Best Practice & Research Clinical Obstetrics and Gynaecology

Transcription

1 Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) 7 23 Contents lists available at ScienceDirect Best Practice & Research Clinical Obstetrics and Gynaecology journal homepage: 1 Hormonal changes and biomarkers in late reproductive age, menopausal transition and menopause G.E. Hale, MD, PhD, Clinical Associated Lecture a, *, H.G. Burger, MD, Director b a Department of Obstetrics and Gynaecology, Queen Elizabeth II Research Institute for Mothers and Infants, University of Sydney, NSW, Australia b Department of Endocrinology, Prince Henry s Institute, Clayton, Victoria, Australia Keywords: menopause perimenopause menopausal transition climacteric reproductive aging inhibin anti-mullerian hormone This chapter describes current definitions of the climacteric, perimenopause, menopausal transition and menopause, and discusses the 1 Stages of Reproductive Aging (STRAW) criteria in relation to more recently proposed categorization criteria for reproductive aging. Data from endocrine studies on women throughout the menopausal transition are discussed from earliest to most recent. The earlier studies focused on the changes in levels of steroid hormones and gonadotrophins, and established that follicle-stimulating hormone undergoes the first detectable change while menstrual cycles remain regular. Erratic and less predictable changes in steroid hormones follow, especially with the onset of irregular cycles. Later serum hormone studies on the inhibins and anti-mullerian hormone established that diminishing ovarian follicle number contributes to the endocrine changes with advancing reproductive age. A classification system of cycle types incorporating all available endocrine data and their associated menstrual cycle patterns is proposed, and the application of biological markers as diagnostic tools for reproductive staging is discussed. Ó 8 Published by Elsevier Ltd. Definitions A number of terms including climacteric, perimenopause, menopausal transition, postmenopause and menopause have been used to refer to the stages of reproductive aging surrounding the final menstrual period (FMP). The terms menopausal transition and perimenopause were * Corresponding author. Tel.: þ ; Fax: þ address: ghale@med.usyd.edu.au (G.E. Hale) /$ see front matter Ó 8 Published by Elsevier Ltd. doi:1.116/j.bpobgyn.8.1.1

2 8 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) 7 23 recommended for use in place of the term climacteric in 1996, with release of the definition by the World Health Organization (WHO). Although these two terms were initially used interchangeably, there is a slight difference in their definition according to the STRAW (Stages of Reproductive Aging) criteria, and the term menopausal transition is now used in favour of both perimenopause and climacteric when referring to the stages of reproductive age in a scientific context. 1 The menopausal transition or perimenopause was first officially referred to by WHO in 1996 as that period of time immediately before menopause when the endocrinological, biological and clinical features of approaching menopause commence. 2 Given that the endocrinological, biological and clinical features referred to in the WHO definition remained undefined, many investigators proposed their own classification criteria to categorize study subjects for research purposes. 3 5 In the prospective Massachusetts Women s Health Study (MWHS), a self-report of between 3 and 12 months of amenorrhoea was used for entry into the perimenopause. 3 Similar criteria were used for entry into the late perimenopause in the Seattle Midlife Women s Health Project. 4 A self-report of a change in menstrual flow, duration of menstrual flow or cycle length was used for entry into the early perimenopause, and new-onset variability in cycle length (two or more consecutive cycles that differ in length by at least 7 days) was used as entry into the mid-perimenopause. 4 In a major longitudinal study of the experiences of menopause, Burger et al and Dennerstein et al also used self-reported new-onset cycle irregularity as the marker of entry into the early perimenopause, and absence of menses for between 3 and 11 months for entry into the late perimenopause. 6,7 The 5-year longitudinal MWHS (257 women aged years) produced important information on the mean length of time that women usually spend passing through the menopausal transition. 8 The authors reported a median age at inception of perimenopause of 47.5 years and a mean duration of the perimenopausal transition of approximately 4 years. The natural menopause occurred at a mean of 51.3 years, and a mean of 49.5 years if currently smoking. The STRAW criteria The first standardized classification guidelines for female reproductive aging were proposed in 1 at the Stages of Reproductive Aging Workshop (STRAW). 1 These guidelines were proposed in order to develop a relevant and useful staging system, to revise the nomenclature and to identify gaps in knowledge that should be addressed by the research community (Table 1). The stages were nominated using the FMP as a reference point, and were based on changes in the pattern of menstrual cycles levels of follicle-stimulating hormone (FSH). Subjective data such as changes in menstrual flow or vasomotor Table 1 The Stages of Reproductive Aging (STRAW) criteria. Soules MR et al. J Womens Health Gender-Based Med 1; 1:

3 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) symptoms were thought to be too variable or too unreliably recorded to be used in the criteria. The proposed STRAW criteria were deemed unsuitable for application in women who smoked, were obese or underweight (body mass index <18 or >3), participated in intense regular physical training, had chronic menstrual irregularities or had abnormal reproductive anatomy. 1 The STRAW classification system includes five stages ( 5to 1) prior to the FMP (Stage ), and two stages (Stages þ1 and þ2) after the FMP (Table 1). The stages are primarily based on the characteristics of the menstrual cycle and secondarily on follicular phase FSH levels. Stages 5 and 4 (early reproductive age) cover the reproductive years and are characterized by regular menstrual cycles and normal (within the reference range for women of mid-reproductive age) follicular phase FSH levels. Stage 3 (late reproductive age) is characterized by regular menstrual cycles and elevated follicularphase FSH levels (higher than two standard deviations above the mean levels for mid reproductive Stage 4 subjects). 1 Stages 2 and 1 are the early and late menopausal transitions, and Stages þ1 and þ2 are early and late postmenopause. Stage 2 Entry into Stage 2, the early menopausal transition, is defined when the difference in cycle length between consecutive cycles is 7 days, or when cycle length becomes either 7 days longer or shorter than normal. Whether this normal cycle length pertains to the individual s usual cycle length (in midreproductive age) or a population mean cycle length was not clarified in the STRAW criteria. On the basis of the work by Taffe and Dennerstein 9, Landgren et al suggested that the early menopausal transition commences on the date of the first of more than two cycles in any consecutive series of 1 cycles, the length of which lies outside the normal reproductive range of days. 1 Given that some women may have regularly experienced cycle lengths shorter than 23 days or longer than 35 days for most of their reproductive life, prior knowledge of their normal cycle length would remain important when assigning reproductive stage. Stage 1 Entry into Stage 1, the late menopausal transition, occurs when an intermenstrual interval equal to or greater than two skipped cycles or > days is experienced in the setting of middle age in the absence of external factors that may interrupt normal menstrual cyclicity. Stage 1 ends and Stage þ1, the postmenopause, begins at the time of the FMP, which is only known to be the FMP 12 months after it is experienced. Whereas the menopausal transition ends at the time of the FMP, the late perimenopause ends 12 months after the FMP. There was no specific definition given for the menopause vs postmenopause, and it was stated that menopause is the anchor point that is defined after 12 months of amenorrhoea, indicating that the two terms could probably be used interchangeably. 1 Endocrine changes with advancing female reproductive age One of the earliest descriptions of the hormonal changes occurring with advancing reproductive age was by Sherman and Korenman. 11 Daily blood samples during a single cycle were taken from regularly cycling women (five aged 41 years, six aged years and a control group aged 18 3 years) at several time intervals over a period of years. The mean cycle length was 3, 25.4 and 23.5 days for the control, 41-year-old and year-old groups, respectively. In the oldest group, the decrease in cycle length was due to a shorter interval between the onset of menses, and the luteinizing hormone (LH) peak and serum oestradiol (E2) concentrations were lower during all three phases of the cycle than in the youngest group. FSH was higher throughout the cycle, despite the attainment of levels of E2 that might be expected to suppress its secretion, while LH remained indistinguishable from normal. 11 The irregular cycles of older women were marked predominantly by substantial increases in FSH and LH concentrations, with variable and erratic E2 levels and lower luteal-phase progesterone. The major features of reproductive aging in this first description were a monotropic rise in FSH secretion, folliculogenesis with variable evidence of ovulation, and periods of hypo-oestrogenaemia associated with substantial elevations in FSH but little change in LH. 11 The differential changes in FSH

4 1 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) 7 23 and LH, in particular, led the authors to hypothesize that there must be an ovarian regulating hormone (an inhibin) exerting a negative feedback effect specifically on FSH secretion, which decreases in the years before menopause due to a diminished number of follicles. In a second study by the same authors of eight regularly cycling women aged years and two women with cycles of variable length 12, the authors observed normal follicular maturation and corpus luteal function in the presence of high menopausal levels of LH and FSH and diminished E2 and progesterone levels. They hypothesized that the variability in cycle length during the menopausal transition was due to either irregular maturation of residual follicles with diminished gonadotropin stimulation, or anovulatory vaginal bleeding that may follow oestrogen withdrawal without evidence of corpus luteal function. An extensive and comprehensive longitudinal study of weekly urinary hormone excretion in 38 women aged > years was reported by Metcalf et al In 178 premenopausal women with regular cycles, FSH excretion was >5 IU/24 h (the upper limit observed in young reproductive-age women in the follicular phase, in contrast to a level >1 IU/24 h in postmenopausal women) in 31.5% of women aged >52 years, and 19.5% of women with a mean age of 42 years with a lesser but significant rise in LH were also identified. Only 52% of the irregular cycles from the older group (mean cycle length 29 days, range 18 2 days) were ovulatory. Anovulatory cycles occurred in 8% of cycles longer than days, compared with % of cycles shorter than days. Although anovulatory cycles increased in incidence towards the FMP, the last cycle before the FMP in four women was ovulatory. In this same study, irregular cycles were characterized by marked variations in oestrogen excretion (both persistently decreased and persistently increased excretion) and persistently increased FSH and LH excretion. Metcalf stated that there is no evidence for a gradual decline in ovarian function during the transition and the only generalization which can safely be made about menstrual cycles in perimenopausal women is that they are richly varied. Unpredictability is the norm, and is in marked contrast to the regular succession of ovulatory cycles observed in premenopausal women. 16 Metcalf also observed that with progression of the menopausal transition, cycles demonstrated progressively lower pregnanediol excretion and progressively higher gonadotropin excretion of different patterns. It was concluded that anovulatory cycles and cycles with elevated levels of FSH and LH are common in the perimenopause and rare in premenopausal women. In the British longitudinal FREEDOM study of daily urinary hormones in 112 women aged 3 58 years, five sequential endocrine stages of reproductive aging based on cycle length and changes in urinary hormone levels with advancing age were proposed The first stage was characterized by regular menstrual cycles with a mean early-follicular-phase FSH excretion of <5 IU/L, the second stage by regular cycles and FSH > 5 IU/L, the third stage by menstrual irregularity with the appearance of what Miro et al termed delayed-response cycles, the fourth stage by acyclical ovarian activity (but no evidence of ovulation or luteinization), and the fifth stage by ovarian quiescence with low oestrogen and high gonadotropin excretion. In the delayed-response cycles, the time between Day 1 of the cycle and the follicular phase increase in oestrogen excretion was prolonged. The second stage 19 was associated with a rise in early-cycle FSH levels and a decrease in follicular phase length; a finding noted by previous authors 22 and suggested to be the first indicator of reproductive aging (while menstrual cycles remain regular). In the third stage, the appearance of menstrual irregularity was associated with cycles of increased length due to a delayed (ovarian) response or lag phase ; an observation suggestive of a progressive loss in the ovary s ability to respond to FSH. 19 The authors proposed that a progressive decrease in ovarian responsiveness is associated with the decline in pregnanediol glucuronide (PdG) excretion and rise in oestrogen excretion during the luteal phase (following the lag phase ), and underlies the majority of the cycle elongation in the menopausal transition. 18 Shideler et al also observed low urinary oestrogen and high urinary FSH during the follicular phase of elongated cycles, followed by an increase in oestrogen excretion (to levels two to three times normal) and decreased PdG excretion during the luteal phase. These hormonal changes are demonstrated clearly in the illustration of two menstrual cycles from a single menopausal transition subject (Fig. 1). 23 Although the hormone features of the shortened follicular phase observed in late reproductive age have been associated with an increase in oestrogen excretion 24, there are few data on the hormonal features within shorter than normal cycles that are associated with new-onset cycle irregularity (as

5 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) Fig. 1. Data from Shidler et al Oestrogen (E1 conj) and pregnanediol gluconate (PdG) excretion in a single subject spanning two ovulatory cycles. The first cycle appears normal with a biphasic oestrogen curve, a normal PdG rise, and a simultaneous fall in oestrogen and PdG at the end of the cycle. In the second cycle, there is a -day interval ( lag phase ) between the menses and follicular-phase rise in oestrogen excretion. The oestrogen peak is normal and is followed by an initially normal-in-appearance rise in PdG. During the luteal phase, however, oestrogen rises again, forming a peak that is almost twice as high as the prior follicularphase peak. In addition, PdG levels only reach approximately half of the levels in the first cycle. This same excretion pattern has been fully described but not illustrated in the elongated cycles in the FREEDOM study publications outlined in the STRAW definition for Stage 2, the early menopausal transition). Between 5% and % of menstrual cycles in the menopausal transition have been observed to be shorter than days and some are as short as 14 days (Fig. 2). 9,18,25 This decrease in cycle length is greater than that associated with late reproductive age 26,27, where the follicular phase is decreased by a mean of 2 3 days 28 3 and has little effect on cycle regularity. The shortened follicular phase associated with late reproductive age has been postulated to be due to an advancement in (rather than a quicker) follicle development 27,31, and to be a result of an earlier rise in FSH levels 27,32,33 and increased FSH bioavailability 34 during the luteal follicular transition. On the other hand, a cycle length between 14 and days has often been assumed to be due to breakthrough anovulatory bleeding 35 38, and has not been described comprehensively until the recent serum hormone study by Hale et al 39 (see section headed Changes in hormone secretion patterns and ovulation ). Role of inhibins in the hormonal changes associated with advancing female reproductive age Before the discovery and measurement of inhibins, endocrine studies had established that the first indicators of advancing reproductive age were early-cycle monotropic elevations in FSH and a shortened follicular phase, with minimal effects on steroid hormone levels and menstrual cyclicity. However, study findings from women with irregular menstrual cycles in the menopausal transition have been more complex and variable. To summarize, there are substantial elevations in FSH, elevations in LH, unpredictable changes in E2 (higher or lower than normal), falls in progesterone and irregular menstrual cycles. 41 Since these earlier studies, assays of immunoreactive inhibin and, later,

6 12 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) Menstrual cycle length & hormonal characteristics before & after entry into the early & late transitions B B Cycle length (days) D MT1 D MT2 D D B Cycles counted back from FMP -1 Fig. 2. Menstrual-cycle length as a function of cycle number from final menses in a single subject. The onset of STRAW Stage 2, the early menopausal transition, is shown as MT1, and the onset of STRAW Stage 1, the late transition, is shown as MT2. Note the occurrence of ovulatory cycles (D), cycles with a defective luteal phase (C) and anovulatory cycles (B) after entry into the transition. Source: Shideler SE et al. Maturitas 1989; 11: inhibin A, inhibin B and anti-mullerian hormone (AMH) have been helpful in elucidating the underlying mechanisms for these reproductive hormone changes. The inhibins The inhibins are part of the complex hypothalamo-pituitary-ovarian axis that is a closed loop negative feedback system. Secretion of the pituitary gonadotropins is predominantly regulated by the inhibins and ovarian steroids. 42,43 Inhibin B is a product of antral follicle granulosa cells 44, and levels fall with the decline in follicle numbers with reproductive aging. 45,46 The fall in inhibin B in late reproductive age has been shown to trigger the monotropic rise in follicular phase FSH. 22,47 The elevations in FSH, in turn, maintain and sometimes increase the production of E2 from granulosa cells. 22,24,48 Inhibin was first isolated from bovine follicular fluid in , and both bovine and human inhibins were cloned soon afterwards. 5 It was undetectable in postmenopausal serum and in the serum of women who had had bilateral oophorectomies. Purified FSH was demonstrated to stimulate the ovarian production of inhibin 51, and it was subsequently shown that this stimulation was dose dependent during the follicular phase of the menstrual cycle. 52 Both inhibin A and B respond to exogenous FSH in the follicular phase of the human menstrual cycle. 53 The role of immunoreactive inhibin in reproductive aging was initially explored in early-follicular (cycle days 4 7) and mid-luteal phase (3 12 days prior to next menses) serum samples in women aged years. 54 Mean early-follicular phase immunoreactive inhibin was significantly lower in the year age group than in younger age groups (128 U/L in the year age group compared with 239, 235 and 7 U/L in the 29-, and 44-year age groups, respectively). Mean FSH levels were significantly higher in the oldest age group (13. IU/L, compared with 4.9, 5.5 and 5.2 IU/L in the three younger groups, respectively). E2 levels were similar in the , 29- and 44-year age groups, supporting the concept of differential feedback. A significant negative correlation between serum inhibin and FSH (r ¼.45, P <.5) was demonstrated, and there was also a significant negative relationship between inhibin and age. As age increased, mean FSH levels showed a two-phase linear increase with the change point estimated at approximately 43 years. The results were consistent with a role for inhibin, in addition to E2, in the regulation of FSH during the follicular phase of the menstrual cycle as a function of increasing age.

7 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) The Melbourne Women s Midlife Health Project 55 was the first major longitudinal study of the experiences of women transitioning from late reproductive age through to the FMP. Early-follicular phase serum samples indicated that falls in early-cycle inhibin occurred before changes in E2, and that while inhibin decreased from late reproductive age, the fall in E2 was most significant in the late menopausal transition group (women with amenorrhoea for >3 months). There was marked variability in FSH, E2 and inhibin concentrations in all the reproductive groups and a substantial degree of overlap. With inclusion of the late transition group, FSH was negatively correlated with E2 (r ¼.3) and inhibin (r ¼.39), whilst inhibin was positively correlated with E2 (r ¼.45). It was concluded that an increase in serum FSH and a decrease in E2 and inhibin were the major endocrine changes associated with the menopausal transition. 6,55 Shortly after the development of assays specific for dimeric inhibins A and B 56,57, Klein et al showed that the monotropic FSH rise seen in older ovulatory women was associated with a decrease in the follicular phase levels of inhibin B but not inhibin A. 58 Inhibin B assays performed on the serum samples from the third year of the Melbourne Women s Midlife Health Project showed a marked decrease in follicular-phase inhibin B in women entering the early menopausal transition, without any change in E2 or inhibin A. 22 Welt et al also showed that follicular phase (early, mid and late) inhibin B was lower and E2 was higher in older women (aged vs aged <35 years). 59 Data from the longitudinal component of the study suggested that reproductive aging is accompanied by a fall in both inhibin A and B, and that the decrease in inhibin B preceded any decrease in inhibin A or E2 and may even be associated with a rise in E2. They proposed that loss of inhibin B negative feedback on FSH is the most important factor in the increase in FSH with advancing reproductive age. 59 Muttikrishna et al studied two groups of older regularly cycling women, one with normal FSH (<8 IU/L, n ¼ 1) and one with raised FSH values (>8 IU/L, n ¼ 6), and compared daily serum hormone levels taken throughout the cycle with a group of young women aged years. The older group with elevated FSH levels had lower concentrations of inhibin B in the early follicular phase and lower concentrations of inhibin A prior to the mid-cycle LH surge and in the mid-luteal phase than the older women with normal FSH concentrations. They concluded that the rise in early-follicular phase FSH in older women was associated with a decrease in circulating inhibin B concentrations in the early follicular phase, and that lower circulating inhibin A concentrations in the luteal phase may also contribute. Inihibin A is a product of the corpus luteum, and its role in inhibition of gonadotrophin secretion is minor if present. Its function within the hypothalamo-pituitary-ovarian axis has yet to be clarified. 61,62 In a recent study of 77 women classified into STRAW Stages 4, 3, 2 and 1, mean ovulatorycycle FSH, LH and E2 levels increased with progression of STRAW stage and luteal phase progesterone decreased. 28 There were two, zero, one and nine anovulatory cycles in the Stage 5 and 4 (n ¼ 21), Stage 3 (n ¼ 16), Stage 2 (n ¼ 17) and Stage 1 (n ¼ 23) subjects, respectively. With inclusion of the anovulatory cycles in the comparative analysis, the increase in FSH and LH across the groups was accentuated but the increase in E2 was no longer detectable. Early-cycle and mean-cycle (ovulatory and anovulatory included) inhibin B decreased progressively across the STRAW stages, with the lowest levels detected in the elongated ovulatory cycles and anovulatory cycles in the late menopausal transition. 28 Inhibin A levels followed those of E2 levels, peaking during E2 peaks and increasing during ovulatory cycles across the STRAW stages. These studies indicate that a decrease in inhibin B and not inhibin A is the major factor influencing the elevations in FSH and later LH with advancing reproductive age. The falls in inhibin B are a reflection of the diminishing ovarian follicle pool. Role of AMH in advancing female reproductive age AMH is a dimeric glycoprotein and a member of the family of growth and differentiation factors that are best known for their effects on sexual differentiation. 63 Under the influence of AMH, the gonads differentiate into testes; AMH is not expressed during female sexual differentiation. After development of the testes and ovaries, AMH is produced by Sertoli cells in the male or granulosa cells in the female. It is produced steadily by primary follicles until they grow to approximately 8 mm in diameter. 64 Most AMH is seen in the granulosa cells of the pre-antral and small antral follicles (<4 mm), with almost no

8 14 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) 7 23 expression in follicles with a diameter of >8 mm. 65 AMH inhibits FSH-dependent follicle growth in a time-dependent manner, mainly as a result of reduced granulosa cell proliferation, and may play a role in follicle recruitment and selection. 66 Unlike the gonadotrophins and inhibins, AMH does not appear to vary throughout the menstrual cycles or through pregnancy. 67 In addition, there does not appear to be any correlation between AMH, FSH or E2 levels in the mid reproductive age menstrual cycle 68 or any effect on AMH synthesis by FSH. 67 In one of the few studies (238 women aged years with FSH levels <1 IU/L) on the changes in AMH with reproductive aging, AMH on Day 3 5 was stable between the ages of 18 and 29 years, then fell more than 1-fold across the year age groups with minimal change in FSH. 63 In another study of 77 women between the ages of 21 and 55 years, AMH levels showed a -fold decline between the year and 45 5 year age groups. 69 In subjects aged >45 years (including STRAW Stage 3, 2 and 1 groups), the FSH:inhibin B ratio rather than the individual s age was inversely correlated with AMH (Fig. 3). No change in AMH was observed throughout individual menstrual cycles, and AMH did not appear to be influenced by the ovulatory status of a cycle. Changes in hormone secretion and ovulation patterns over the menopausal transition A number of factors are likely to contribute to the marked variability in serum and excretion hormone levels seen in endocrine studies, including variation in normal levels between individuals, variations in the categorization or classification of subjects, and variations in the secretion patterns of hormones. Changes in excretion patterns in elongated but ovulatory cycles occurring during the menopausal transition have been illustrated 18,23 and the mechanisms discussed at length by other authors. 32,7 The relatively small study by Hale et al was designed specifically to explore serum hormone (including inhibins and AMH) secretion patterns in individual cycles during both short and long cycles in the menopausal transition. 28 Serum samples from 77 women were obtained three times weekly during the menstrual, follicular and luteal phases of one cycle and the menstrual and follicular phases of the subsequent cycle. Subjects were categorized as STRAW Stage 5 and 4 (n ¼ 21), Stage 3 (n ¼ 16), Stage 2 (n ¼ 16) and Stage 1 (n ¼ 23). 28 Examples of two normal ovulatory cycles (a Stage 4 and a Stage 2 subject) and two anovulatory cycles (a Stage 4 and a Stage 1 subject) are illustrated in Fig. 4. AMH (ng/ml) Cycle Type 1 Cycle Type 2 Cycle Type 3 MRA FSH:Inhibin B ratio Fig. 3. Regression plot of early-follicular-phase follicle-stimulating hormone (FSH):inhibin B ratio and anti-mullerian hormone (AMH) in ovulatory cycles of STRAW Stage 5/ 4 (mid-reproductive age) and Types 1, 2 and 3 (correlation coefficient r ¼.83, P <.1). The dotted linesrepresent the upperand lowercut-off points of thefsh:inhibin B ratio (.25) and AMH (.25 ng/ml) asa meansto distinguish Type 1 from Type 2 and 3 cycles. Source: Hale GE et al. JCEM 7; Robertson DM et al. Menopause 8 (in press). 69

9 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) Luteal out-of-phase (LOOP) event In the first report, increased ovulatory cycle E2 and decreased luteal-phase progesterone were observed with progression of STRAW stage ( 4 to 1). 28 The E2 levels were particularly variable (normal to high) during the menstrual and luteal phases of the cycles. In individual ovulatory cycles, these high menstrual- and luteal-phase E2 levels were all found to be associated with a common aberrant E2 secretion pattern. The aberrant E2 secretion pattern was termed a luteal out-of-phase (LOOP) event 39, and was similar to the rise in luteal-phase oestrogen excretion described during elongated ovulatory cycles in the FREEDOM study 18 and in a single subject illustrated by Shideler et al. (Fig. 1). 23 The LOOP events occurred in six subjects (three STRAW Stage 2 and three STRAW Stage 1) and were all characterized by a second high peak in E2 at the mid-luteal point, causing markedly elevated E2 levels during the luteal phase of the current ovulatory cycle and the menstrual phase of the subsequent cycle (Fig. 5). In a further five STRAW Stage 2 and 1 ovulatory cycles (total of 11 cycles), E2 levels were >35 pmol/l during the first menstrual phase, indicating that a LOOP may have occurred late in the cycle prior to testing. In four of these 11 cycles, the LOOP event triggered an LH peak and an ovulatory episode early in the subsequent cycle, resulting in a markedly shortened (14 19 days) ovulatory cycle (Fig. 5A, B and D). Prior to the initiation of a LOOP event, the cycles were characterized by higher FSH and lower inhibin B in comparison with the normal biphasic STRAW Stage 2 and 1 ovulatory cycles. 39 The FSH levels, in particular, remained high and the inhibin B levels were mostly undetectable until mid-cycle. The FSH levels dropped by the time of the mid-luteal rise in E2, and were very low by the beginning of the following menstrual phase. LOOP events were associated with all cycles <21 days in length and were responsible for the group differences in E2 and progesterone that was observed in the original group comparison. 28 The only elongated ovulatory cycle captured (68 days) in a STRAW Stage 1 subject concluded with a LOOP event. Two publications of the British FREEDOM study 17,18 and one from the US SWAN study 71 reported within-cycle hormone excretion patterns consistent with LOOP events. The FREEDOM data reported that elongated cycles were more likely to have high E1G and lower PdG excretion during the second half of an elongated cycle (the same features in the late menopausal transition with a superimposed LOOP event illustrated in Fig. 4D). There was also reference made to eight unusually short menstrual cycles (mean 19 days) that exhibited very high basal levels of E1G excretion 18, but these cycles were excluded from the group analyses and were not described any further. In the SWAN study, functional data analysis of across-cycle urinary data was performed in order to ascertain characteristic changes in excretion patterns associated with changes in cycle length. The analysis revealed second rises in E1G excretion during the luteal phase of the more elongated cycles (consistent with a LOOP event) and higher early-cycle E1G excretion in the short cycles. 71 The occurrence of LOOP events associated with markedly high FSH and low inhibin B levels suggests that the higher and more sustained elevations in FSH can lead to selection of a dominant follicle earlier than normal during a follicular phase 27, and sometimes as early as the mid-luteal phase of an existing cycle. The FSH levels would have to be sufficiently high to overcome suppression from inhibin B, E2 and progesterone during the time when the dominant follicle (responsible for the LOOP event) is undergoing its FSH-dependent growth (possibly around the time of mid-cycle). In addition, a LOOP event would require the presence of at least one antral follicle ready for dominant follicle selection during the luteal phase of a cycle. The consistent timing of the initiation of the LOOP events during the luteal phase suggests that ovarian follicles can undergo FSH-driven development around mid-cycle. This may relate to Baerwald et al s theory on the consistency and position of follicle recruitment waves within the normal menstrual cycle. 39,72 Paradoxically, although FSH levels are high and sustained and inhibin B levels are very low during the follicular phase before the LOOP event, FSH levels fall back to mid-reproductive levels during the menstrual and follicular phases of the subsequent cycle. The LOOP observation therefore provides an explanation for the variability in early cycle FSH levels during reproductive aging, and also an explanation regarding why early cycle or random single FSH measurements are less reliable than menstrualcycle patterns in predicting reproductive stage, especially during irregular cycles. 73

10 16 E2 LH INHA A B C D Lag phase Lag phase Lag phase Day of cycle Day of cycle Day of cycle Day of cycle Fig. 4. Illustration of luteal out-of-phase (LOOP) events that were observed to occur in four ovulatory cycles in four STRAW Stage 2 subjects (A C) and one STRAW Stage 1 subject (D). The broken vertical lines represent mid-cycle in the first cycle, and the shaded areas represent the menstrual and follicular phases of the second cycle. The LOOP event is characterised by a second rise in E2 at the mid-luteal point. The peak in E2 occurs around the time of the subsequent menstrual phase, and triggers an LH peak and ovulatory episode early in the subsequent cycle in Cycles A, B and D but not in Cycle C. Individual A: First cycle, 35 days; second cycle, 19 days. Individual B: First cycle, 27 days; second cycle, 14 days. Individual C: First cycle, 26 days; second cycle, 34 days. Individual D: First cycle, 68 days; second cycle, days. In Cycles A D, there are high and sustained FSH levels throughout the follicular phase of the first cycle. FSH and LH are up to two-fold higher in the elongated STRAW Stage 1 cycle (D) compared with the normal-length cycles (A C). Inhibin B levels are variable, not predictable or cyclical, and are often undetectable (<12.8 pg/ ml), particularly in the elongated cycle (D). Inhibin A, on the other hand, tends to mirror the rise and falls in E2 secretion P FSH INHB G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) 7 23

11 Age 22: normal ovulatory 8 Age 47: normal ovulatory 3 E2 pmol/l E2 pmol/l Menses Day of cycle Age 22: anovulatory Day of cycle Menses P nmol/l & FSHIU/L P nmol/l & FSHIU/L E2 pmol/l E2 pmol/l Day of cycle Age 49: anovulatory Day of cycle Fig. 5. Individual hormone profiles in four menopausal transition subjects. E2, oestradiol; FSH, follicle-stimulating hormone P nmol/l & FSHIU/L P nmol/l & FSHIU/L G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9)

12 18 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) 7 23 The LOOP phenomenon is likely to represent one of the phases in the normal progression of reproductive aging, and may be a common endocrinological event that heralds the onset of irregular menstrual cycles and the early menopausal transition. Once FSH reaches a critically high level, it appears that FSH is capable of causing out-of-phase follicle development and disruption of the normal cyclical follicle recruitment and ovulatory process. The markedly high FSH levels are most likely the result of low or absent levels of inhibin B, which in turn are due to the diminishing numbers of ovarian follicles. 74 Follicular depletion accelerates dramatically in the last decade of menstrual life, and the mean number of primordial follicles in the ovaries of regularly menstruating women are around 1392þ/ 355, 1-fold greater than in mid-life women with irregular menses. 45 Classification of the endocrine changes from late reproductive age to menopause On the basis of the most recent AMH, inhibin and secretion pattern data, Robertson et al proposed a classification system based on four cycle types. 69 This system builds upon Stages 1 4 proposed by Miro et al based on the FREEDOM urinary data, and includes the serum inhibin and AMH data. Type 1 was characterized by an isolated and age-related decrease in AMH (no elevation in FSH); Type 2 by low inhibin B levels, elevated early cycle FSH and further falls in AMH; Type 3 by falls in progesterone, further elevations in FSH and elevations in LH; and Type 4 by anovulation, undetectable inhibin A and B levels, and further marked elevations in FSH and LH. 69 Although the type number tends to increase through the four STRAW stages, no cycle type is strictly associated with any one STRAW stage (Table 2). It is likely that individuals progress through the cycle types in a mostly forward and occasionally a back-and-forward fashion, or may even skip a stage or more altogether. 69 Given the comprehensive urinary and more recent serum hormone data on women across the menopausal transition, the hormonal and menstrual features of each stage can be summarized as follows. Mid to late reproductive age; STRAW 4 Regular menstrual cycles Decreased AMH Normal FSH Late reproductive age; STRAW 3 Regular menstrual cycles Further falls in AMH Elevations in early-cycle FSH Falls in early-cycle inhibin B Table 2 Distribution of cycle types within the STRAW stages. Study group MRA LRA EMT LMT Total no. of cycles Corresponding STRAW stage Stage 4 n ¼ 21 Stage 3 n ¼ 16 Stage 2 n ¼ 17 Stage 1 n ¼ 22 Type 1 n (%) 19 (1) 3 () 2 (9) 14 Type 2 n (%) 16 (1) 1 (67%) 8 (36) 24 Type 3 n (%) 2 (13%) 3 (14) 5 Anovulatory n (%) 9 (41) 9 MRA, mid-reproductive-age group (two anovulatory cycles not categorized); LRA, late-reproductive-age group (no anovulatory cycles); EMT, early-menopausal-transition group (two anovulatory cycles not categorized); LMT, late-menopausal-transition group (nine anovulatory cycles). Sources: Hale GE et al. J Clin Endocrinol Metab 7; 92: Robertson DM et al. Menopause 8 (in press). 69

13 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) Early menopausal transition; STRAW 2 Further falls in AMH and inhibin B Further elevations in FSH and first observable elevations in LH Falls in luteal-phase progesterone levels Initial appearance of LOOP cycles associated with more marked falls in luteal-phase progesterone, and variable to high menstrual- and/or luteal-phase E2 Cycle irregularity with a predominance of short and normal-length cycles Late menopausal transition; STRAW 1 As in 3A but further elevations in FSH and LH Continuing LOOP events but more with increasing lag phase and less frequent high peaks in E2 Cycle irregularity with a predominance of elongated menstrual cycles Menopause; STRAW þ1 Anovulatory cycles and increase in mean cycle length Further elevations in FSH and LH Undetectable inhibins Variable to low E2 Biomarkers of reproductive aging When the STRAW guidelines were proposed, FSH was considered to be the most suitable and readily available biomarker to indicate the onset of late reproductive age, despite the limitations in its ability to predict the stage of menopausal transition or the FMP. 73 Although an early-cycle FSH level of > IU/L is an independent marker of the late menopausal transition, it is less predictive of the late transition than menstrual bleeding markers, such as amenorrhoea for > days. 75,76 Other biomarkers such as inhibin B 77 and AMH 63 appear to change earlier in reproductive aging than FSH, but like FSH, inhibin B tends to vary with the ovulatory status of the cycle, particularly in the menopausal transition. 28 AMH, on the other hand, remains unchanged throughout the menstrual cycle 78, is not dependent on the ovulatory status of a cycle, is more predictive of the number of early antral follicles than either FSH or inhibin B 79, and starts to decline from approximately 3 years of age. 8 AMH levels appear to represent both the quantity and quality of the ovarian follicle pool, independent of the timing of menstrual cyclicity. 8,81 More data on AMH levels and proximity to final menstrual period are needed to assess its ability however to predict reproductive stage. Summary Female reproductive aging is a challenging area to research and define because of the variability of the physiological and menstrual manifestations, and the complexity of the hypothalamo-pituitaryovarian axis. The 1 STRAW reproductive staging criteria have been helpful in providing standardized nomenclature and methods of grouping research subjects into late reproductive age, menopausal transition and menopause. These staging criteria are intended to be guidelines rather than strictly applied diagnoses, and it is understood that every stage may not occur in individuals, and if they do occur in any one individual, they may not occur in the exact sequence provided. The endocrine and menstrual cycle features associated with each of the STRAW stages are summarized above. The features in each of the stages are all a result of the decline in ovarian follicle numbers and the consequent fall in inhibin B. The resultant high FSH levels appear to be responsible for the shortened follicular phase in late reproductive age, and disruption of the normal cyclical follicle recruitment and cycle irregularity in the menopausal transition. With the onset of the menopausal transition, ovarian follicles have reached a critically low level, and FSH has reached a critically high level where it can precipitate LOOP events. As a result, cycles become either shorter, longer or (more

14 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) 7 23 commonly) a mixture of the two. In menstrual cycles where the early-cycle FSH is sufficiently high to trigger a LOOP event, the following cycle is paradoxically characterized by low early-cycle FSH because of the high late-luteal and menstrual phase E2. Thus, LOOP events may explain the lack of reliability in early cycle FSH levels, especially with the onset of the menopausal transition. In addition, the markedly high luteal and menstrual phase E2 levels may trigger E2-related symptoms including headache, mood swings and heavy menstrual loss.,82 Although the reason remains unclear, luteal-phase progesterone becomes lower than mid-reproductive-age levels during the menopausal transition, and even more so during ovulatory cycles with LOOP events. The number of ovarian follicles that remain available when FSH reaches the level at which it disrupts normal cyclical folliculogenesis probably determines the length of time during which an individual will experience dramatic swings in E2 and erratic alternating short and long cycles. The fewer ovarian follicles left, the less likely it is that the markedly high FSH will be able to trigger LOOP events, and the more likely it is that cycle irregularity will progress to low E2 levels, elongated cycles and anovulation. The exact mechanisms involved in the LOOP events remain unknown and will only be fully elucidated with careful detailed ultrasonographic studies on follicle activity during irregular cycles. AMH appears to be a stable biomarker of the remaining ovarian follicles (and, possibly, their capacity to lead to ovulation), and may become a reliable indicator of the reproductive stage and proximity to the FMP in the future. Practice points the STRAW criteria were proposed in 1 and are based on menstrual-cycle characteristics and early-cycle FSH levels late reproductive age (STRAW Stage 3) is characterized by elevated early-cycle FSH levels, decreased AMH and the maintenance of regular menstrual cycles the early menopausal transition (STRAW Stage 2) is characterized by elevated FSH, further decreases in AMH and inhibin B, and the appearance of menstrual-cycle irregularity associated with LOOP events (high and erratic E2 levels, low luteal-phase progesterone), causing alternating normal, short and longer cycles from to 55 days in length the late menopausal transition (STRAW Stage 1) is characterized by further elevations in FSH, elevations in LH, and further falls in AMH and inhibin B. Cycle irregularity remains a feature but with an increase in mean cycle length as markedly elevated FSH levels fail to trigger normal cyclical follicular development or LOOP events. As a result, normal ovulatory cycles become less frequent, as do periods of LOOP-related high erratic E2 levels. Anovulatory cycles predominate as the FMP approaches the early menopause (STRAW Stage þ1) is characterized by amenorrhoea, high and stable FSH and LH levels, undetectable inhibins and variable to low E2 levels the late postmenopause (STRAW Stage þ2) is characterized by amenorrhoea, high and stable FSH and LH levels, undetectable inhibins and low stable E2 levels Research agenda reproductive endocrine studies have established the critical role of ovarian follicle depletion, falls in inhibin B and increases in gonadotrophin in the hormonal and physiological characteristics associated with the progression from late reproductive age to menopause. The effect of elevated FSH in disrupting the normal cyclical folliculogenesis remains to be clarified by ultrasonographic studies. In addition, the role of falls in AMH with reproductive aging in hormonal dynamics and physiological processes remains to be clarified

15 G.E. Hale, H.G. Burger / Best Practice & Research Clinical Obstetrics and Gynaecology 23 (9) early-cycle measurements of FSH and inhibin B are relatively stable from cycle to cycle in late reproductive age and the postmenopause. However, during the early and late transitions, they are less consistent and FSH, in particular, can alternate between low (mid-reproductive age levels) and high (menopausal levels) in consecutive cycles. Single measurements of either hormone are therefore unreliable in predicting STRAW stage. AMH levels do not appear to be affected by the menstrual cycle, the ovulatory status of a cycle or the cycle length. It therefore promises to be an accurate marker of reproductive aging, but further large studies on AMH levels in women at all STRAW stages are needed to clarify this. References *1. Soules MR, Sherman S, Parrott E et al. Stages of Reproductive Aging Workshop (STRAW). J Womens Health Gender-Based Med 1; 1: WHO Scientific Group. Research on the menopause in the 199 s. A report of the WHO Scientific Group. Geneva: World Health Organization, pp Brambilla DJ, McKinlay SM & Johannes CB. Defining the perimenopause for application in epidemiologic investigations. Am J Epidemiol 1994; 1: *4. Mitchell ES, Woods NF & Mariella A. Three stages of the menopausal transition from the Seattle Midlife Women s Health Study: toward a more precise definition. Menopause ; 7: Dennerstein L. Well-being, symptoms and the menopausal transition. Maturitas 1996; 23: *6. Burger HG, Dudley EC, Robertson DM et al. Hormonal changes in the menopause transition. Rec Prog Horm Res 2; 57: Dennerstein L, Dudley EC, Hopper JL et al. A prospective population-based study of menopausal symptoms. Obstet Gynecol ; 96: McKinlay SM, Brambilla DJ & Posner JG. The normal menopause transition [see comments]. Maturitas 1992; 14: Taffe JR & Dennerstein L. Menstrual patterns leading to the final menstrual period. Menopause 2; 9: Landgren BM, Collins A, Csemiczky G et al. Menopause transition: annual changes in serum hormonal patterns over the menstrual cycle in women during a nine-year period prior to menopause. J Clin Endocrinol Metab 4; 89: Sherman BM & Korenman SG. Hormonal characteristics of the human menstrual cycle throughout reproductive life. J Clin Investig 1975; 55: Sherman BM, West JH & Korenman SG. The menopausal transition: analysis of LH, FSH, estradiol, and progesterone concentrations during menstrual cycles of older women. J Clin Endocrinol Metab 1976; 42: Metcalf MG, Donald RA & Livesey JH. Pituitary-ovarian function in normal women during the menopausal transition. Clin Endocrinol 1981; 14: Metcalf MG, Donald RA & Livesey JH. Classification of menstrual cycles in pre- and perimenopausal women. J Endocrinol 1981; 91: Metcalf MG, Donald RA & Livesey JH. Pituitary-ovarian function before, during and after the menopause: a longitudinal study. Clin Endocrinol 1982; 17: Metcalf MG. The approach of menopause: a New Zealand study. NZ Med J 1988; 11: Miro F, Parker SW, Aspinall LJ et al. Relationship between follicle-stimulating hormone levels at the beginning of the human menstrual cycle, length of the follicular phase and excreted estrogens: the FREEDOM study. J Clin Endocrinol Metab 4; 89: *18. Miro F, Parker SW, Aspinall LJ et al. Origins and consequences of the elongation of the human menstrual cycle during the menopausal transition: the FREEDOM study. J Clin Endocrinol Metab 4; 89: Miro F, Parker SW, Aspinall LJ et al. Sequential classification of endocrine stages during reproductive aging in women: the FREEDOM study. Menopause 5; 12: Reyes FI, Winter JA & Faiman C. Pituitary-ovarian relationships preceding the menopause. A cross-sectional study of serum follicle-stimulating hormone, luteinizing hormone, prolactin, estradiol, and progesterone levels. Am J Obstet Gynecol 1977; 129: Ahmed Ebbiary NA, Lenton EA & Cooke ID. Hypothalamic-pituitary ageing: progressive increase in FSH and LH concentrations throughout the reproductive life in regularly menstruating women. Clin Endocrinol 1994; 41: Burger HG, Cahir N, Robertson DM et al. Serum inhibins A and B fall differentially as FSH rises in perimenopausal women [Erratum appears in Clin Endocrinol (Oxf) 1998; 49: 55]. Clin Endocrinol 1998; 48: *23. Shideler SE, DeVane GW, Kalra PS et al. Ovarian-pituitary hormone interactions during the perimenopause. Maturitas 1989; 11: Santoro N, Brown JR, Adel T et al. Characterization of reproductive hormonal dynamics in the perimenopause. J Clin Endocrinol Metab 1996; 81: Burger HG, Robertson DM, Baksheev L et al. The relationship between the endocrine characteristics and the regularity of menstrual cycles in the approach to menopause [see comment]. Menopause 5; 12: van Zonneveld P, Scheffer GJ, Broekmans FJ et al. Do cycle disturbances explain the age-related decline of female fertility? Cycle characteristics of women aged over years compared with a reference population of young women. Hum Reprod 3; 18: Klein NA, Harper AJ, Houmard BS et al. Is the short follicular phase in older women secondary to advanced or accelerated dominant follicle development? J Clin Endocrinol Metab 2; 87: