Special Feature

Understanding the Perimenopause

By Sarah L. Berga, MD

The perimenopause is a time of confusion. bleeding patterns become erratic. Hormone levels fluctuate unpredictably. Symptoms come and go. To make matters worse, the physician is often at a loss for effective therapies and may dismiss the symptom complex because of diagnostic or therapeutic uncertainty. The way out of this limbo, for both patients and physicians, is to have a clear understanding of what the peri-menopausal transition entails.

Perimenopause refers to those years in the reproductive life cycle during which ovarian cyclicity becomes irregular because of a decline in oocyte competence and count. Generally, this phase occurs during the years from ages 40-55.1 The reproductive life cycle is characterized by a progression of states that involve the interplay between oocyte competence and hypothalamic-pituitary functioning. In a sense, the perimenopause is the "mirror image" of puberty. The timing and tempo of puberty is primarily a function of hypothalamic events and the reactivation of GnRH pulsatility, but, for puberty to be clinically manifested, the ovaries must contain responsive follicles. In contrast, the timing and tempo of perimenopause reflects follicular availability and responsivity, but, for the perimenopause to be clinically apparent, the hypothalamic-pituitary unit must provide gonadotropin drive.

The cause of perimenopause and menopause is progressive follicular depletion. The store of primordial follicles is fixed in utero, and, thereafter, the number of primordial follicles declines with age. Across each interval of time, a certain percentage of these resting follicles die or become atretic. It has been estimated that menopause occurs when the number of primordial follicles falls below 1000.2 The usual age for this degree of follicular depletion is 51 years. The age of menopause is not thought to be influenced by socioeconomic factors, but rather, it has been found to be a relatively universal biological constant.3 It is estimated that the number of primordial follicles at birth is about 1 million and that the rate of atresia is relatively constant until the remaining follicles number about 25,000. The customary age at which the rate of follicular atresia accelerates is estimated to be 37.5 years.2 The probability of being menopausal increases with duration of amenorrhea and age.4

Considerable evidence suggests that serum FSH and serum inhibin levels are biomarkers for the number or quality of remaining follicles.4 However, in the perimenopause, each menstrual cycle is a relatively independent event,5 so measuring a single, isolated FSH level does not help to predict when menopause will occur and an untimed FSH level is a poor discriminator of follicular reserve. It remains to be determined if inhibin levels, which also change with follicular development, will predict ovarian reserve. While the biological basis for changing rates of atresia is not known, there is some empirical support for the notion that elevated FSH accelerates the rate of atresia.6 This might be one reason that rates of atresia accelerate in the perimenopause. According to one analysis, smokers tend to reach menopause an average of 1.74 years earlier than nonsmokers.7 However, family history is a good predictor of early menopause. In one study, women with a family history of early menopause were not more likely to have inborn errors of galactose metabolism or stigmata of Turner’s syndrome, but they were less likely to have brothers.8 The strength of this familial association is such that a history of menopause before age 46 in a mother or sister increases the probability that a woman will have her menopause before age 46 years from 5-25%. Given that the cause of ovarian failure in girls with Turner’s syndrome and those with deletions of the long arm of the X chromosome is thought to be an accelerated rate of atresia, Cramer and associates suggested that the low incidence of male siblings suggested that microdeletions of the long arm of the X chromosome might account for these observations.8 The exact gene product encoded for at the distal end of the long arm of the X chromosome is unknown, but it presumably influences oocyte longevity, and, thus, it would be potentially useful to identify. More information about the molecular events regulating follicular apoptosis also is expected to aid our understanding of the "ovarian clock."9

Clinical Consequences

One of the main reasons for wanting to understand the timing of menopause relates to concerns about predicting the end of fertility. If we could estimate the remaining follicle reserve, we might be able to individually forecast fecundity and menopause. This might be helpful to those who were interested in conceiving as well as to those seeking to avoid fertility. Although the mean age of sterility is estimated to be 44 years and is relatively constant worldwide,3 it is difficult to tell an individual woman when she can stop using contraception or when she can expect a fertility procedure to result in a conception. Maintaining or extending follicular reserve might also be an objective some would desire. Understanding and reversing gonadal dysgenesis might be an outcome that could be achieved if the biology of follicular atresia was better understood. Some experts believe that the ovarian clock reflects the individual’s biological as opposed to chronological age and that menopause is a marker for the aging process in a given individual.4 Forestalling aging might seem to be an unrealistic goal, but extending disease-free life may not be. Aging confers vulnerability to a wide variety of disorders that challenge individuals and society; thus, our forays into predicting or retarding the aging process could have multiple dividends.

Another reason for needing to understand the physiology of the menopause is that the changes in ovarian function are associated with a number of clinically apparent symptoms. Santoro and colleagues have shown that perimenopausal cycles are characterized by hyper-estrogenism and luteal phase insufficiency.5 These alterations in steroidal secretion may lead to endometrial proliferation and menorrhagia. Endometrial sloughing depends on exposure to an orderly sequence of hormonal excursions. Should these not occur, the endometrium may not slough in an orderly fashion and the usual hemostatic mechanisms, particularly the release of prostaglandins and subsequent provocation of uterine "contractions" may be blunted, leading to arteriole dilation. Relative hyperestrogenism may provoke fibroid growth in predisposed individuals, moodiness in those so prone, and mastodynia. Because each menstrual cycle in the perimenopause is an independent event, cycles are irregularly irregular and associated symptoms may come and go. Lack of predictability alone may provoke concern and annoyance, fear, or frustration. It is even more maddening for women seeking medical consultation for these symptoms to be subjected to a dismissive attitude that abrogates their own sense of observation and concern. Thus, the prudent clinician is advised to heed the woman’s own description of the changes she notes and help her to understand that they likely reflect erratic ovarian function that is physiologic. The physician’s role is also to determine when these symptoms herald organic disorders and warrant further assessment and treatment.

Treatment Strategies

Clinically, the major problem is recognizing the perimenopause. Symptoms may be episodic because ovarian function waxes and wanes in an unpredictable fashion. Therefore, management must be based on the symptom complex. While many symptoms may be attributable to erratic ovarian function and amplified hormonal fluctuations, care must be taken not to prematurely assign all symptoms to this physiological event. Organic conditions must be excluded. In short, this is a good time to take a health inventory and to make sure that the perimenopausal woman is up-to-date on health screening. If the past personal or family history indicates problem areas, screening can be tailored or diagnostic studies undertaken.

A common clinical problem is heavy menses at unpredictable intervals (i.e., poor endogenous cycle control). Once conditions such as cancer, polyps, and fibroids have been excluded, the optimal solution for unpredictable ovarian function is the use of low-dose oral contraceptive pills containing 20 mcg of ethinyl estradiol. However, oral contraceptive use must be confined to nonsmokers and women without risk factors for or a history of thromboembolic events. The progestin dominance of ultra-low-dose oral contraceptives may help to reverse any nascent endometrial hyperplasia and because the dose of steroids in oral contraceptives is sufficient to suppress central GnRH drive, endogenous ovarian sex steroid secretion will also be suppressed. Standard hormone replacement regimens used for menopausal women also may have use, but given the propensity of the perimenopausal ovary to oversecrete estrogen, care must be taken to provide sufficient progestin along with estrogen exposure to avoid further stimulating the endometrium and increasing the chance of endometrial hyperplasia. Also, standard HRT regimens may not confer adequate cycle control, because the doses are insufficient to adequately suppress hypothalamic GnRH drive, and, thus, the endometrium may be exposed to both endogenous and exogenous steroid excursions. In perimenopausal women with recent onset of depression, studies suggest that antidepressant therapy has minimal efficacy if estrogen levels are inadequate, so institution of hormonal therapy is suggested prior to a trial of antidepressants.10 Perimenopausal ovarian function may provide high estrogen levels at some times followed by very low levels after the demise of a dominant follicle and before the recruitment of another. Therefore, the goal of HRT in the perimenopause is to provide "background" sex steroid levels upon which endogenous ovarian steroids fluctuate. Because progestins inhibit GnRH drive, there also may be some slight intermittent suppression of central input and possibly some blunting of ovarian hypersecretion of estrogen. Women with low bone mass can receive concurrent therapy with bisphosphonates, if needed, although low-dose oral contraceptives tend to promote bone accretion.


Understanding the physiology of the perimenopause helps the patient and clinician to put subtle clinical symptoms, such as mood swings, intermittent hot flashes, and heavy, irregular bleeding, into context. Conditions such as uterine pathology and hypothyroidism need to be excluded, however. Relative hyperestrogenism and luteal phase insufficiency characterize perimenopausal ovarian function, although there also may be windows of relative hypoestrogenism between the demise and the recruitment of poorly responsive follicles. Menometrorrhagia due to hormonal aberrations may respond to hormonal interventions, particularly oral contraceptives containing 20 mcg of ethinyl estradiol. The erratic nature of perimenopausal ovarian function means that the patient’s history is one of the best diagnostic tools available and prudent clinicians are advised to elicit from the patient a prioritization as to which symptoms require intervention vs. reassurance. This is a good time to assess health risks and behaviors and to provide well-care screening.


1. Treloar AE, et al. Int J Fertil 1967;12:77-126.

2. Faddy MJ, et al. Hum Reprod 1992;7:1342-1346.

3. Rahman O, Menken J. Chapter 5 in Biomedical and Demographic Determinants of Reproduction. Graywith R, Leridon H, Spira A, eds. Oxford, GB: Clarendon Press; 1993:63-84.

4. Richardson SJ. Baillieres Clin Endocrinol Metab 1993;7:1-16.

5. Santoro N, et al. J Clin Endocrinol Metab 1996;81:1495-1501.

6. Nelson JF, et al. Neurobiol Aging 1995;16:837-843.

7. McKinlay SM, et al. Ann Intern Med 1985;103:350-356.

8. Cramer DW, et al. Fertil Steril 1995;64:740-745.

9. Hsueh AJ, et al. Recent Prog Horm Res 1996;51:433-455.

10. Schneider LS, Fluoxetine Collaborative Study Group. Am J Geriatr Psychiatry 1997;5:97-106.