Epidemiology

Most estimates of age at natural menopause are based on samples of Caucasian women in Western societies. In one large, comprehensive, prospective cohort study of mid-aged, Caucasian U.S. women (the Massachusetts Women’s Health Study [MWHS]) the age at natural menopause occurred at 51.3 years,⁸ confirming prior reports.⁹ The Study of Women’s Health Across the Nation (SWAN), a multicenter, multiethnic, community-based cohort study of women and the menopausal transition, reported the overall median age at natural menopause to be 51.4 years, after adjustment for other factors.¹⁰ Studies performed outside the United States suggest that Africans,¹¹ African Americans,¹² and Hispanics of Mexican descent¹³ experience menopause at an earlier age than Caucasian women, as opposed to Japanese¹⁴ and Malaysian¹⁵ women, who report a similar median age of menopause to women of European descent.

Late menopause is defined as an FMP that occurs after 54 years, and early menopause occurs at ages 40 to 45. Both are present in about 5% of women. Approximately 1% of women experience hypergonadotropic amenorrhea or premature ovarian failure (POF) before age 40. Many factors affect age at natural menopause (Table 135-1).

Table 135-1. Etiologic Factors in Early Menopause

Lower educational attainment and unemployment have been independently associated with earlier age at menopause^10,16,17 and may be markers for elevated biopsychosocial stress. Women who are separated, divorced, or widowed have been shown to have an earlier menopause than women who are married.¹⁰ Age at natural menopause for parous women has been reported to occur significantly later than for nulliparous women.^8,19–22 Gold et al. and Cramer et al. observed a trend of increasing age at menopause with increasing number of life births, and that prior use of oral contraceptives was associated with earlier age at natural menopause¹⁶; however, a slight prolongation of the reproductive life-span has been associated with oral contraceptive use.^24,25 The proposed mechanism by which parity and use of oral contraceptives may result in later age at natural menopause involves reducing ovulatory cycles earlier in life and thus preserving oocytes longer, resulting in later menopause.¹⁰ Some studies show that women with a lower body mass index (BMI) experience an earlier menopause¹⁷; other studies have not confirmed this finding.^27,28

Genetic factors may influence the timing of menopause. The blepharophimosis gene, located on chromosome 3,¹⁸ and X chromosome deletions such as POF 1 and POF 2 genes,¹⁹ have been shown to predispose women to an earlier menopause. In one study by Tibilette et al., pedigree analysis revealed a dominant pattern of inheritance of early menopause and POF through maternal or paternal relatives; because POF and early menopause have been shown to share the same genetic features, they may actually represent variable expression of the same genetic disease.²⁰ Fragile X syndrome premutation carriers are prone to premature menopause. The PvuII polymorphic allele for estrogen receptor alpha has been associated with a slightly earlier age at natural menopause and a twofold increase in risk for hysterectomy.²¹

Environmental toxicants may play a role in early menopause. A large body of literature shows that current smokers tend to experience menopause at an earlier age (1 to 2 years) than nonsmokers^31–34 and may have a shorter menopausal transition.⁸ It has been shown that polycyclic hydrocarbons in cigarette smoke are toxic to ovarian follicles²² and may lead to their loss and thus an earlier menopause in smokers. Irradiation and chemotherapy with alkylating cytostatics have also been implicated as causes of early menopause.²³ Evidence suggests that galactose consumption through the ingestion of high-lactose dairy foods may be a dietary risk factor, and that galactose metabolism, as measured by galactose-1-phosphate uridyl transferase, may be a genetic risk factor for early menopause.²⁴

Harlow et al. observed that women with a history of medically treated depression had a 20% increased rate of entering perimenopause sooner than women with no depression history, after adjustment for age, parity, age at menarche, education, cigarette smoking, and BMI.²⁵

Epidemiology gives answers about populations, whereas clinical medicine deals with patient samples and individuals. For example, in population-based studies,⁸ no global increased prevalence of depression has been associated with the menopause transition, whereas in clinical samples,²⁶ depression around menopause has reportedly increased. Furthermore, symptoms vary among women, and the distinction between populations versus individuals must be made when one is evaluating epidemiologic factors related to menopause.

Pathogenesis

The basis of reproductive senescence in women is oocyte/follicle depletion in the ovary. Developmentally, a woman attains her peak oocyte complement at 20 weeks’ gestational age.²⁷ Between 20 and 40 weeks’ gestation, two thirds of a woman’s oocyte complement is lost, and total oocyte counts drop from a mean of about 6 to 8 million to 1 to 2 million.²⁸ The most massive wave of atresia (rate of follicle loss) that a woman ever experiences happens before she is born. At the onset of puberty, germ cell mass has been reduced to 300,000 to 500,000 units.²⁹ Subsequent reproductive aging consists of a steady loss of oocytes through atresia or ovulation and does not necessarily occur at a constant rate.

Atresia is an apoptotic process. During the 35 to 40 years of reproductive life, 400 to 500 oocytes will be selected for ovulation. By menopause, only a few hundred follicles remain.²⁹ The relatively wide age range (42 to 58 years) for menopause in normal women seems to indicate that women may be endowed with a highly variable number of oocytes, or that the rate of oocyte loss varies greatly.⁷

All levels of follicular loss are influenced by molecular genetics. Genetic modifiers of oocyte/follicle endowment, development, and atresia have been shown to play a major role in the mouse ovary.³⁰ The KIT receptor, present on oocyte and theca cells in ovarian follicles, and its ligand, KIT LIGAND, produced by granulosa cells, are encoded at the Kit gene and the Mgf gene, respectively. Mutations in these genes alter the expression of KIT and KIT LIGAND proteins, resulting in alterations in granulosa cell proliferation and/or oocyte growth in mice.³¹ Growth differentiation factor-9 (GDF-9), a growth factor that is secreted by oocytes in growing ovarian follicles, has been shown to promote the growth, development, and survival of human ovarian follicles in organ culture.³² Finally, FOXL2 (forkhead transcription factor gene) is thought to be a highly conserved regulator of vertebrate ovarian development and has been implicated as an etiologic factor in the ovarian failure noted in patients with blepharophimosis ptosis epicanthus inversus syndrome (BPES).³³ It is interesting to note that FOX03 shunts follicles through development at a more rapid pace and causes follicle depletion by promoting presumptive development and apoptosis. Furthermore, genetic modifiers indicate that there are acquired ways to change the size of the follicle pool.

Concurrent with the loss of ovarian follicles as a woman transitions to menopause are hormonal changes in the hypothalamic-pituitary-ovarian axis. Follicle-stimulating hormone (FSH) is an established indirect marker of follicular activity; as follicle numbers decline, FSH levels increase.³⁴ An elevated level is often the first clinically measurable sign of reproductive aging. Large cross-sectional studies have reported a progressive, quantitative rise in FSH with age.³⁵

In the late reproductive years, initial elevations in FSH are most prominent in the early follicular phase of the menstrual cycle but are intermittent and do not occur in every cycle.³⁶ This increase is first detectable some years before any clinical indications of approaching menopause are evident.³⁴ The rise in FSH appears to be the result of a decline in inhibin B, a dimeric protein that reflects the fall in ovarian follicle numbers. In reproductive life, inhibin serves to selectively inhibit FSH by binding to receptors on the anterior pituitary.³⁷ Estradiol is stable or even elevated during the earlier menopause transition; closer to the final menstrual period, a decline is clearly observed.³⁸ Findings from the Melbourne Women’s Midlife Health Project, a cohort of women followed through the menopause transition, confirm that a decline in inhibin B precedes the increase in FSH and the decline in estradiol that occur later in the transition.³⁹ The remaining follicles are less likely to function normally, which may lead to erratic follicular development and dysregulation of folliculogenesis.^52–54

Although FSH and estradiol vary near menopause, steroidogenic enzymes appear to be completely absent in the postmenopausal ovary after all functional follicles are lost.⁴⁰ Between the ages of 20 and 40 years, concentrations of total testosterone have been reported to fall by about 50%.⁴¹ This age-related decline does not change further during the transition years.⁴² Similarly, dehydroepiandosterone (DHEA) and its sulfate, DHEAS, decline with age.^57,58 Because circulating sex hormone–binding globulin (SHBG) decreases across the menopausal transition, free androgen levels actually rise, as indicated by a small increase in free androgen index (T ÷ SHBG × 100).⁴²

Androstenedione, which remains relatively stable during the transition, is converted to estrone in extraglandular tissue. This accounts for almost all the estrogen in circulation after menopause. When ovulation stops, concurrent with a woman’s FMP, serum progesterone levels are invariably low.⁴³ Luteinizing hormone (LH) eventually increases, although at a slower rate than FSH.

Clinical Features

Impending menopause is clinically evident with menstrual cycle changes. A normal menstrual cycle ranges anywhere from every 21 to 35 days.⁴⁴ Anovulatory cycles become more common as women progress through the menopause transition, and cycle lengths become increasingly variable. An analysis of prospectively kept menstrual calendars has shown that cycle lengths rise in mean length to more than 35 days during the last 10 cycles before cessation of menses. After reaching this level, cycles decrease in length for some and continue to increase for others.⁴⁵

Women who have skipped more than three but less than 12 menses are very likely (estimated probability, 0.93) to stop menstruating within the next 4 years.⁴⁶ A woman over 45 whose difference between lengths of the longest and shortest cycles is >42 days has fewer than 20 cycles remaining before her FMP.⁴⁵

“Dysphoric mood” is a common complaint in the early menopause transition and is reported less frequently in the later phases.⁴⁷ Menstrual migraines peak during the transitional years,⁴⁸ as do often unrecognized hot flushes around menses.⁴⁹ In the early menopause, breast soreness-tenderness is significantly more common, but as the menopause transition progresses, more hypoestrogenic symptoms such as hot flashes, vaginal dryness, and night sweats occur with greater frequency.⁵⁰

Neuromuscular symptoms have been found to be stable throughout the menopausal transition, suggesting that other events in a midlife woman’s life (i.e., life stress and acute and chronic illnesses) account for the common complaint of joint pain.⁵¹

Hot flashes are considered to be a hallmark symptom of the menopausal transition.⁵² Hot flashes are commonly defined as transient periods of intense heat in the upper arms and face, which often are followed by flushing of the skin and profuse sweating.⁵³ Many hot flashes are followed by chills and can be accompanied by palpitations and a sense of anxiety.⁵³ Approximately 40% to 70% of menopausal women experience hot flashes, and 10% to 20% of these women obtain medical attention for treatment of their hot flashes.⁵⁴ Most midlife women will experience hot flashes for several months up to 5 years, but some will have hot flashes for up to 30 years.⁵³ Although the origin of hot flashes is not entirely clear, studies have suggested that changes in core body temperature regulation or changes in endogenous hormone levels or both are associated with the onset of hot flashes.⁵² Hot flashes negatively affect quality of life for women by causing sleep disturbances, which often result in fatigue, irritability, forgetfulness, acute physical discomfort, and negative effects on work.⁵⁵

Reproductive and Somatic Aging

Reproductive aging is confounded by somatic aging. All women do not age at the same rate. When interpreting menopausal data, we have to take the overall aging backdrop into account.

Menopause has been regarded historically as a primary ovarian event, with associated changes in pituitary gonadotropin secretion occurring secondary to the decline in ovarian sex steroid and protein production. However, increasing evidence suggests that aging is associated with dynamic changes in the hypothalamic and pituitary components of the reproductive axis that are independent of changes in gonadal hormone secretion.⁵⁶ Using gonadotropin free α subunit (FAS) as a surrogate marker of gonadotropin-releasing hormone (GnRH) pulse frequency in postmenopausal women, Hall et al. demonstrated a 35% decline in GnRH pulse frequency between the fifth and eighth decades with aging, providing evidence of slowing of hypothalamic GnRH pulse generator activity.⁵⁶ Other studies in which LH was used as a marker for GnRH secretion have shown variable results. Although estimates of pulse frequency were lower than in the previously stated study, Rossmanith et al. showed a decline in LH pulse frequency with age when comparing naturally postmenopausal women aged 49 to 57 years versus those 78 to 87 years.⁵⁷ Santoro et al. did not observe different patterns of LH secretion in postmenopausal women who were compared with young women with premature ovarian failure, but older women were much more readily suppressed by exogenous estradiol.⁵⁸ Whether these age-related changes in GnRH pulse frequency are due to lesser secretion of GnRH or to changes in pituitary responsiveness to GnRH is uncertain. Using prematurely menopausal women as controls allows separation of the process of ovarian failure from the hypothalamic-pituitary changes that accompany aging.

The somatopause begins prior to the menopause with aging. Twenty-four hour growth hormone (GH) secretion decreases in women before the onset of estrogen deficiency, and insulin-like growth factor-1 (IGF-1) levels in adults progressively decline with age.⁵⁹ Because IGF-1 is a co-factor for FSH action, additional gonadal compromise may be imposed by reduced somatotrophic axis function.⁶⁰ Body composition changes, including an increase in body fat in the visceral/abdominal compartment and a decrease in lean body mass, as well as adverse changes in lipoproteins, a relative increase in insulin resistance, and a reduction in aerobic capacity, are thought to be secondary to changes in the somatotrophic axis.⁶¹

As with other endocrine systems, aging is associated with significant changes in the anatomy of the thyroid gland and modifications in the physiology of the hypothalamic-pituitary-thyroid axis. Direct age-related changes should be distinguished from indirect alterations caused by other physiologic and pathologic states that occur with age. Subtle changes in the hypothalamo-pituitary-thyroid axis suggest decreased hypothalamic stimulation of thyroid function.⁶² The thyroid-stimulating hormone (TSH) nocturnal surge may be lost with increasing age, thus providing evidence for hypothalamic dysfunction.⁶³ Moreover, thyroid hormone clearance decreases with age, which explains the reduced daily replacement doses of thyroxine in hypothyroid elderly subjects.⁶⁴

With aging, minimal and mean cortisol plasma concentrations increase, with no alteration noted in pulsatile frequency, and the diurnal amplitude of cortisol and adrenocorticotropic hormone (ACTH) relative to the 24 hour mean of hormones shows an age-related decline.⁶⁵ In addition, the evening cortisol quiescent period is shortened in the elderly, suggesting increasingly impaired circadian function with aging.

In relation to bone and aging around the menopausal transition, SWAN reported that in premenopausal and early perimenopausal women, higher FSH concentrations were positively associated with greater bone turnover before the last menstrual period, as assessed by higher serum osteoclacin and urinary N-telopeptide of type I collagen concentrations.⁶⁶ SWAN also showed that serum FSH was inversely correlated with bone mineral density (BMD), and that BMD tended to be lower in women in the late perimenopause or early postmenopause versus the early perimenopause.⁶⁷ Finally, SWAN found that more sport and home physical activity, as compared with work and active living activity, was statistically significantly associated with greater BMD,⁶⁸ suggesting the importance of preserving activity in middle-aged women.

Cognitive function declines with age. Research studies in young, oophorectomized animals serve as classic endocrine ablation-replacement models to support a role for estrogen in decreasing cerebral ischemia and Alzheimer’s disease. Aging predicts cognitive decline, and leisure activity mitigates those effects. In a prospective, cohort study, Verghese et al. found that reading and playing board games were associated with reduced risks for dementia and Alzheimer’s disease.⁶⁹ In the Nun study, nuns who exhibited complex verbal skills early in life maintained cognitive function longer as they aged and were protected against the development of Alzheimer’s disease.⁷⁰ Although hormone therapy has been associated with decreased risk for dementia, neither the Women’s Health Initiative nor the Heart and Estrogen/Progestin Replacement Study (HERS) reported effects of hormone therapy on cognition and dementia.

The aging process is associated with predictable anatomic and physiologic alterations in the cardiovascular system. Cardiac mass increases approximately by 1 to 1.5 g yearly.⁷¹ A gradual increase in the circumference of all four cardiac valves is noted.⁷² Studies that have excluded subjects with hypertension show consistent increases in left ventricular wall thickness in conjunction with reduced diastolic compliance,^92,93 and the duration of contraction of the myometrium is prolonged. Histologic changes detected in the vasculature include increased intimal thickness, elastin fragmentation, and increased collagen content of the arterial wall.⁷³ Finally, the natural history of heart disease generally is adversely affected by age.⁷³ None of these changes are abruptly affected by menopause; all are gradual.

Diagnosis

Despite the epidemiologic trend toward elevated FSH and decreased estradiol with progression through the transition, measurement of FSH, inhibin, and estradiol provides at best an unreliable guide to the menopausal status of an individual woman.^96,97 A more rational approach to diagnosing menopause would include an assessment of the longitudinal symptoms of a woman who presents with perimenopausal complaints.⁷⁴ Hormone profiles correlate well with symptoms and cycle features.⁷⁵ Thus, if a woman is >45 years old and has had a recent disruption in her menstrual pattern and symptoms suggestive of transient hypoestrogenemia, it is likely that she has entered her menopausal transition.⁶⁰ That being said, the clinician should take care to rule out other pathologies that can be masked by common complaints associated with the menopausal transition. At minimum, a screening TSH level should be performed, as menstrual irregularity may be the only manifestation of thyroid dysfunction.

Women with a prior history of depression appear to be at excess risk for menopause-related depression^101,102; therefore screening for depression may be indicated. Three newer studies indicate an increased risk for new onset of depression associated with menopause.^76,77 In a 5 year follow-up of the SWAN study, women were significantly more likely to score > 16 on the Center for Epidemiologic Studies Depression (CES-D) scale for depressive symptoms when perimenopausal or postmenopausal than when premenopausal, and the odds ratio was found to be higher in the late than in the early menopause transition. In addition, women with CES-D scores < 16 at baseline were more likely to have scores > 16 in the later perimenopausal or postmenopausal period, highlighting our incomplete understanding of the link between menopause and depression. Chinese women had a 50% lower chance of having CES-D scores > 16, suggesting possible ethnic variability of the association between menopause and depression.⁷⁸ Many women experience a loss of “wellness” and dysphoric mood during the transition; these syndromes should be formally distinguished from the more severe and life-threatening risk that clinical depression may pose.

Atypical hot flushes can be due to non–estrogen-related causes, including new-onset diabetes with autonomic dysfunction, carcinoid, and pheochromocytoma.⁴⁹ Poor sleep can result from non–estrogen-related causes such as sleep apnea. Women who appear to be entering the menopausal transition earlier than age 40 years should have a more extensive workup, including screening appropriate for patients with premature ovarian failure.⁷⁹