Menarche and menopause mark the onset and offset of ovarian and endocrine activity associated with reproduction, and early menarche and late menopause are known to increase risk of breast cancer. In a recent Lancet Oncology article, the Collaborative Group on Hormonal Factors in Breast Cancer reported findings of a meta-analysis that assessed the strength of these risk associations and determined whether the associations differ by characteristics of affected women or tumor subtypes.1
The meta-analysis included individual data from 117 epidemiologic studies involving 118,964 women with invasive breast cancer and 306,091 women without such disease, of whom none had used menopausal hormone therapy. The studies were performed in 35 countries, mostly in Europe or North America. The median year of birth of women with breast cancer was 1939, and the median age at cancer diagnosis was 54 years.
Among controls, mean age at menarche was 13.1 years, with 65% reporting menarche at 12 to 14 years and 16% having menarche at 11 years or younger. Breast cancer risk increased by 5% for each year younger at menarche (relative risk [RR] = 1.050 [P < .0001] for each year) when results were stratified by study, age, year of birth, parity and age at first birth, height, current body mass index, smoking, and alcohol consumption. Additional adjustment by ethnic origin, hormonal contraceptive use, and family history of breast cancer altered the excess relative risk estimate by less than 1%.
On subgroup analyses, the relative risk for each year younger at menarche was reduced by current body mass index ≥ 25 kg/m2, although the effect was significant only for postmenopausal women (RR = 1.031, P = .006 for heterogeneity with lower body mass index); by greater attained age in postmenopausal women (1.039 for age ≥ 65 and 1.044 for age 55–64, P = .04 for heterogeneity among age groups); and by family history of breast cancer (1.001, P = .01 for heterogeneity with no family history). Heterogeneity across groups was not significant for any other factor examined (ie, year of birth, ethnic origin, parity, age at first birth, height, body mass index as young adult, oral contraceptive use, menopausal status, and age at menopause in postmenopausal women).
Information on tumor characteristics was available from 85 studies. The association with age at menarche was significantly stronger for lobular tumors vs ductal tumors (RR = 1.073 vs 1.035 for each year younger at menarche, P = .0001 for heterogeneity). There was no significant difference in relative risk according to estrogen receptor status.
Women Aged 45 to 54 Years
Among postmenopausal controls, mean age at natural menopause was 49.3 years, with menopause occurring in 15% before age 45 and 75% between 45 and 54 years. There was little association between age at menarche and age at menopause. Late menopause was weakly associated with year of birth, childbearing history, body mass index, and alcohol consumption, with the strongest association being a negative association with smoking.
The short-term effect of reduced ovarian hormone reduction on breast cancer risk was assessed by restricting analysis to 31,000 cases and 70,000 controls aged 45 to 54 years. Among these women, risk was 43% greater among premenopausal women than among postmenopausal women of identical age (RR = 1.43, P < .0001), indicating a rapid decline in risk soon after menopause; risk in perimenopausal women was intermediate between these two groups.
Subgroup analyses showed little heterogeneity, except for an attenuation of risk in premenopausal women by adiposity. Premenopausal women with body mass index ≥ 25 kg/m2 had lower risk than leaner women, whereas the reverse was found for postmenopausal women. The relative risk for breast cancer decreased significantly more rapidly after menopause (P < .0001) in lean vs overweight or obese women, likely reflecting at least in part differences in estradiol concentrations between the two groups.
The proportion of breast tumors that were estrogen receptor–positive and of lobular histology increased with age in both premenopausal and postmenopausal women, with a sudden decrease occurring for both estrogen receptor–positive and lobular tumors around the time of menopause. Heterogeneity between premenopausal women and postmenopausal women aged 45 to 54 years was significant for both estrogen receptor status and tumor histology.
In analysis restricted to postmenopausal women, relative risk for breast cancer increased by 2.9% for each year older at menopause (RR = 1.029, P < .0001 for each year older at menopause) with stratification for study, year of birth, age, parity and age at first birth, height, current body mass index, smoking, and alcohol consumption. Additional adjustment for ethnic origin, age at menarche, family history of breast cancer, and hormonal contraceptive use altered the excess relative risk estimate by less than 1%. Relative risks did not differ significantly between women with natural menopause and those with bilateral oopherectomy.
The relative risk for breast cancer per year older at menopause was reduced by current body mass index ≥ 25 kg/m2 (RR = 1.024, P = .008 for heterogeneity between higher and lower body mass index) and age < 25 years at first birth (1.023, P = .03 for heterogeneity between younger and older age at first birth); the relative risk was not significantly affected by the other patient characteristics examined. The relative risk was significantly increased for estrogen receptor–positive disease (1.031, P = .01 for heterogeneity with estrogen receptor–negative disease) and for lobular histology (1.036, P = .006 for heterogeneity with ductal histology). Among postmenopausal women, the relative risk for each year younger at menarche (1.045) was significantly greater than the relative risk for each year older at menopause (P = .001 for heterogeneity).
As summarized by the authors, “Breast cancer risk increased by a significantly greater factor for every year younger at menarche than for every year older at menopause, indicating that menarche and menopause may not affect breast cancer risk merely by extending women’s total reproductive years. Endogenous ovarian hormones are more relevant for estrogen receptor–positive disease than for estrogen receptor–negative disease and for lobular than for ductal tumors.” ■
1. Collaborative Group on Hormonal Factors in Breast Cancer: Menarche, menopause, and breast cancer risk: Individual participant meta-analysis, including 118,964 women with breast cancer from 117 epidemiological studies. Lancet Oncol 13(11):1141-1151, 2012.