Secondary Logo

Share this article on:

A Prospective Population-Based Study of Menopausal Symptoms



Objective To identify symptoms that change in prevalence and severity during midlife and evaluate their relationships to menopausal status, hormonal levels, and other factors.

Methods In a longitudinal, population-based study of 438 Australian-born women observed for 7 years with an 89% retention rate, 172 advanced from premenopause to perimenopause or postmenopause. Annual measures included a 33-item symptom check list; psychosocial, lifestyle, and health-related factors; menstrual status; hormone usage; and blood levels of follicle-stimulating hormone and estradiol (E2).

Results Increasing from early to late perimenopause were the number of women who reported five or more symptoms (+14%), hot flushes (+27%), night sweats (+17%) and vaginal dryness (+17%) (all P < .05). Breast soreness-tenderness decreased with the menopausal transition (−21%). Trouble sleeping increased by +6%. The major change in prevalence was from early to late perimenopause, except for insomnia, which showed a gradual increase. Those variables most related to onset of hot flushes were number of symptoms at early perimenopause (P < .05), having an unskilled or no occupation (P < .05), more than 10 pack-years of smoking (P < .01), and decreased E2 (P < .01). The onset of night sweats increased with the change in E2 (P < .05). The onset of vaginal dryness decreased with more years of education (P < .05). Trouble sleeping was predicted by prior lower well-being (P < .01), belief at baseline that women with many interests hardly notice menopause (P < .01), and hot flushes (P < .01).

Conclusion Although middle-aged women are highly symptomatic, the symptoms that appear to be specifically related to hormonal changes of menopausal transition are vasomotor symptoms, vaginal dryness, and breast tenderness. Insomnia reflected bothersome hot flushes and psychosocial factors.

Hot flushes, night sweats, vaginal dryness, and breast tenderness are directly related to the menopausal transition and underlying hormonal changes.

Office for Gender and Health, Department of Psychiatry; Centre for Genetic Epidemiology, The University of Melbourne, Parkville, Australia; and the Prince Henry's Institute for Medical Research, Monash Medical Centre, Clayton, Victoria, Australia.

Lorraine Dennerstein, MBBS, PhD Office for Gender and Health Department of Psychiatry University of Melbourne Royal Melbourne Hospital, Charles Connibere Building Parkville, Vic 3050 Australia. E-mail:

Funding sources: Public Health Research Project Grant from the National Health and Medical Research Council, Victorian Health Promotion Foundation, Australasian Menopause Society. Prince Henry's Institute for Medical Research received a grant from Organon (Aust) Pty Ltd to measure the hormone levels.

Received January 3, 2000, Received in revised form April 3, 2000, Accepted April 20, 2000.

The midlife years of 45–55 coincide with those of menopausal transition, and studies of middle-aged women suggest that they are highly symptomatic.1,2 Much research effort has gone into trying to identify which symptoms are related to hormonal changes underlying menopause and which are caused by aging, other health states, psychosocial factors, or lifestyle. In 1972, Utian3 determined the “true clinical features of postmenopause” by comparing prevalences of symptoms between women who had hysterectomies with and without ovarian removal, and studying the alleviation of symptoms when oral estrogen was given in a single blind crossover trial. The only symptoms directly associated with oophorectomy and oral estrogen alleviation were vasomotor symptoms (hot flushes and perspiration) and those consistent with atrophic vaginitis (dyspareunia, blood-stained vaginal discharge). Randomized controlled trials (RCTs) of hormone replacement therapies (HRT)4,5 followed, and larger ones are still in progress; however, study samples often included women who had surgical menopause, and the estrogens administered were usually synthetic. Thus, RCTs have not informed us of the relationship of symptoms to natural menopausal transition. Although very suggestive of which symptoms are linked to hormonal change, the RCT approach is complemented by findings of prospective, population-based observational studies of the natural menopausal transition.

The main methodologic issues that limit inferences from epidemiologic studies of the menopausal transition include lack of longitudinal studies that use hormone measures6–8 and short follow-up.7–9 For studies of changes in symptom prevalence with menopause, the major issue is validity and reliability of symptom measures. The standard measure used has been a checklist of symptoms, but that introduces the problem of elicitation, with more positive responses to a checklist than open-ended questions.10 When frequency or severity of complaints is included, prevalence of reported symptoms is typically reduced.11

The Melbourne Women's Midlife Health Project2 used a number of strategies to overcome those problems in study design, including population-based samples; asking about bothersome symptoms in the previous 2 weeks to minimize recall bias; longer follow-up (7 years from baseline); annual blood samples for measuring hormone levels; and collecting annual data on menstrual status and hormone usage to determine phase of natural menopause. We reported on the factorial validity of the symptom checklist2 with the use of baseline interview data. The present analyses use data drawn from baseline interviews and seven subsequent annual follow-up interviews. The aim was to determine which symptoms change in prevalence and severity during midlife and which factors best explain changes in those symptoms.

Back to Top | Article Outline

Materials and Methods

The study was approved by the Human Research Ethics Committee of The University of Melbourne, and procedures were in accordance with the ethical standards of the National Health and Medical Research Council. Subjects provided written informed consent for participation.

The study began in 1991 with the use of random telephone digital dialing to baseline interview 2001 Australian-born women between 45 and 55 years old and residing in Melbourne (71% response rate).2 Women at baseline who had menses in the prior 3 months and who were not taking oral contraceptives (OCs) or hormone therapy were invited to participate in a longitudinal study, and 438 (56%) among those eligible agreed. Volunteers for the longitudinal study were more likely than eligible nonparticipants to report better self-rated health, employment, more than 12 years of education, having ever had Papanicolaou smears, exercising at least once a week, and having ever had D&C12 (Table 1). The retention rate by the end of year 7 of follow-up was 89% (n = 387).

Table 1

Table 1

The current report is based on 172 women who were premenopausal at baseline and, by the end of the seventh year of annual follow-up, had advanced to perimenopause or postmenopause. If women had surgical menopause through hysterectomy, bilateral oophorectomy, or endometrial ablation (n = 10) or took HRT during the study (n = 53), only observations before that medical intervention were used.

We used questionnaires developed by Kaufert and Syrotuik13 and McKinlay et al6 for studies in North America. Additional references are given for specific variables and those questions or instruments that supplemented the core questionnaire. Once each year, follow-up questionnaires were given and physical measures and blood samples were taken in women's homes by fieldworkers. Variables selected for analysis were related to symptom reporting in the baseline cross-sectional study2 or identified as important by review of the literature.14

To the North American symptom checklist14 we added shortness of breath on exertion, chest pain on exertion, vaginal dryness in the second year of follow-up, and breast soreness-tenderness in the third year of follow-up. Women were asked whether they had been bothered by each symptom in the last 2 weeks. Each symptom was scored on a four-point scale in which 0 = not present, 1 = symptom causes minor irritation, 2 = symptom interferes with normal life, 3 = symptom is debilitating.

An effect of order within the symptom list was apparent by moving the symptom “lack of energy” from second position in year 4, to 28th position in year 5, which caused a sharp drop (−12.9%) in the proportion of women who reported lack of energy and was more than four standard deviations (SDs) outside the mean change in other symptoms ( = + 0.31, SD = 3.09); P < .001). Other symptoms were always presented in the order shown in Table 2.

Table 2

Table 2

Variables from baseline telephone interviews included sociodemographic factors (years of education, type of occupation); premenstrual complaints2; attitudes about aging13; attitudes about menopause15; and health status (chronic conditions, prescription and non-prescription medications).2

Variables at annual assessments included sociodemographic factors (age, marital status, presence or absence of children in the home, work status); well-being (Affectometer-216); self-rated health17; lifestyle factors (exercise, smoking, and alcohol consumption); daily hassles18,19 and interpersonal stress2; feelings for partner (factor 1 of the Personal Experiences Questionnaire20); and satisfaction with work. Annual assessments also recorded menopausal status based on women's self-reports of perceived changes in menstruation. Women were asked to compare regularity of their current menstrual cycles to that in the 12 months before.21 Premenopausal status was assigned to women who reported no changes in menstrual frequency. Early perimenopausal status was assigned to women who reported changes in menstrual frequency but had menstruated in the previous 3 months. Late perimenopausal status was assigned when there had been 3 to 11 months of amenorrhea.21 Women were assigned to postmenopausal status when there had been amenorrhea for at least 12 months. Late perimenopause and postmenopause were verified by fieldworkers who viewed women's prospective menstrual calendars. Women who had induced amenorrhea or were taking HRT were excluded from those analyses.

Blood samples were collected annually between days 4 and 8 of menstrual cycles for those still cycling, or after 3 months of amenorrhea. Follicle-stimulating hormone and estradiol (E2) were measured by radioimmunoassays as described.12

The means of symptom severity scores (μ1) over the premenopausal and early perimenopausal years were compared with those over the late perimenopausal and postmenopausal years by using paired t tests. Differences were considered nominally significant at P < .05 and are illustrated by 95% confidence intervals (CIs) not crossing the 0 line in Figure 1. Menopausal categories were dichotomized because previous analyses identified late perimenopause as the phase of maximum change in levels of follicle-stimulating hormone and E222 (Figure 2).

Figure 1

Figure 1

Figure 2

Figure 2

Logistic regression was used to determine predictors of symptoms in late perimenopause. Women who reported symptoms in early perimenopause were excluded. Explanatory variables associated with developing symptoms at the P = .10 level of significance were retained for model building. A backward stepwise elimination method based on the likelihood ratio test and the significance level of P = .05 was used to determine a parsimonious model. Levels of follicle-stimulating hormone and E2 were log-transformed to reduce the influence of extreme points. Breast soreness-tenderness was not included because it was not introduced until the third year of follow-up.

Back to Top | Article Outline


The proportion of women who reported a particular symptom as bothersome in the previous 2 weeks for each category of menopausal status is given in Table 2.

By postmenopause, almost all women were reporting at least one symptom, and most were reporting five or more. The average total number of symptoms increased by 17%, from 4.2 to 4.9 (P < .001).

Figure 1 summarizes changes; 0 indicates no change. Severity scores for breast soreness were reduced in late perimenopause and postmenopause compared with breast soreness scores before and in early perimenopause (P < .001). Symptom severity of trouble sleeping (P < .05), vaginal dryness (P < .001), night sweats (P < .01), and hot flushes (P < .001) were all increased in late perimenopause and postmenopause. No other symptom showed any significant change with menopausal status at P < .05.

Sample proportions of symptom reporting were calculated for each of the six categories of menopausal status (Table 2). The magnitude between categories was assessed. The transition of most change was clearly between early and late perimenopause for hot flushes (+27%), night sweats (+17%), vaginal dryness (+17%), and breast soreness-tenderness (−21%). The symptom of trouble sleeping appears to follow a more gradual increase across menopausal categories (+6% between early and late perimenopause).

The number of women who reported five or more symptoms increased by 14% between early and late perimenopause (P < .05). Table 3 lists variables related to hot flushes, night sweats, vaginal dryness and trouble sleeping in late perimenopause in women free of those symptoms in early perimenopause.

Table 3

Table 3

Back to Top | Article Outline


The findings of the present study were consistent with those of previous studies that middle-aged women were very symptomatic according to a checklist. Vasomotor symptoms were not the most frequently reported until 3 years after final menstrual periods. Our findings were similar to those of Avis et al,1 who reported the percentages of Massachusetts women who had five or more core symptoms as 23% for premenopausal women, 35% for perimenopausal women, and 32% for naturally postmenopausal women. Our figures for the same groups using the same set of symptoms were 22%, 22%, and 36%, respectively. In this Australian study, only a few women reported no or only one symptom, and that declined as women went through the transition, whereas the number of symptoms reported increased by 17%.

There were some limitations in our study design, which might have caused underestimation of symptom prevalence for this population. Volunteers for the longitudinal phase of the study differed at baseline from nonparticipants in certain variables (self-rated health, education level, work, exercise participation) that were shown in our own baseline cross-sectional analysis to be related to symptom reporting.2 During the longitudinal study, there also was attrition related to women taking HRT. In the second year of follow-up, we asked about reasons for HRT use among 108 of 438 women who began the longitudinal study, who were then current users of HRT. The most common reasons were hot flushes, sweats (45%); moods, depression, anxiety (16%); headaches (4%); prevention of osteoporosis and heart disease (22%), bleeding problems (7%); no reason or non-classified reason (6%). Thus, most who took HRT did so because of bothersome symptoms, which would be expected to lower the point prevalence of symptoms, but should not affect the aim of this paper, which was to measure change in symptoms during the natural menopausal transition.

The prevalence of symptom reporting might be related to the order in which symptoms were presented in the checklist. Four of nine symptoms reported as bothersome by 30% or more of women were among the first nine symptoms presented in the checklist, and their high prevalence might indicate early position on the checklist, which should be considered when comparing prevalence rates from different studies. The position of all symptoms except lack of energy was held constant during the baseline and 7 subsequent years, so we were able to detect reliably change in prevalence of symptom reporting, our aim.

One hundred seventy-two women (all premenopausal at baseline) contributed to the analysis, but not all supplied data points for each menopausal phase, as shown in Table 2. That does not indicate subject attrition but does show that it can take many years to observe women through menopausal transition.

In the current study, fieldworkers verified women's self-reports of 3 months' or more amenorrhea or of having reached final menstrual period against menstrual calendars recorded daily. Late perimenopause and date of final menstrual period from which postmenopause was calculated were ratified objectively. There was no such ratification of women's reports of whether or not there had been changes in their menstrual frequency over the prior 12 months. Our earlier21 and current results suggest that premenopause and early perimenopause can be combined.

Regarding which symptoms truly relate to endocrine change rather than aging, this study found that only vasomotor symptoms, breast tenderness, and vaginal dryness changed significantly with menopause. Many studies also noted a marked temporal relationship between vasomotor symptoms and menopause.23 We confirmed other studies that found vasomotor symptoms significantly elevated in late perimenopause and remained elevated for some years.11,24–26 Several studies, including our earlier factor analysis,2 found an association between hot flushes and night sweats,27 and some an association between those vasomotor symptoms and insomnia.28

The present study confirmed that hot flushes contribute to reported trouble sleeping, which appeared to be related to depression (low well-being) or busy lifestyle. The pattern of change seen for trouble sleeping was different from other symptoms that changed with menopause, showing a small, gradual change rather than a large incremental one. This suggesting that the change in trouble sleeping was not a direct effect of hormonal changes of menopause.

We found a five-fold increase in the prevalence of vaginal dryness as women advanced through menopause (see also Oldenhave and Jaszmann25). Despite that, most observational studies neglected to include that symptom in their checklists. Breast soreness-tenderness (mastalgia), the prevalence of which we found decreased by more than half, was also absent from most checklists in population-based studies of menopause. It is not clear why those two important symptoms were omitted because both have been known for many years to be related to menopause (Neugarten and Kraines29) and in clinical trials were related to estrogen-progestin levels.5,30 We suggest those symptoms should be included in future studies.

Population-based studies have been less consistent on the relationship of psychologic symptoms to menopause. Some cross-sectional studies24,25 reported small but transient increases in non–vasomotor symptoms in perimenopause. We found increased symptoms such as nervous tension and sadness in late perimenopause, but the increase was not statistically significant. When validated mood scales were used in longitudinal studies,7,19 no significant increase in depression or negative moods was directly associated with menopause.

On the effect of physical, psychosocial, and lifestyle factors in those symptoms related to menopause, cross-sectional studies explored associations between symptom experience and a wide range of other factors. Cross-sectional studies that incorporated hormone levels found vasomotor symptoms associated with decreased E231,21 and increased follicle-stimulating hormone.31

Longitudinal population-based studies are best able to establish likely relationships between symptoms and physical, psychosocial, and lifestyle factors. The present study is one of few longitudinal population-based studies to include hormone levels. We found that E2 was the best predictor for vasomotor symptoms. The development of hot flushes in late perimenopause was increased for lower levels of E2. A particularly high odds ratio (OR) (20.5) was observed for the middle E2 category, 30–100 pmol/L; however, the OR was not significantly greater than the last category, less than 30 pmol/L. One interpretation is that hot flushes are precipitated by falling levels of E2, so are closely related to the phase of late perimenopause when E2 levels are changing. Length of smoking (in pack-years) contributed to vasomotor symptoms of hot flushes, and smoking has adverse effects on ovarian function.33 Although increased vaginal dryness was associated with menopause, logistic regression did not find a relationship between it and hormone levels, and it showed a different pattern in relation to menopause. Vaginal dryness was the only symptom that appeared to increase exponentially with time from the late perimenopause, indicating that it was a later consequence of hormonal changes during menopause. Women with tertiary education were less likely to report that symptom, suggesting that increased knowledge might be related to factors that helped women with it maintain sexual function.

Back to Top | Article Outline


1. Avis NE, Kaufert PA, Lock M, McKinlay S, Vass K. The evolution of menopausal symptoms. Baillieres Clin Endocrinol Metab 1993; 7:17–32.
2. Dennerstein L, Smith AMA, Morse CA, Burger H, Green A, Hopper JL, et al. Menopausal symptoms in Australian women. Med J Aust 1993;159:232–6.
3. Utian W. The true clinical features of postmenopause and oophorectomy, and their response to oestrogen therapy. S Afr Med J 1972;46:732–7.
4. Campbell S, Whitehead M. Oestrogen therapy and the menopausal syndrome. Clin Obstet Gynecol 1977;4:31–47.
5. Greendale GA, Reboussin BA, Hogan P, Barnabel VM, Shumaker S, Johnson S, et al. Symptom relief and side effects of postmenopausal hormones: Results from the postmenopausal estrogen/progestin interventions trial. Obstet Gynecol 1998;92:982–8.
6. McKinlay J, McKinlay S, Brambilla D. The relative contributions of endocrine changes and social circumstances to depression in mid-aged women. J Health Soc Behav 1987;28:345–3.
7. Kaufert P, Gilbert P, Tate R. The Manitoba Project: A reexamination of the link between menopause and depression. Maturitas 1992;14:143–55.
8. Bromberger J, Matthews K. A longitudinal study of the effects of pessimism, trait anxiety, and life stress on depressive symptoms in middle-aged women. Psychol Aging 1996;11:207–13.
9. Woods N, Mitchell E. Patterns of depressed mood in midlife women: Observations from the Seattle Midlife Women's Health Study. Res Nurs Health 1996;19:111–23.
10. Wright A. On the calculation of climacteric symptoms. Maturitas 1981;3:55–63.
11. Holte A. Prevalence of climacteric complaints in a representative sample of middle-aged women in Oslo, Norway. J Psychosom Obstet Gynaecol 1991;12:303–17.
12. Burger H, Dudley E, Hopper J, Shelley J, Green A, Smith AMA, et al. The endocrinology of the menopausal transition: A cross-sectional study of a population based sample. J Clin Endocrinol Metab 1995;80:3537–45.
13. Kaufert P, Syrotuik J. Symptom reporting at the menopause. Soc Sci Med 1981;15:173–84.
14. Dennerstein L. Well-being, symptoms and the menopausal transition. Maturitas 1996;23:147–57.
15. Neugarten B, Wood V, Kraines R, Loomis B. Women's attitudes toward the menopause. Vita Humana 1963;6:140–51.
16. Kammann R, Flett R. Affectometer 2: A scale to measure current level of general happiness. Aust J Psychol 1983;35:259–65.
17. Smith AMA, Shelley JM, Dennerstein L. Self-rated health: Biological continuum or social discontinuity? Soc Sci Med 1994;39:77–83.
18. Kanner A, Coyne J, Schaefer C, Lazarus R. Comparison of two modes of stress measurement: Daily hassles and uplifts versus major life events. J Behav Med 1981;4:1–39.
19. Dennerstein L, Lehert P, Burger H, Dudley E. Mood and the menopausal transition. J Nerv Ment Dis 1999;187:685–91.
20. Dennerstein L, Dudley E, Hopper J, Burger H. Sexuality, hormones and the menopausal transition. Maturitas 1997;26:83–93.
21. Dudley E, Hopper J, Taffe J, Guthrie J, Burger H, Dennerstein L. Using longitudinal data to define the perimenopause by menstrual cycle characteristics. Climacteric 1998;1:18–25.
22. Burger HG, Dudley E, Groome N, Guthrie J, Hopper J, Green A, et al. Prospectively measured levels of serum follicle-stimulating hormone, estradiol and the dimeric inhibins during the menopausal transition in a population-based cohort of women. J Clin Endocrinol Metab 1999;84:4025–30.
23. Greene JG. Constructing a standard climacteric scale. Maturitas 1998;29:25–31.
24. Jaszmann L, van Lith ND, Zaat JCA. The perimenopausal symptoms: The statistical analysis of a survey. Part A. Med Gynecol Soc 1969;4:268–77.
25. Oldenhave A, Jaszmann L. The climacteric: Absence or presence of hot flushes and their relation to other complaints. In: Schonbaum E, ed. The climacteric hot flush. Prog Basic Clin Pharmacol, Basel Karger, 1991;6:6–39.
26. McKinlay S, Brambilla D, Posner JG. The normal menopause transition. Maturitas 1992;14:103–15.
27. McKinlay S, Jeffreys M. The menopausal syndrome. Br J Prev Soc Med 1974;28:108–15.
28. Thompson B, Hart SA, Durno D. Menopausal age and symptomatology in a general practice. J Biosoc Sci 1973;5:71–82.
29. Neugarten B, Kraines J. “Menopausal symptoms” in women of various ages. Psychosom Med 1965;27:266–73.
30. de Lignieres B, Vincens M. Differential effects of exogenous oestradiol and progesterone on mood in post-menopausal women: Individual dose/effect relationship. Maturitas 1982;4:67–72.
31. Guthrie J, Dennerstein L, Burger H, Hopper J. Hot flushes, menstrual status and hormone levels in a population-based sample of midlife women. Obstet Gynecol 1996;88:437–42.
32. Wilbur J, Miller A, Montgomery A, Chandler P. Sociodemographic characteristics, biological factors, and symptom reporting in midlife women. Menopause 1998;5:43–51.
33. Cramer DW, Harlow BL, Xu H, Fraer C, Barbieri R. Cross-sectional case controlled analyses of the association between smoking and early menopause. Maturitas 1995;22:79–87.

Cited By:

This article has been cited 6 time(s).

Obstetrics & Gynecology
Premenstrual Syndrome as a Predictor of Menopausal Symptoms
Freeman, EW; Sammel, MD; Rinaudo, PJ; Sheng, L
Obstetrics & Gynecology, 103(5): 960-966.
PDF (216) | CrossRef
Obstetrics & Gynecology
Gabapentin, Estrogen, and Placebo for Treating Hot Flushes: A Randomized Controlled Trial
Reddy, SY; Warner, H; Guttuso, T; Messing, S; DiGrazio, W; Thornburg, L; Guzick, DS
Obstetrics & Gynecology, 108(1): 41-48.
PDF (290) | CrossRef
Obstetrics & Gynecology
Phytoestrogens for Treatment of Menopausal Symptoms: A Systematic Review
Krebs, EE; Ensrud, KE; MacDonald, R; Wilt, TJ
Obstetrics & Gynecology, 104(4): 824-836.
PDF (361) | CrossRef
Obstetrics & Gynecology
Association of Change in Estradiol to Lower Urinary Tract Symptoms During the Menopausal Transition
Gopal, M; Sammel, MD; Arya, LA; Freeman, EW; Lin, H; Gracia, C
Obstetrics & Gynecology, 112(5): 1045-1052.
PDF (228) | CrossRef
Obstetrics & Gynecology
Accuracy of Subjective Hot Flush Reports Compared With Continuous Sternal Skin Conductance Monitoring
Carpenter, JS; Monahan, PO; Azzouz, F
Obstetrics & Gynecology, 104(6): 1322-1326.
PDF (157) | CrossRef
Obstetrics & Gynecology
Eszopiclone in Patients With Insomnia During Perimenopause and Early Postmenopause: A Randomized Controlled Trial
Soares, CN; Joffe, H; Rubens, R; Caron, J; Roth, T; Cohen, L
Obstetrics & Gynecology, 108(6): 1402-1410.
PDF (303) | CrossRef
Back to Top | Article Outline
© 2000 The American College of Obstetricians and Gynecologists