TRUST998.FINAL were supplied by TREMIN in March 1993.)

MWMHP began in 1991 with a cross-sectional survey of 2001 Australian-born women of Anglo-European heritage identified by random telephone digital dialing (10). From these women, MWMHP enrolled 438 women aged 45–55 yr who had menstruated in the prior 3 months and were not using hormones into a longitudinal study (13). This analysis includes records from 216 participants who maintained calendars up to 8 yr (14).

SMWHS is an ongoing longitudinal study begun in 1990 (3). The original population-based sample of 508 was obtained by telephone screening of all households in specified census tracts to identify English-speaking women aged 35–55 yr, with at least one menstrual period in the previous 12 months, at least one intact ovary, and not pregnant/lactating. The sample includes women using hormone therapy at enrollment and was 76% Caucasian, 11% Asian-American/Pacific Islander, 8% African-American, and 5% other. This analysis includes records of 279 women contributing up to 13 yr of menstrual calendars.

SWAN is an ongoing multiethnic, multisite, longitudinal, community-based study of middle-aged women begun in 1995 (8). A cross-sectional survey randomly selected women from lists including a large managed health care plan, community census, utility households, and registered voters or by random digit dialing. For the longitudinal phase, each site recruited about 450 women including Caucasian women and women from one minority group (African-Americans at four sites; and Japanese, Chinese, and Hispanic women at one site each) who were aged 42–52 yr with an intact uterus, not using hormones, and had a menstrual period in the previous 3 months. The sample of 3302 is 28% African American, 47% Caucasian, 8% Chinese, 9% Hispanic, and 8% Japanese. This analysis includes records of 2270 women contributing up to 6 yr of menstrual calendars.

To be eligible for ReSTAGE, women had to provide at least 10 consecutive, nonmissing, untreated bleeding segments. (A bleeding segment is analogous to a menstrual cycle but acknowledges that menstrual and nonmenstrual bleeds are not easily distinguished in menstrual calendar data.)

Methods

Each cohort in this analysis included a menstrual calendar and asked women to record each day they had menstrual bleeding/spotting on the day it occurred. Although slightly different in format, calendars identified the date of bleeding onset and days of bleeding in a comparable manner. Date of onset of late transition and FMP are ascertained from these calendar data.

In TREMIN, the menstrual calendar card covered 1 yr (9). At the end of each year, women were asked whether they would continue participating and completed a short questionnaire on marital status, medical treatments for menstrual difficulties, and pregnancies. Questions about menopause were added in 1952 and oral contraceptive use in 1963.

In MWMHP, the menstrual calendar card covered 1 yr (14). Participants had annual assessments, which included symptom experience, measured height and weight, hormone use, experience of hot flashes (15), and blood sampling to measure FSH and estradiol. In SMWHS (3), the menstrual calendar card covers 1 yr. Height and weight were measured in 1997 and 2000. A yearly health questionnaire includes information on self-reported weight and hormone use. Data on hot flashes were collected through symptom diaries. First morning urine specimens are collected quarterly on d 5–7 and analyzed for FSH.

In SWAN (8) women are provided monthly menstrual calendars. Assessments at baseline and each annual follow-up visit include symptom experience, measured height and weight, hormone use, and blood sampling to measure FSH and estradiol.

Defining menstrual segments

Using definitions recommended by the World Health Organization (16) and modified by ReSTAGE, a bleeding episode is a period of consecutive bleeding days; a bleeding-free interval is a period of consecutive bleeding-free days; and a bleeding segment is a bleeding episode and the subsequent bleeding-free interval. A single day of bleeding as well as consecutive days of spotting/bleeding were coded as bleeding episodes. Bleeding-free intervals had to consist of at least 3 d; thus, the shortest bleeding segment was 4 d. One or 2 bleed-free days between 2 bleed days were considered part of the bleeding episode. SMWHS judged some bleeding episodes to be midcycle bleeding based on careful observation of women's patterns and ignores these episodes in calculating segment length. Pregnancies and the first three segments after birth/abortion are considered nonmenstrual intervals and are excluded from analyses (17).

For each bleeding episode and subsequent bleeding-free interval, we indicated whether a woman was using hormone therapy or oral contraceptives, fertility medications, selective estrogen receptor modulators, injectable contraceptives, contraceptive implants, or a short course of hormones to treat a menstrual disorder. Gaps in the menstrual record were coded as missing. All untreated bleeding segments of each eligible woman are included in these analyses, with women contributing up to 321 eligible bleeding segments (Table 1). From 41 to 51% of the women had intermittent missing data with a median of two gaps per woman.

TABLE 1.

Summary of final status of participants and characteristics of menstrual calendar data by cohort

	TREMIN (n = 735)		MWMHP (n = 216)		SMWHS (n = 279)		SWANa (n = 2270)
	n	%	n	%	n	%	n	%
Final status
FMP observed	211	28.7	42	19.4	52	18.6	132	5.8
Hysterectomy	68	9.3	11	5.1	9	3.2	17	0.8
Still menstruating	0	0	53	24.5	34	12.2	953	42.0
Stopped calendar/still menstruating	119	16.2	48	22.2	86	30.8	187	8.2
Endometrial ablation	0	0	3	1.4	0	0	0	0
Began hormone therapy	334	45.4	59	27.3	94	33.7	981	43.2
Began radiation/chemotherapy	3	0.4	0	0	4	1.4	0	0
No. of bleeding segments
Median	145		36		69		35
Maximum	321		107		190		91
Intermittent missing
Proportion with any missing data	301	41.0	90	41.7	141	50.5	1159	51.1
Median number of gaps	2		2		2		2

^aAnalysis to date of the SWAN calendar is based on the first 6 yr of menstrual calendar data.

Defining bleeding criteria for onset of the late menopausal transition

This paper compares three proposed bleeding criteria (3-6) for onset of late menopausal transition with the criterion of 90-d amenorrhea (7). We specify the bleeding episode that marks when each bleeding criterion is met (hereafter the marker event). For all criteria, except the running range, we identify the bleeding episode that marks the first observed occurrence of the criterion as well as the episode that marks its persistent occurrence. We define persistence when a marker, once observed, occurs again within 10 segments. For the running range, we identify the bleeding episode that marks the time when the running range exceeds 42 d.

We calculated the following marker events for late transition: the first and persistently observed 1) segment of at least 90 d (3); 2) segment of at least 60 d (6); 3) skipped segment, defined as a segment that exceeded twice the median segment length of the previous 10 segments (3); and 4) a running range more than 42 d cumulative since enrollment and over a 10-segment sequence (4). We adapted the definition of a skipped segment (3) using the median as opposed to the mode and using a moving reference window of 10 segments to facilitate calculation of the marker across studies. Menopause was defined as the date of the FMP recorded in the menstrual calendars established retrospectively after 12 months of amenorrhea.

Analysis

Women were censored when they had a hysterectomy or bilateral oophorectomy or when they began using hormonal birth control, hormone therapy, or chemotherapy (Table 1). Segments during which a short course of hormonal medications, including fertility medications, was used to treat a menstrual disorder were treated as gaps in the menstrual record.

For each woman, we identified the age at occurrence of each marker, defined as the first day of the bleeding episode that meets the criterion. We compared the four bleeding criteria within each study and assessed robustness and replicability of results across studies. We first determined the proportion of women in whom each bleeding criterion was observed before FMP among the subgroup of uncensored women with observed FMP. The age distribution of each marker event was then calculated using Kaplan-Meier survival analysis to include information from censored and uncensored women, and side-by-side box plots were generated. We then calculated within-individual differences in age at markers and assessed whether systematic differences existed in the mean age at occurrence using a paired t test.

To evaluate how well the criterion predicted FMP, we modeled the hazard function of FMP given age at marker and assessed the association between observation of a marker and time to FMP. We used a varying-coefficient Cox model (18) that incorporates censored and uncensored women. Both age at marker and age at menopause events were subject to censoring, and their association varies with age at the marker (6). Estimation proceeded using the regression spline method, with the regression coefficient being a function of age at the marker event, the biological parameter of interest, as opposed to a function of follow-up time (18). For the 60-d marker, we plotted the log-relative hazard of reaching FMP and also plotted the estimated time from marker to menopause by age at marker for both 60- and 90-d markers. To test whether the hazard curves differed across markers, we calculated the difference between any two estimated log-relative hazard curves and used a bootstrap approach to calculate the variance of the difference by age.

Intermittently missing data were a concern because we may fail to observe transitions that occur during gaps in the menstrual record. We developed an approach to impute missing menstrual segments (Little, R., M. Yosef, K. Cain, B. Nan, S. D. Harlow, unpublished data). Because results from imputed data differed little from results from unimputed data (Little, R., M. Yosef, K. Cain, B. Nan, S. D. Harlow, unpublished data), we present only the latter. Gaps less than 2 yr are ignored, and women with gaps longer than 2 yr were censored at the gap.

Results

Mean age of participants at baseline was 41.3 yr in SMWHS, 45.9 yr in SWAN, and 48.7 yr in MWMHP with all participants aged 35 yr in TREMIN. Mean body mass index was 25.3 kg/m² in SMWHS, 27.6 kg/m² in SWAN, and 25.5 kg/m² in MWMHP. Thirteen to 22% reported having hot flashes at enrollment.

Among women observed through FMP, the segment of 60 d or more and the 42-d running range were observed in 90–100% of women (Table 2). The criterion of 90 d or more was least likely to occur, with a maximum of 91% of participants with an observed FMP ever experiencing this criterion. Requiring persistence reduced substantially the proportion of women observed to experience each marker; thus, subsequent analyses considered first occurrence only.

TABLE 2.

Proportion of women in whom marker events for the late menopausal transition are observed among all women and among women with observed FMP by cohort

	All women			Women with FMP observed
	Total	First occurrence		Total	First occurrence		Persistent occurrence
	n	n	%	n	n	%	n	%
At least 60-d segment
TREMIN	735	325	44.2	211	202	95.7	187	88.6
SMWHS	279	112	40.1	52	52	100	50	96.2
MWMHP	216	127	58.8	42	40	95.2	30	71.4
SWAN	2270	923	40.7	132	121	91.7	105	79.5
Skipped segment
TREMIN	735	327	44.5	211	200	94.8	179	84.8
SMWHS	279	105	37.6	52	48	92.3	42	80.8
MWMHP	216	111	51.4	42	32	76.2	26	61.9
SWAN	2270	749	33.0	132	98	74.2	77	58.3
Ten-segment running range > 42 d
TREMIN	735	329	44.8	211	203	96.2	NA	NA
SMWHS	279	108	38.7	52	52	100	NA	NA
MWMHP	216	125	57.9	42	39	92.9	NA	NA
SWAN	2270	976	43.0	132	124	93.9	NA	NA
At least 90-d segment
TREMIN	735	254	34.6	211	192	91.0	129	61.1
SMWHS	279	77	27.6	52	47	90.4	31	59.6
MWMHP	216	87	40.3	42	33	78.6	21	50.0
SWAN	2270	549	24.2	132	107	81.1	65	49.2

Figure 1 presents Kaplan Meier estimates of the distribution of age at marker for each bleeding criteria within each cohort, using information from censored and uncensored women (see also Table 2). Within each study, the median age at occurrence of a skipped segment, 10-segment running range greater than 42 d, and at least 60-d segment are similar and occurred 1–2 yr earlier than the median age of the at least 90-d segment marker. For example, in the SMWHS, median ages for a skipped segment, running range greater than 42 d, and 60-d segment were all 51.1 yr as compared with 52.0 yr for the 90-d segment marker. Similarly in MWMHP, the median ages were 52.0–52.2 vs. 53.1 yr, respectively. Systematic differences in median age at marker across studies are a function of the age at enrollment, with TREMIN being the youngest and MWMHP being the oldest and in SWAN the relatively shorter duration of follow-up.

Fig. 1.

Box plot distributions of Kaplan Meier estimates of age at marker for four bleeding-related markers of the late menopausal transition by cohort (bold line represents 50th percentile, box represents 25th and 75th percentiles, dashed line represents 10th and 90th percentiles).

The 90-d marker occurred on average 0.5–1.5 yr later than the other markers, although for roughly one third of women, the 90-d marker occurred at the exact same age, i.e. on the same bleeding episode, as the other markers (Table 3). The 60-d segment marker and 10-segment running range greater than 42 d occurred on the exact same bleeding episode in 65–74% of women and were within 1 yr of each other in 86–94% of women with no significant difference in the mean age of occurrence in any of the four studies. Mean age of the skipped segment was significantly earlier than the running range or 60-d markers in TREMIN and SWAN with the age difference 0.1–0.8 yr. Yet age at skipped segment and other markers coincided in 49–67% of women.

TABLE 3.

Distribution of difference in age at onset of markers of the late menopausal transition within each cohort study

	TREMIN		SMWHS		SWAN		MWMHP
	Exact same age, %	Difference (yr) mean (se)	Exact same age, %	Difference (yr) mean (se)	Exact same age, %	Difference (yr) mean (se)	Exact same age, %	Difference (yr) mean (se)
Skipped vs. ≥ 60-d segment	53.8	−0.64 (0.16)a	60.0	−0.09 (0.12)	51.6	0.12 (0.04)b	66.7	0.01 (0.10)
Skipped segment vs. running   range > 42 dc	49.2	−0.81 (0.16)a	55.3	−0.21 (0.15)	48.9	0.14 (0.04)d	60.8	0.12 (0.09)
≥60-d segment vs. running   range > 42 dc	67.8	−0.16 (0.12)	73.6	−0.13 (0.09)	68.2	−0.03 (0.02)	64.7	0.05 (0.05)
Skipped vs. ≥ 90-d segment	21.2	−2.39 (0.21)a	30.9	−1.19 (0.22)a	33.9	−0.52 (0.05)a	30.6	−0.32 (0.11)b
≥60 vs. ≥ 90-d segment	30.3	−1.65 (0.19)a	36.4	−0.97 (0.18)a	39.2	−0.69 (0.03)a	36.1	−0.49 (0.07)a
Running range > 42 d vs.   ≥ 90-d segmentc	28.9	−1.52 (0.16)a	36.4	−0.84 (0.13)a	39.2	−0.64 (0.03)a	36.1	−0.55 (0.08)a

^aP < 0.0001.

^bP < 0.01.

^cTen-segment running range.

^dP < 0.001.

Occurrence of a marker before age 40 yr was not informative of the proximity of the FMP because median time to menopause was greater than 10 yr in the two studies with relevant data (TREMIN and SMWHS). After age 40 yr, the median time from marker to menopause was on average 1.5–2.3 yr for the 90-d marker and 2.6–3.3 yr for the other three markers. Time to FMP decreased with increasing age at marker, and the curve was steepest for the 90-d marker.

Figure 2 presents an example of the estimated probability of reaching FMP by time since marker and age at marker for the 60- and 90-d amenorrhea criteria using data from TREMIN. For all bleeding criteria, the hazard of FMP was higher if the marker event had occurred relative to women who had not yet experienced the marker event. Figure 3 presents the log-relative hazard of FMP with 95% confidence intervals for the 60-d segment criterion by age for each study. The log-relative hazard declined with age at marker and was broadly similar across data sets. However, steepness of the decline depended on age at cohort enrollment and appears to be sensitive to left censoring and truncation. At any given age, the log-relative hazard was similar for each of the four markers, although the hazard for skipped segment was lower at some ages (P < 0.05) than the hazard for the other markers in SWAN and TREMIN (data not shown).

Fig. 2.

Estimated probability of reaching FMP by time since bleeding marker event occurred and age at marker for different markers of the late menopausal transition in TREMIN.

Fig. 3.

Log-relative hazard of FMP by age at 60-d marker adjusted for age at enrollment. (To achieve convergence of the model in SWMHS, marker events after age 53 yr are ignored; n = 4.)

Discussion

This paper evaluated several proposed bleeding criteria for defining the onset of the late menopausal transition in four large cohort studies. Three of the four criteria, a skipped segment, a 10-segment running range greater than 42 d, and a segment of at least 60 d, identified a similar time in women's reproductive lives, with the latter two identifying the same date in two thirds of women. These markers occurred in a greater proportion of women than the 90-d marker and were equally predictive of the FMP, although they occurred 1–2 yr earlier. Findings were robust across the four cohorts that represent diverse populations.

Establishing a staging system for reproductive aging akin to the Tanner-Marshall stages of puberty (19) would be of great scientific and clinical interest. The goal is to define objective, reliable criteria that demarcate distinct physiological phases of the reproductive life span, with a particular focus on the period of transition to reproductive senescence (1). STRAW recommended that the primary criteria for defining the stages of the transition should be based on menstrual bleeding. STRAW made a preliminary recommendation as to how to define onset of the late transition, i.e. two or more skipped cycles and an interval of amenorrhea of 60 d or more.

A reliable bleeding criterion for the late transition would be expected to occur, before FMP, in almost all women; predict a relatively short time to FMP; predict a shorter time to FMP for women who have experienced the marker, compared with women who have not; and be associated with FSH levels and symptoms of the transition, such as hot flashes. Taken together our data provide substantial evidence that an interval of amenorrhea of 60 d or more defines the onset of the late menopausal transition. In selecting an optimal marker, consideration should be given to feasibility of measurement by women, clinicians, and researchers, given the similarity in timing and behavior of the other markers.

When we required persistence, i.e. that the marker be observed to occur again within 10 segments, considerably fewer women met the criterion for each marker across studies. Persistence may not be observed if women spend little time in the late transition, which may be more likely in older women. However, persistence might be important at younger ages to distinguish women truly in transition from women experiencing a single aberrant cycle. A future paper will address this question.

Retrospective recall of specific menstrual events is unreliable (20-23). These analyses are based on similarly designed menstrual calendars with prospectively recorded data on menstrual bleeding. However, nonparticipants in the menstrual calendar component of these studies were older, heavier, and more likely to smoke currently. Thus, selection bias cannot be ruled out. Left censoring and left truncation occurs in three of the studies. Left censoring, when a participant has begun the transition before enrollment, can be compared with the situation in which a woman presents to a clinician in her 40s, and the clinician must base her evaluation on subsequent observations. Left truncation occurs when women who began the transition at an earlier age are excluded from the cohort. Both left censoring and left truncation are likely to bias upward estimates of age at marker and age at menopause. The impact of these biases requires further assessment. Intermittent missing data, although of potential concern, did not have a large influence on our results.

Currently proposed staging criteria (1, 3-5) apply only to the case in which women are not using exogenous hormones. Omission of hormone users is problematic both because they represent a significant proportion of women and hormone users are of interest, given that use is more frequent among symptomatic women (24, 25). Future studies should extend criteria for staging reproductive aging to apply to hormone users. Although STRAW explicitly limited their proposal to nonsmoking women with normative body weight, these analyses included smokers and women across the spectrum of body size. Future studies should evaluate whether systematic differences exist in the relationship between timing of the transition and these characteristics (26, 27).

In conclusion, this paper provides information necessary to refine the STRAW recommendations and facilitate development of a consensus regarding which of the extant criteria may be optimal for defining onset of the late transition. Evaluation of the concordance between approaches and their relative ability to predict the approach of the FMP across multiple cohort studies demonstrates that several of the extant criteria are similar and suggests that they robustly define a critical time in reproductive life that is likely to correspond to the onset of the late transition. Notably, the newer proposed markers each describe a shorter duration of amenorrhea than the 90-d criterion in current use.

Our findings are consistent with similar comparisons of questionnaire-based criteria (28) and support the STRAW recommendation that a shorter duration of amenorrhea, i.e. 60 d as opposed to 90 d, be used as the bleeding criterion for the late transition. Given the concordance in timing of the running range, skipped segment, and 60-d amenorrhea criteria, selection of the most appropriate criterion should incorporate logistical and biological considerations. The 60-d criterion is most easily calculated by women, clinicians, and researchers. A separate paper examining the association between these bleeding criteria and changes in serum FSH levels suggests that the 60-d criterion also correlates well with underlying hormonal and other changes that characterize the transition (29). Taken together these findings support adoption of 60 d of amenorrhea as the bleeding criterion for the onset of the late menopausal transition.

Abbreviations

FMP

Final menstrual period

MWMHP

Melbourne Women's Midlife Health Project

SMWHS

Seattle Midlife Women's Health Study

STRAW

Stages of Reproductive Aging Workshop

SWAN

Study of Women's Health Across the Nation