Racial and Ethnic Differences in the Physiology and Clinical Symptoms of Menopause

Gloria Richard-Davis, MD, FACOG; Melissa Wellons, MD


Semin Reprod Med. 2013;31(5):380-386. 

In This Article

Natural Menopause—Biological Differences

Racial Difference in the Timing of Natural Menopause

Knowledge of the timing of menopause is important for many reasons. The timing of this key reproductive milestone has implications for fertility and family planning. It also has chronic health implications, as it is associated with diseases such as coronary heart disease (associated with early menopause) and breast cancer (associated with later menopause).

Most of our knowledge of the timing of menopause has come from studies of white women (≥ 95% white) from industrialized countries. In these studies, the median age of menopause is reported as between 50 and 52.[1–4] Obviously, these studies lack ethnic and racial diversity. A further limitation of these studies is the frequent use of the cross-sectional survey design. This design lends itself to bias, since women are often asked to recall their age at menopause at a time many years distant from when menopause occurred. Therefore, ethnically diverse, longitudinal studies of reproductive aging are needed. To date, longitudinal studies of the menopause transition with racial/ethnic diversity include the Healthy Women Study (HWS), the Study of Women Across the Nation (SWAN) and the Pennsylvania Ovarian Aging Study (POAS).

Healthy Women Study. The HWS provided intriguing data suggesting that race/ethnicity may serve as an important factor with regard to the timing of menopause. HWS began recruiting in 1983 and recruited 541 women between the ages of 42 and 50 years from Allegany County, PA. Of these 541 women, 48 (9%) were African American. HWS's purpose was to investigate the biologic and behavioral risk factor changes in healthy premenopausal women as they subsequently experience menopause. One study regarding this longitudinal cohort included only women who were 48 or younger at baseline (n = 185) and determined their age at menopause as reported 7 to 9 years after study entry. In this group, the average age of menopause was 51.5 years. HWS researchers found that African Americans had an average age of menopause of 48.4 years. Thus, African American race was found to be a significant determinant of the timing of menopause in this small study.[5]

Study of Women Across the Nation. The racial differences found in HWS likely influenced the development of the longitudinal SWAN. Begun in 1996, SWAN seeks to describe the chronology of the biological and psychological characteristics of menopausal transition in a much larger and more ethnically diverse US population of women. SWAN's goal is to determine the effect of this transition on subsequent health and risk factors for age-related chronic disease.

SWAN recruited women from five racial/ethnic groups (African American, Caucasian, Chinese, Hispanic, and Japanese). SWAN recruited women from seven sites across the United States. Its recruitment process involved two stages: a cross-sectional survey that engaged more than 16,000 women ages 40 to 55 and a second stage that recruited 3,000 racially/ethnically diverse women for the longitudinal study (women ages 42–52).[6]

Cross-Sectional Survey Findings. In the cross-sectional survey, the median age at menopause was 51.4 as determined by multivariable Cox-proportional hazards models that took race/ethnicity into account. African American, Chinese, and Hispanic women had similar median ages at menopause to Caucasian women (51.4, 51.5, and 51.0 years vs. 51.4 years, respectively), while Japanese women were found to have a statistically significantly later median age at menopause (51.8 vs. 51.4, p < 0.05).[7] In an analysis that specifically focused on premature ovarian failure (POF), defined as the cessation of menses for 12 months before the age of 40 for no discernible cause, the prevalence of POF was 1.1%. This prevalence differed by race/ethnicity with African American, Caucasian, and Hispanic women reporting POF statistically significantly more often than Japanese women (1.4, 1.0, and 1.4% vs. 0.1%). The proportion of Chinese and Japanese women with POF was similar (0.5 vs. 0.1%).[8]

Longitudinal Findings. Longitudinal analyses of SWAN data suggest similar trends as compared with the SWAN cross-sectional analyses. In one SWAN analysis focused on changes in coronary artery risk factors, after 10 years of follow-up, there were slight racial/ethnic differences in the proportion of women experiencing a final menstrual period (FMP) as compared with remaining premenopausal. Japanese women remained premenopausal most often and Hispanic women least often; however, these differences did not achieve statistical significance in univariate analyses[9] (and multivariable analyses were not performed).

To date, investigations solely focused on racial/ethnic differences in the timing of menopause in SWAN are not available. This is an important missing piece of evidence on the relationship between race/ethnicity and the timing of menopause. However, it should be noted that there are inherent limitations with regard to race/ethnicity data in SWAN. Several of the racial/ethnic groups in SWAN lack regional diversity. For example, Japanese American women were recruited from Los Angeles, California exclusively. Thus, it is impossible to distinguish a racial versus a geographic contributor to differences seen in the timing of menopause in the racial/ethnic groups recruited from single sites in SWAN.

Pennsylvania Ovarian Aging Study. The POAS began in 1996 and used sample stratification to recruit equal numbers of African American and white women from Philadelphia, PA. A total of 401 women were recruited for the baseline examination and these women were 35 to 48 years of age and reported regular menstrual cycles at that time. After 14 years of follow-up, in univariate analyses, race was not associated with the timing of menopause which, when it occurred, occurred at a mean age of 50.9. However, after adjustment, African American race was associated with a slightly later age at menopause and this racial difference reached statistical significance.[10]

Timing of Menopause: Summary

Overall, the patterns of racial differences in the timing of menopause are not consistent. While Japanese women appear to have a later menopause and less POF than women of other racial and ethnic groups in the SWAN study, it is unclear as to whether this is a racial/ethnic difference or a geographic difference. However, cross-sectional studies in other cohort studies, although often designed for other purposes such as for cancer epidemiology, do appear to corroborate this later age at menopause for Japanese women.[11]

While it would be easy to conclude that African American women have an earlier menopause (especially given that the chronic diseases that disproportionally burden African American women, such as coronary heart disease, are also associated with earlier menopause), the available data on an African American-specific difference in reproductive aging are far from persuasive. In fact, as reported above, POAS found a later menopause in African American women. Thus, a more appropriate conclusion is that more data are needed regarding the relationship between race and the timing of menopause with regard to African American women.

Racial Differences in Hypothalamic-Pituitary-Ovarian Relationship

Hormone levels are inextricably linked with menopausal stages. Some researchers have investigated racial/ethnic differences in hormones within menopausal stages and hormonal trajectories and in transition through multiple menopausal stages. There are advantages to analyses that focus on differences in ovarian hormones and biomarkers of ovarian function and compared with other factors related to menopause. For example, biomarkers are not subject to recall bias that plagues studies of self-reported variables. Therefore, studies of racial/ethnic differences in hormone levels, both cross-sectional and longitudinal studies, may provide important information on potential differences in women attributable to race/ethnicity.

Studies with Data on African American Women. Several studies on differences in hormone levels in African American women versus Caucasian women have been performed. The results suggest that the differences are likely subtle, but may still have relevance to the development of racial differences in hormonally related diseases. Several studies have focused on differences in estradiol (E2) levels in cycling premenopausal women. In five such studies, estradiol levels were higher in African American women as compared with Caucasian women.[12–16] However, several of these studies did not adjust for body mass index (BMI) (and African American women often have significantly higher BMI).

In a recent, comprehensive study by Marsh et al that included daily blood draws for an entire monthly cycle in which women were matched for age and BMI, estradiol levels were higher in African American women (vs. Caucasian).[15] This difference was most pronounced during the late follicular, early luteal, and late luteal phases. However, a racial difference in estradiol was not seen during the early follicular phase. In contrast to estradiol, androstenedione levels were similar between African Americans and Caucasians during all phases of the menstrual cycle.

Because of the racial/ethnic differences in estradiol and the lack of differences in androstenedione, the authors attributed the findings to a potential racial/ethnic difference in aromatase. Supporting this premise, a study investigating aromatase gene polymorphisms in SWAN found that an aromatase gene polymorphism associated with a high estradiol-to-testosterone ratio (measured during the follicular phase) was more prevalent in African American women.[17]

Analysis from POAS tells a different story. In a subsample of women from POAS who were initially 35 to 47 years, POAS found that premenopausal estradiol levels were lower in African American women as compared with Caucasian women. Estradiol is inversely associated with BMI in premenopausal women. However, even after control for BMI, African American women still had significantly lower estradiol than Caucasian women during premenopause in cross-sectional analyses.[18,19]

Longitudinal POAS data show that African American versus Caucasian differences in POAS are strongly mediated by BMI. In longitudinal analyses in POAS, African American women had lower estradiol than Caucasian women during premenopause and higher estradiol than Caucasian women in postmenopause. However, in these longitudinal analyses, BMI and estradiol were inversely related in premenopause (i.e., women with higher BMI had lower estradiol), while they were positively related in postmenopause (women with higher BMI had higher estradiol). Since African American women were heavier in POAS, BMI explained much of the estradiol level differences observed between African American and Caucasian women. However, even after adjustment for BMI, African American women still had lower estradiol levels during premenopause.[19]

Why would African American women in POAS have a racial difference in estradiol (i.e., lower estradiol in during premenopause) that is potentially opposite of that seen in the study of Marsh et al?[15] One potential explanation is the cycle phase of the sex-steroid measurement, as the study of Marsh et al[15] did not see a racial difference in estradiol levels when estradiol was measured during the early follicular phase. Another explanation is that racial differences in young premenopausal women in their 20s and early 30s are distinct from potential difference in premenopausal women at ages closer to menopause (40s). Thus, there may be complex racial differences that vary with cycle phase, menopausal stage, and age at that menopausal stage that warrant further investigation.

Study of Women Across the Nation. In cross-sectional analyses at baseline in SWAN, serum estradiol levels were inversely correlated with BMI. In univariate analyses, Caucasian and Hispanic women have the highest premenopausal estradiol levels (80 and 81 pg/mL, respectively), while African American, Chinese, and Japanese women's estradiol levels were 73, 62, and 68 pg/mL, respectively (ANOVA p < 0.05). These racial differences, however, appeared highly confounded by BMI, and after adjustment for BMI, any racial/ethnic difference was eliminated.[20]

In longitudinal analyses, differences among racial/ethnic groups were present. Longitudinal analyses were performed with adjustment for both BMI and menopausal stage. SWAN researchers found that for any given BMI, African American and Caucasian women had similar estradiol levels throughout the menopausal transition from premenopause to postmenopause. However, throughout this transition, follicle-stimulating hormone (FSH) levels were significantly higher in African American women as compared with Caucasian women of the same BMI. SWAN researchers attributed the racial differences in FSH levels to potential racial/ethnic differences in the pituitary's sensitivity to estradiol.[21]

Summary. Studies of racial/ethnic differences between African American and Caucasian women in hormone levels and patterns during the menopausal transition suggest as a highly complex picture. The findings of racial differences in aromatase polymorphisms, estrogen/testosterone ratios, and FSH versus estradiol levels between African Americans and Caucasians, which appear distinct from differences in adiposity, are intriguing. However, they are difficult to reconcile. More research is needed to replicate and validate these findings.

Asian Women. In SWAN, in cross-sectional analyses, estradiol levels were significantly lower in both Chinese and Japanese women as compared with Caucasian women. Differences were still present, but were not statistically significant, after BMI was taken into account in baseline cross-sectional analyses.[20] In the longitudinal study, where both BMI and menopausal stage were taken into account, findings were more pronounced. Chinese and Japanese women had lower estradiol concentrations as compared with Caucasians throughout the transition through menopause.[21] In contrast to African American women, their FSH levels were similar to Caucasian women throughout the menopausal transition.[21]

Obesity Impact

Obesity is in epidemic proportion in the United States, but disproportionately affects African American and Hispanic women. There is evidence of an increase in obesity with midlife transition. Whether this increase is due to aging, menopausal transition, or both remains unclear. In addition, it is unclear whether weight gain and/or development of abdominal obesity induces changes in reproductive hormones or vice versa. Weight gain increases the incidence of diabetes, cancer, and cardiovascular disease. These diseases associated with obesity increases morbidity and mortality for African American and Hispanic women. An improved understanding of the temporal associations between weight gain and changing hormone levels may lead to new strategies to curb weight gain during midlife. SWAN reports on longitudinal associations between changes in selected hormones and changes in waist circumference and body weight in a multiethnic cohort of women undergoing the menopause transition.

Obesity markedly attenuated the rise in FSH, particularly in the period after the FMP. This was noted in Caucasian and African American women reported by Randolph et al.[20] This is consistent with the cross-sectional report of premenopausal and early perimenopausal values reported from SWAN.[20] Similarly, obesity was negatively associated with premenopausal E2 levels but then reversed with the FMP and became positively associated with serum E2, an effect also previously reported in SWAN.[21] This suggests a postmenopausal shift from ovarian E2 secretion to a compensatory source most likely in fat, a known site of the enzyme aromatase and peripheral production of estrogens from androgen precursors. Given higher ovarian estrogen production in premenopausal women, it overshadows peripheral production.