- Email: [email protected]
Ages at menarche and menopause, and mortality among postmenopausal women
Maturitas 130 (2019) 50–56
Contents lists available at ScienceDirect
Maturitas journal homepage: www.elsevier.com/locate/maturitas
Ages at menarche and menopause, and mortality among postmenopausal women
T
Xi Zhanga, Luyang Liub, Fengju Songb, Yiqing Songc, Hongji Daib,⁎ a
Clinical Research Unit, Xin Hua Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, PR China Department of Epidemiology and Biostatistics, National Clinical Research Center for Cancer, Key Laboratory of Molecular Cancer Epidemiology of Tianjin, Tianjin Medical University Cancer Institute and Hospital, Tianjin Medical University, Tianjin, PR China c Department of Epidemiology, Richard M. Fairbanks School of Public Health, Indiana University, Indianapolis, IN, USA b
ARTICLE INFO
ABSTRACT
Keywords: Menarche Menopause Mortality Cause-specific mortality
Objectives: Although both age at menarche and age at menopause may independently affect the risk of cardiovascular diseases and all-cause mortality, their joint association with mortality is less clear. The objectives of this study were to address the relationship between ages at menarche and at menopause with mortality among postmenopausal women. Study design: The study included 75,359 U.S. postmenopausal women aged 50–78 years from the Prostate, Lung, Colorectal, and Ovarian (PLCO) cohort. Information on ages at menarche and menopause was self-reported and collected at baseline, by questionnaire. Main outcome measures: All-cause, cardiovascular and cancer mortality. Results: After a median follow-up of 13 years, we identified 7826 deaths among 75,359 women in the PLCO cohort. Compared with women with an age at menarche of 12–13 years and an age at menopause of 45–54 years, the adjusted hazard ratios (95% confidence interval) for all-cause mortality for women with early menarche (≤11 years) and menopause (≤44 years) and those with late menarche (≥14 years) and menopause (≥55 years) were 1.20 (1.09, 1.32) and 0.82 (0.71, 0.96), respectively. This association remained significant in a sensitivity analysis that excluded women who did not undergo natural menopause. The indexes for the additive effect of the combined association showed no excess risk due to an interaction. Conclusions: Early menarche and early menopause seemed to have an exactly additive effect on all-cause mortality. The findings suggest that it is important to evaluate ages at both menarche and menopause rather than to consider either variable on its own in assessing the risk of mortality.
1. Introduction Menarche and menopause are two reproductive milestones in a woman’s life and strongly reflect endogenous hormonal changes in puberty and adulthood. Early onset of menarche or pre-mature menopause has been reported linked to risks of cardiovascular diseases (CVD) and several cancers [1–3]. The U-shaped relation between age at menarche and risk of coronary disease was observed in the previous study [4]. Most [5–11], but not all [12], relevant studies reported that an earlier age at menarche was associated with a higher risk for all-cause mortality or selected cause-specific mortalities. Meanwhile, higher risks of major cardiovascular diseases and all-cause mortality have been observed among women with early menopause [10,13–15]. Although previous studies shown early menarche or menopause may be independently associated with mortality, their joint association ⁎
with mortality was less clear. It is largely unknown which menstrual factor could be a key indicator for mortality among post-menopausal women. Two previous studies have shown conflicting information regarding the association between reproductive life span (defined as the duration from menarche to menopause) and risk of mortality. The Rotterdam Study including 4076 postmenopausal women reported longer reproductive life span was significantly harmful for all-cause mortality [16]. A larger study of Women’s Health Initiative including 16,251 postmenopausal women shown that longer reproductive lifespan was significantly associated with increased longevity [17]. Of note, it remains controversial whether the reproductive lifespan used in these studies sufficiently reflected the combined effect of these two menstrual factors, especially for populations with short follow-up period or limited number of postmenopausal women. Ages at onset of menarche and menopause are heritable and under
Corresponding author at: Tianjin Medical University Cancer Institute and Hospital, Tianjin Medical University, Tianjin, 300060, PR China. E-mail address: [email protected] (H. Dai).
https://doi.org/10.1016/j.maturitas.2019.10.009 Received 8 August 2019; Received in revised form 19 September 2019; Accepted 13 October 2019 0378-5122/ © 2019 Elsevier B.V. All rights reserved.
Maturitas 130 (2019) 50–56
X. Zhang, et al.
race, body mass index (BMI) at baseline (in kg/m2), current smoking status (never, current, and past), current alcohol consumption (never, current, and past), marital status (married or living as married, widowed, divorced or separated, and never married), education level (12 years or less, post high school, and college graduate or postgraduate), hormone replacement therapy use (never, current, former, and unknown use), CVD history, number of live births (0, 1–2, and ≥3), age at first birth (< 16, 16–19, 20–24, 25–29, 30–34, and 35–39 years) and randomized arms (intervention and control groups), reason for menstrual period stopped (natural, unnatural and unknown reason). Information about physical activity, including current physical activity status at baseline (none or < 1 day/week, 2–3 days/week, 4–5 days/ week, and 6–7 days/week) and changes in physical activity in past 10 years (more active, about the same, and less active), were obtained from supplemental questionnaires which contains additional data not collected at baseline.
genetic and nutritional influences [18–20]. Changes in both, the average ages at menarche and menopause, have implications on the total length of exposure to high levels of circulating estrogens during reproductive years [21]. Evaluating the combination of age at menarche and age at menopause in relation to all-cause and cause-specific mortality risks may provide further insights into how reproductive ages influence the general life-course health of women. Therefore, in this study, using data from a large-scale and long-term prospective study, the Prostate, Lung, Colorectal, and Ovarian (PLCO) cohort, we aimed to comprehensively evaluate the combined associations of ages at menarche and menopause with risks of all-cause and cause-specific mortality among postmenopausal women. 2. Methods 2.1. Study population The PLCO study is a randomized, controlled trial of screening tests for prostate, lung, colorectal and ovarian cancers in the United States. Detailed information about the PLCO study design and procedures has been reported previously [22]. Briefly, all eligible participants enrolled and randomly assigned to either usual care arm or screening arm between November 1993 and July 2001. A total of 78,215 women completed a baseline questionnaire which contained information on demographics, social economics, lifestyle, and medical history. Participants with missing information on race (n = 27), age at menarche or age at menopause (n = 2829) were excluded from the analyses. Finally, a total of 75,359 women were included into the analyses.
2.5. Statistical analysis Person-years of follow-up for each participant were calculated from the date of the baseline interview through the date of death, last contact or December 31, 2009. We divided women into three categories by menstrual variables based on their frequency distributions in the study population. We chose ages of 12–13 years and 45–54 years as the reference groups for age at menarche and age at menopause, respectively. The reference groups were chosen as they cover the medians of each menstrual variable. We evaluated the associations of menarche (early menarche, age ≤11; late menarche, age ≥14 years) and menopause (early menopause, age ≤44; late menopause, age ≥55 years) with mortality risk using multivariable adjusted Cox proportional hazard regression models and estimated the hazard ratios (HRs) and 95% confidence intervals (CIs). To assess the combined associations, we estimated adjusted HRs with 95% CIs for the association between 9 categories of age at menarche and age at menopause with risk of mortality and used the Gail-Simon method to test the interactions. We also calculated the indexes for additive effect, including synergy index (SI), relative excess risk due to interaction (RERI), and the attributable proportion due to interaction (AP) and their 95% CIs to evaluate the interaction in additive scale [25]. In order to assess the magnitude of associations among women with similar menarche ages, we further did stratified analyses for combined associations of age at menarche and age at menopause with all-cause and cause-specific mortality. Covariates adjusted in all multivariable analyses included age, race/ethnicities, BMI, smoking status, alcohol consumption, physical activity, education levels, hormone replacement therapy use, live births, CVD history, reasons for menstrual period stopped, and intervention arms. As a sensitivity analysis, we excluded women who were not naturally menopaused (e.g. by surgery, radiation or drug therapy) and evaluated the combined association of menarche and natural menopause with mortality risk. All statistical analyses were conducted using Statistical Analysis System software (SAS, version 9.4; SAS Institute, Cary, NC). All statistical tests were two tailed, and the significance level was set at P < 0.05.
2.2. Assessment of exposure Information on ages of menarche and menopause was self-reported and collected at baseline questionnaire. Age of menarche was defined as the age at first menstrual period. Age at menopause was defined as the age at final menstrual period, because of any reasons (natural and unnatural). Both of these two variables were originally categorized into five groups (age at menarche: < 10, 10–11, 12–13, 14–15, and ≥16 years; age at menopause: < 40, 40–44, 45–49, 50–54, and ≥55 years). 2.3. Assessment of outcome All participants were asked to complete follow-up questionnaires to report cancer diagnosis. Follow-up was censored at study withdrawal, last annual questionnaire interview, the 13th year of follow-up, or end of study through December 31, 2009, whichever came first. Deaths were ascertained primarily according to mail annual study update forms, newsletters and birthday cards; additionally, if there was no response to repeated attempts to contact, these participants would be included for an annually National Death Index search. All deaths were reassessed and their medical records were reviewed according to the Death Review Process [23,24]. Cause-specific deaths including deaths from CVD and cancers, which were selected according to the Ninth Version of the International Classification of Diseases (ICD-9). Cardiovascular mortality included deaths from ischemic heart disease (410414), heart rhythm disturbance (426, 427), heart failure (428), and cerebrovascular disease/stroke (430-438). Cancers included the ICD-9 codes 140-208. Women specific cancers included cancers of breast (174), cervix (180), ovary (183), vagina (184.0) and endometrium (179,182).
3. Results A total of 7826 deaths were reported during a median follow-up of 12.5 years (range: 5.25–13.0 years). The total mortality rate is 9.17 per 1000 person-years. Baseline characteristics of study participants are described in Table 1. Mean age of participated women was 62.6 ± 5.4 years. The mean BMI was 27.1 ± 5.6 kg/m2. Among a total of 75,359 women, a majority of them had their first menstruation at 12–13 years of age (54%) and their last time of menstruation at 45–54 ages (60%). About 20% women experienced early menarche at age ≤11 years, and 29% women experienced early menopause at age ≤44 years. When
2.4. Potential confounders Socioeconomic and environmental condition may have potential biological associations between menstrual characteristics and mortality risk. Therefore, we obtained the following information as covariates in the statistical models from the baseline questionnaire: baseline age, 51
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Table 1 Characteristics of participants according to age at menarche and age at menopause. Age at menarche, years
≤11 (n = 15,345)
12-13 (n = 40,491)
≥14 (n = 19,523)
Total (n = 75,359)
Age, years Non-Hispanic Whites, % Baseline BMI, km/m2 Current smokers, % Current alcohol drinking, % Educational years > 12, % Physical activity ≥4–5 days/week, % Number of live births ≥3 Current hormone therapy users, % Age at the first birth ≥35 years, % Age at menopause ≥55 years, % History of CVD, % Natural menopause, %
61.9 (5.27) 13,482 (87.9) 28.6 (6.19) 1551 (10.1) 8067 (66.8) 10,347 (67.5) 2215 (14.37) 8954 (58.4) 7582 (49.5) 236 (1.71) 1594 (10.4) 6251 (40.7) 8854 (58.2)
62.5 (5.38) 36,228 (89.4) 27.0 (5.41) 3826 (9.45) 22,299 (68.8) 26,843 (66.3) 5736 (14.21) 23,856 (59.0) 20,013 (49.5) 723 (1.98) 4420 (10.9) 14,536 (35.9) 25,290 (62.9)
63.0 (5.48) 17,083 (87.5) 26.3 (5.09) 1947 (9.97) 10,489 (68.2) 12,382 (64.4)) 2783 (14.21) 11,889 (61.0) 9261 (47.5) 375 (2.11) 2370 (12.1) 6632 (34.0) 12,435 (64.1)
62.5 (5.40) 66,793 (88.6) 27.1 (5.55) 7324 (9.72) 40,855 (68.3) 49,572 (65.8) 10,724 (14.2) 44,699 (59.4) 36,856 (49.0) 1334 (1.96) 8384 (11.1) 27,419 (36.4) 46,579 (62.2)
Age at menopause, years
≤44 (n = 21,537)
45–54 (n = 45,438)
≥55 (n = 8384)
Total (n = 75,359)
Age, years Non-Hispanic Whites, % Baseline BMI, km/m2 Current smokers, % Current alcohol drinking, % Educational years > 12, % Physical activity ≥4–5 days/week, % Number of live births ≥3, % Current hormone therapy users, % Age at the first birth ≥35 years, % Age at menarche ≤11 years, % History of CVD, % Natural menopause, %
62.4 (5.25) 18,777 (87.2) 27.6 (5.71) 2658 (12.3) 10,927 (64.8) 13,315 (61.9) 2786 (12.93) 12,802 (59.6) 11,836 (55.0) 224 (1.15) 5102 (23.7) 8678 (40.3) 4771 (22.2)
62.7 (5.48) 40,469 (89.1) 26.9 (5.46) 4178 (9.20) 25,152 (69.5) 30,356 (66.8) 6582 (14.48) 26,843 (59.2) 20,491 (45.1) 964 (2.35) 8649 (19.0) 15,802 (34.8) 35,511 (78.2)
61.8 (5.23) 7547 (90.0) 27.2 (5.56) 488 (5.82) 4776 (70.4) 5901 (70.4) 1356 (16.19) 5054 (60.4) 4529 (54.1) 146 (1.90) 1594 (19.0) 2939 (35.1) 6297 (79.8)
62.5 (5.40) 66,793 (88.6) 27.1 (5.55) 7324 (9.72) 40,855 (68.3) 49,572 (65.8) 10,724 (14.2) 44,699 (59.4) 36,856 (49.0) 1334 (20.4) 15,345 (11.1) 27,419 (36.4) 45,679 (62.2)
considered in combination, about 6.8% women experienced both early menarche and early menopause. Women with late menarche, but not late menopause, tended to be older and have lower BMI. Around 10% of women reported to be current smokers while 68% be current alcohol user at baseline survey. More than half (59%) of the participants had more than 2 live births. Nearly half (49%) of the participants were hormone therapy users. Age at menarche was significantly associated with risks of all-cause and cancer mortality, but not CVD or women specific cancer mortality
(Table 2). Compared with women with an age at menarche of 12–13 years, women with early menarche had 11% higher risks of all-cause (HR = 1.11, 95% CI 1.05–1.18) and cancer mortality (HR = 1.11, 95% CI 1.01–1.22), and women with late menarche tended to have 7% lower risk of all cause (HR = 0.93, 95% CI 0.88–0.99) and 12% lower risk of cancer mortality (HR = 0.88; 95% CI 0.80–0.96). There was a trend towards decreased risks of all-cause, cancer, and women specific cancer mortalities as increasing age at menarche (all P-values for trend < 0.05).
Table 2 Associations between ages at menarche and all-cause, CVD, cancer, and women specific cancer mortality. Age at menarche
No. of women (mortality ratea)
All-cause mortality ≤11 years 15,345 (9.91) 12–13 years 40,491 (8.89) ≥14 years 19,523 (9.18) P for trend 75,359 (9.17) CVD mortality ≤11 years 15,345 (1.95) 12–13 years 40,491 (1.74) ≥14 years 19,523 (2.04) P for trend 75,359 (1.86) Cancer mortality ≤11 years 15,345 (3.90) 12–13 years 40,491 (3.56) ≥14 years 19,523 (3.41) P for trend 75,359 (3.59) Women specific cancer mortality ≤11 years 15,345 (0.84) 12–13 years 40,491 (0.71) ≥14 years 19,523 (0.70) P for trend 75,359 (0.73)
No. of deaths
HRb
95% CI
HRc
95% CI
HRd
95% CI
1707 4083 2036 7826
1.20 1.00 0.97 < 0.0001
1.13, 1.27 Ref. 0.92, 1.03
1.12 1.00 0.93 < 0.0001
1.06, 1.18 Ref. 0.88, 0.98
1.11 1.00 0.93 < 0.0001
1.05, 1.18 Ref. 0.88, 0.99
335 797 452 1584
1.24 1.00 1.08 0.13
1.09, 1.40 Ref. 0.96, 1.21
1.10 1.00 1.04 0.54
0.96, 1.25 Ref. 0.92, 1.17
1.09 1.00 1.04 0.60
0.96, 1.24 Ref. 0.92, 1.17
671 1633 756 3060
1.15 1.00 0.92 < 0.0001
1.05, 1.26 Ref. 0.85, 1.01
1.11 1.00 0.88 < 0.0001
1.01, 1.22 Ref. 0.8, 0.96
1.11 1.00 0.88 < 0.0001
1.01, 1.22 Ref. 0.80, 0.96
145 326 155 626
1.24 1.00 0.95 0.03
1.02, 1.50 Ref. 0.79, 1.15
1.16 1.00 0.92 0.06
0.95, 1.42 Ref. 0.76, 1.12
1.17 1.00 0.92 0.05
0.96, 1.43 Ref. 0.76, 1.12
a
Deaths per 1000 person per year. Adjusted for age, race, and arm. c Adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone therapy uses, live birth, history of CVD, reasons for menstrual period stopped, and intervention arms. d Additionally adjusted for age at natural menopause. b
52
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Table 3 Associations between ages at menopause and all-cause, CVD, cancer, and women specific cancer mortality. Age at menopause
No. of women (Mortality ratea)
All-cause mortality ≤44 years 21,537 (10.6) 45–54 years 45,438 (8.93) ≥55 years 8384 (7.04) P for trend 75,359 (9.17) CVD mortality ≤44 years 21,537 (2.20) 45–54 years 45,438 (1.79) ≥55 years 8384 (1.35) P for trend 75,359 (1.86) Cancer mortality ≤44 years 21,537 (3.86) 45–54 years 45,438 (3.61) ≥55 years 8384 (2.81) P for trend 75,359 (3.59) Women specific cancer mortality ≤44 years 21,537 (0.63) 45–54 years 45,438 (0.77) ≥55 years 8384 (0.77) P for trend 75,359 (0.73)
No. of deaths
HRb
95% CI
HRc
95% CI
HRd
95% CI
2538 4613 675 7826
1.24 1.00 0.87 < 0.0001
1.18, 1.30 Ref. 0.80, 0.94
1.12 1.00 0.92 < 0.0001
1.05, 1.18 Ref. 0.85, 1.00
1.11 1.00 0.92 < 0.0001
1.05, 1.18 Ref. 0.85, 1.00
530 924 130 1584
1.32 1.00 0.87 < 0.0001
1.18, 1.47 Ref. 0.72, 1.04
1.14 1.00 0.89 0.008
1.01, 1.30 Ref. 0.73, 1.07
1.14 1.00 0.89 0.009
1.01, 1.30 Ref. 0.73, 1.07
927 1863 270 3060
1.10 1.00 0.83 < 0.0001
1.02, 1.19 Ref. 0.73, 0.94
1.03 1.00 0.89 0.10
0.93, 1.13 Ref. 0.78, 1.02
1.02 1.00 0.90 0.14
0.93, 1.12 Ref. 0.79, 1.02
152 400 74 626
0.83 1.00 1.05 0.05
0.69, 1.01 Ref. 0.82, 1.34
0.77 1.00 1.07 0.02
0.62, 0.96 Ref. 0.83, 1.37
0.76 1.00 1.07 0.02
0.61, 0.95 Ref. 0.83, 1.37
a
Deaths per 1000 person per year. Adjusted for age, race, and arm. c Adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone therapy uses, live birth, history of CVD, reasons for menstrual period stopped, and intervention arms. d Additionally adjusted for age at natural menarche. b
Age at menopause was significantly associated with risks of allcause, CVD, and women specific cancer mortality, but not with total cancer mortality (Table 3). Women with early menopause had 11% and 14% higher risk of all-cause (HR = 1.11, 95% CI 1.05–1.18) and CVD mortality (HR = 1.14, 95% CI 1.01–1.30), respectively, compared with women with menopausal age of 45–54 years. There was a trend towards decreased risk of all-cause and CVD mortality with increasing age at menopause (all P-values for trend < 0.05). For women specific cancers, early menopause tended to have a lower mortality risk (HR = 0.76, 95% CI 0.61–0.95); while late menopause likely to have a higher mortality risk (HR = 1.07, 95% CI 0.83–1.37) (P for trend = 0.02). A combination of early menarche and early menopause was significantly associated with higher risks of all-cause and CVD mortality (Table 4 and Fig. 1); while a combination of late menarche and menopause seemed to have lower risks. Compared with women with an
age at menarche of 12–13 years and an age at menopause of 45–54 years, women with both early menarche and early menopause were among those with the highest risk of all-cause mortality (HR = 1.20, 95% CI 1.09–1.32) and CVD mortality (HR = 1.25, 95% CI 1.02–1.54), while those with late menarche and late menopause were at the lowest risk of all-cause mortality (HR = 0.82, 95% CI, 0.71–0.96) and CVD mortality (HR = 0.86, 95% CI 0.62–1.21). The results from the sensitivity analyses by excluding unnaturally menopaused women did not substantially change with regard to significance, direction, and size of the effect. (Supplemental Table 1). The indexes for additive effect, including SI, RERI and AP, showed no excess risk due to interaction. All 95% CIs of synergy index included 1 and 95% CIs of AP and RERI covered the null value; and the interactions were not significant. The combined effects of ages at menarche and menopause seems to be “exactly additivity” on moralities. There were no significant
Table 4 Combined associations of age at menarche and age at menopause with all-cause, CVD, cancer, and women specific cancer mortality. Age at menarche
All-cause mortality ≤11 years 12–13 years ≥14 years CVD mortality ≤11 years 12–13 years ≥14 years Cancer mortality ≤11 years 12–13 years ≥14 years Women specific cancer mortality ≤11 years 12–13 years ≥14 years
Age at menopause, years
P
≤44
45–54
≥55
1.20 (1.09, 1.32) 1.12 (1.04, 1.21) 1.06 (0.97, 1.17)
1.14 (1.06, 1.23) 1.00 (Ref.) 0.94 (0.87, 1.00)
1.07 (0.89, 1.27) 0.94 (0.84, 1.05) 0.82 (0.71, 0.96)
1.25 (1.02, 1.54) 1.12 (0.95, 1.33) 1.21 (0.99, 1.48)
1.10 (0.92, 1.31) 1.00 (Ref.) 1.03 (0.89, 1.21)
0.87 (0.56, 1.36) 0.95 (0.74, 1.23) 0.86 (0.62, 1.21)
1.06 (0.91, 1.24) 1.04 (0.93, 1.18) 0.94 (0.81, 1.10)
1.16 (1.03, 1.30) 1.00 (Ref.) 0.88 (0.78, 0.98)
1.07 (0.82, 1.40) 0.91 (0.76, 1.09) 0.74 (0.58, 0.95)
0.82 (0.56, 1.19) 0.84 (0.64, 1.12) 0.73 (0.50, 1.06)
1.34 (1.05, 1.72) 1.00 (Ref.) 0.94 (0.74, 1.20)
0.99 (0.55, 1.77) 1.19 (0.84, 1.67) 1.08 (0.69, 1.70)
interaction
0.74
0.93
0.61
0.52
The reference groups were chosen as they cover the medians of each menstrual variable. Cox model adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone replacement therapy use, CVD history, number of live births, reasons for menstrual period stopped, and randomized arms. 53
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Fig. 1. Adjusted hazard ratios for combined association of ages at menarche and menopause with all-cause (a) and cause-specific mortality (b–d). Analyses were adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone therapy use, live birth, history of CVD, and intervention arms.
cancer mortality. This finding could be explained by the opposite role of premenopausal sex hormones in CVD [27] and cancer development in women [3]. In the combined analysis, the association between menarche and menopause seems to have an exactly additive effect on all-cause mortality and CVD mortality. On the contrary, for cancer mortality, especially women-specific cancers, the risk of mortality was significantly reduced among women with early menarche and early menopause. Breast and ovarian cancers were the majority cancer types in womenspecific cancers. These two commonly diagnosed cancers in women can be connected by similar genetic mutations and subjected to similar hormonal therapy. It is reported that a later menopause increased the risk of fatal uterine or ovarian cancer [3,28]. Previous study indicated the effects of menarche and menopause on breast cancer risk might not be acting merely by lengthening women's total number of reproductive years [1]. Potential mechanisms include the interplay between endogenous hormones and environmental exposure which may lead to heterogeneous cancer outcome. We found the risks of all-cause mortality and CVD mortality associated with early menarche might be slightly attenuated by a late menopause. This finding could have important clinical implications. Hormone replacement therapy (HRT) or other drugs could be used as effective tools for delaying the menopausal age. Women with early menarche could be benefit from using HRT in order to reduce their CVD mortality risk and even all-cause mortality risk. Some studies suggested that shorter reproductive life span was associated with specific mortality risk [29]. However, using reproductive years as an indicator has an obvious pitfall. In that case, the reproductive years of a woman with early menarche (e.g. 8 years) and early menopause (e.g. 45 years) was equal to woman with normal menarche (e.g. 13 years) and normal menopause (e.g. 50 years). Women with same reproductive years could have different mortality risks according to our findings. Therefore, we suggest that it is important to evaluate both ages at menarche and menopause rather than to consider each variable separately in order to assess the mortality risk. Our study findings are strengthened by our ability to adjust for most known confounders, e.g. hormone replacement therapy and unnatural
interactions between ages at menarche and menopause with morality risk (Table 4). In stratified analyses by ages at menarche and menopause, the risks of all-cause mortality and CVD mortality associated with early menarche might be slightly attenuated by the effect of late menopause. On the other hand, women with early menopause and late menarche still had a higher risk of all-cause mortality and cancer mortality (Table 5). Additionally, the risk of women specific cancer mortality reduced by half (HR = 0.49, 95% CI 0.32–0.77) among women with early menarche and early menopause, compared to the reference group. 4. Discussion In this large-scale cohort study, we found that early menarche was associated with higher total mortality and cancer mortality. Early menopause was associated with higher risk of mortality from all causes and CVD, but with lower risk of mortality from women specific cancers. Early menarche and early menopause seemed to have an exactly additivity effect on all-cause mortality. Our findings suggest that both age at menarche and age at menopause could be key indicators for all-cause and CVD mortality among women, although further investigations in mechanistic studies and well-designed prospective studies are needed. To our knowledge, the present study was one of the largest studies to date including the largest population of U.S. postmenopausal women. Consistent with most of previous studies [6,7,11,26], we found a significant association between early age at menarche and increased total mortality risk. As for CVD and cancer mortality, the associations were reported to be significant only among specific populations, e.g. younger women and women without major CVD risk factors [11]. In this study, we found a significant association with CVD mortality only in age, race, and intervention arm-adjusted model. However, this association became non-significant when further adjusting with smoking and alcohol drinking status, hormone therapy uses and other covariates. This implicates that age at menarche could be a robust indicator for risk of total mortality rather than CVD or cancer mortality. We also observed a negative association of ages at menopause with CVD and allcause mortality, while a positive association with women specific 54
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Table 5 Stratified analysis of combined associations of age at menarche and age at menopause with all-cause and cause-specific mortalities. Age at menopause, years
No. of women (Mortality ratea)
No. of deaths
HR (95% CI)b
P
≤44 45–54 ≥55 ≤44 45–54 ≥55 ≤44 45–54 ≥55
5102 (11.06) 8649 (9.72) 1594 (7.39) 11214 (10.23) 24857 (8.63) 4420 (7.05) 5221 (10.78) 11932 (8.98) 2370 (6.78)
625 948 134 1284 2443 356 629 1222 185
1.06 1.00 0.93 1.14 1.00 0.94 1.11 1.00 0.89
(0.94, (Ref.) (0.77, (1.05, (Ref.) (0.84, (0.99, (Ref.) (0.76,
1.20)
0.18
1.12) 1.24)
0.0004
1.05) 1.24)
0.009
≤44 45–54 ≥55 ≤44 45–54 ≥55 ≤44 45–54 ≥55
5102 (2.32) 8649 (1.88) 1594 (1.16) 11214 (2.01) 24857 (1.68) 4420 (1.37) 5221 (2.52) 11932 (1.95) 2370 (1.47)
131 183 21 252 476 69 147 265 40
1.20 1.00 0.78 1.09 1.00 0.96 1.20 1.00 0.84
(0.92, (Ref.) (0.49, (0.91, (Ref.) (0.74, (0.95, (Ref.) (0.59,
≤44 45–54 ≥55 12–13 years ≤44 45–54 ≥55 ≥14 years ≤44 45–54 ≥55 Women specific cancer mortality ≤11 years ≤44 45–54 ≥55 12–13 years ≤44 45–54 ≥55 ≥14 years ≤44 45–54 ≥55
5102 (3.93) 8649 (4.02) 1594 (3.14) 11214 (3.86) 24857 (3.55) 4420 (2.87) 5221 (3.79) 11932 (3.43) 2370 (2.49)
222 392 57 484 1004 145 221 467 68
0.85 1.00 0.94 1.10 1.00 0.91 1.03 1.00 0.85
(0.69, (Ref.) (0.71, (0.97, (Ref.) (0.76, (0.85, (Ref.) (0.66,
5102 (0.67) 8649 (0.97) 1594 (0.66) 11214 (0.64) 24857 (0.72) 4420 (0.81) 5221 (0.58) 11932 (0.73) 2370 (0.77)
38 95 12 80 205 41 34 100 21
0.49 1.00 0.70 0.99 1.00 1.19 0.66 1.00 1.18
(0.32, (Ref.) (0.38, (0.74, (Ref.) (0.85, (0.43, (Ref.) (0.73,
Age at menarche All-cause mortality ≤11 years 12–13 years ≥14 years CVD mortality ≤11 years 12–13 years ≥14 years Cancer mortality ≤11 years
trend
1.04) 1.58)
0.06
1.24) 1.31)
0.31
1.24) 1.52)
0.05
1.18) 1.03)
0.25
1.24) 1.26)
0.05
1.08) 1.24)
0.26
1.10) 0.77)
0.07
1.28) 1.35)
0.45
1.67) 1.04)
0.05
1.90)
a
Deaths per 1000 person per year. Cox model adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone replacement therapy use, CVD history, number of live births, reasons for menstrual period stopped, and randomized arms. b
menopause. In addition, due to the large sample size in the PLCO cohort, majority of our subgroups classified by age categories of menarche and menopause still has sufficient power to test the significant of associations. Our study has several limitations. First, we couldn’t examine the relationship separately by subgroups as defined by each CVD or other cancer subtypes, due to limited number of specific-deaths. Second, because the exposures (age at menarche and age at menopause) were collected retrospectively at baseline, there could be potential biases due to possible misclassification. Third, although we adjusted for many potential confounders, we cannot rule out residual confounding by factors that have not been measured. Finally, the original data for age at menarche and menopause were categorical variables. We have no sufficient statistic power to evaluate the possible non-linearity.
Contributors Xi Zhang designed and conducted the research, contributed to data analysis, drafted the manuscript, interpretation of the results, and is responsible for the final content of the manuscript. Luyang Liu contributed to data collection, and interpretation, and revision of the manuscript. Fengju Song contributed to data collection, and interpretation, and revision of the manuscript. Yiqing Song contributed to study design, interpretation and revision of the manuscript. Hongji Dai designed and conducted the research, contributed to data analysis, drafted the manuscript, interpretation of the results, and is responsible for the final content of the manuscript. All authors read and approved the final version.
5. Conclusions In summary, the current study provides prospective evidence of the combined association between ages at menarche and menopause with all-cause mortality risk in a large population of U.S. postmenopausal women. Better understanding the interplay between endogenous hormone exposure and CVD and cancer development may elucidate novel mechanistic pathways in survival. We suggest evaluate both ages at menarche and menopause rather than to consider each variable separately in order to assess the mortality risk.
Funding This work was supported by Natural Science Foundation of China (81703238) and Tianjin Science & Technology Foundation for Selected Overseas Scholar (2016016).
55
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Ethical statement
[9] N.T. Mueller, A.O. Odegaard, M.D. Gross, W.P. Koh, J.M. Yuan, M.A. Pereira, Age at menarche and cardiovascular disease mortality in Singaporean Chinese women: the Singapore Chinese Health Study, Ann. Epidemiol. 22 (10) (2012) 717–722. [10] K. Murakami, H. Metoki, M. Satoh, K. Asayama, M. Hosaka, A. Matsuda, R. Inoue, M. Tsubota-Utsugi, T. Murakami, K. Nomura, M. Kikuya, Y. Imai, T. Ohkubo, Menstrual factors and stroke incidence in Japanese postmenopausal women: the Ohasama Study, Neuroepidemiology 47 (2) (2016) 109–116. [11] L. Yang, L. Li, I.Y. Millwood, S.A.E. Peters, Y. Chen, Y. Guo, Z. Bian, X. Chen, L. Chen, S. Feng, S. Lv, Z. Pang, M. Woodward, Z. Chen, Age at menarche and risk of major cardiovascular diseases: evidence of birth cohort effects from a prospective study of 300,000 Chinese women, Int. J. Cardiol. 227 (2017) 497–502. [12] M.W. Lundblad, B.K. Jacobsen, Is age at menarche associated with total mortality? The Tromso Study, Int. J. Womens Health 10 (2018) 203–209. [13] D. Gong, J. Sun, Y. Zhou, C. Zou, Y. Fan, Early age at natural menopause and risk of cardiovascular and all-cause mortality: a meta-analysis of prospective observational studies, Int. J. Cardiol. 203 (2016) 115–119. [14] B.K. Jacobsen, I. Heuch, G. Kvale, Age at natural menopause and all-cause mortality: a 37-year follow-up of 19,731 Norwegian women, Am. J. Epidemiol. 157 (10) (2003) 923–929. [15] T. Muka, C. Oliver-Williams, S. Kunutsor, J.S. Laven, B.C. Fauser, R. Chowdhury, M. Kavousi, O.H. Franco, Association of age at onset of menopause and time since onset of menopause with cardiovascular outcomes, intermediate vascular traits, and all-cause mortality: a systematic review and meta-analysis, JAMA Cardiol. 1 (7) (2016) 767–776. [16] L. Jaspers, M. Kavousi, N.S. Erler, A. Hofman, J.S. Laven, O.H. Franco, Fertile lifespan characteristics and all-cause and cause-specific mortality among postmenopausal women: the Rotterdam Study, Fertil. Steril. 107 (2) (2017) 448–456 e1. [17] A.H. Shadyab, C.A. Macera, R.A. Shaffer, S. Jain, L.C. Gallo, M.L. Gass, M.E. Waring, M.L. Stefanick, A.Z. LaCroix, Ages at menarche and menopause and reproductive lifespan as predictors of exceptional longevity in women: the Women’s Health Initiative, Menopause 24 (1) (2017) 35–44. [18] F.R. Day, D.J. Thompson, H. Helgason, D.I. Chasman, H. Finucane, P. Sulem, et al., Genomic analyses identify hundreds of variants associated with age at menarche and support a role for puberty timing in cancer risk, Nat. Genet. 49 (6) (2017) 834–841. [19] C. He, P. Kraft, C. Chen, J.E. Buring, G. Pare, S.E. Hankinson, S.J. Chanock, P.M. Ridker, D.J. Hunter, D.I. Chasman, Genome-wide association studies identify loci associated with age at menarche and age at natural menopause, Nat. Genet. 41 (6) (2009) 724–728. [20] L. Stolk, J.R. Perry, D.I. Chasman, C. He, M. Mangino, P. Sulem, et al., Meta-analyses identify 13 loci associated with age at menopause and highlight DNA repair and immune pathways, Nat. Genet. 44 (3) (2012) 260–268. [21] H.B. Nichols, A. Trentham-Dietz, J.M. Hampton, L. Titus-Ernstoff, K.M. Egan, W.C. Willett, P.A. Newcomb, From menarche to menopause: trends among US Women born from 1912 to 1969, Am. J. Epidemiol. 164 (10) (2006) 1003–1011. [22] J.K. Gohagan, P.C. Prorok, P. Greenwald, B.S. Kramer, The PLCO cancer screening trial: background, goals, organization, operations, results, Rev. Recent Clin. Trials 10 (3) (2015) 173–180. [23] A.B. Miller, R. Feld, R. Fontana, J.K. Gohagan, I. Jatoi, W. Lawrence Jr., A. Miller, K.P. Proro, A. Rajput, M. Sherman, G. Welch, P. Wright, S. Yurgalevitch, P. Albertsen, Changes in and impact of the death review process in the prostate, lung, colorectal and ovarian (PLCO) cancer screening trial, Rev. Recent Clin. Trials 10 (3) (2015) 206–211. [24] A.B. Miller, S. Yurgalevitch, J.L. Weissfeld, L.C. Prostate, T. Ovarian Cancer Screening trial project, death review process in the prostate, lung, colorectal and ovarian (PLCO) cancer screening trial, Control. Clin. Trials 21 (Suppl. 6) (2000) 400S–406S. [25] M.J. Knol, T.J. VanderWeele, R.H. Groenwold, O.H. Klungel, M.M. Rovers, D.E. Grobbee, Estimating measures of interaction on an additive scale for preventive exposures, Eur. J. Epidemiol. 26 (6) (2011) 433–438. [26] R. Lakshman, N.G. Forouhi, S.J. Sharp, R. Luben, S.A. Bingham, K.T. Khaw, N.J. Wareham, K.K. Ong, Early age at menarche associated with cardiovascular disease and mortality, J. Clin. Endocrinol. Metab. 94 (12) (2009) 4953–4960. [27] J.E. Manson, T.K. Woodruff, Reproductive health as a marker of subsequent cardiovascular disease: the role of estrogen, JAMA Cardiol. 1 (7) (2016) 776–777. [28] M.E. Ossewaarde, M.L. Bots, A.L. Verbeek, P.H. Peeters, Y. van der Graaf, D.E. Grobbee, Y.T. van der Schouw, Age at menopause, cause-specific mortality and total life expectancy, Epidemiology 16 (4) (2005) 556–562. [29] X. Wu, H. Cai, A. Kallianpur, Y.T. Gao, G. Yang, W.H. Chow, H.L. Li, W. Zheng, X.O. Shu, Age at menarche and natural menopause and number of reproductive years in association with mortality: results from a median follow-up of 11.2 years among 31,955 naturally menopausal Chinese women, PLoS One 9 (8) (2014) e103673.
We confirm that the study was approved by the Institutional Review Board of the National Cancer Institute and the participating centers (https://www.ncicirb.org/). Each participant provided written consent upon enrollment. Provenance and peer review This article has undergone peer review. Research data (data sharing and collaboration) The data set is available via https://biometry.nci.nih.gov/cdas/ datasets/plco/. Declaration of Competing Interest The authors declare that they have no conflict of interest. Acknowledgements We would like to thank the National Cancer Institute (NCI) for access to data collected by the PLCO study group. We also thank Jerome Mabie of Information Management Services, Inc. for data management and support. The statements contained herein are solely those of the authors and do not represent or imply concurrence or endorsement by NCI. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.maturitas.2019.10. 009. References [1] 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) (2012) 1141–1151. [2] M.M. Hassan, G. Botrus, R. Abdel-Wahab, R.A. Wolff, D. Li, D. Tweardy, A.T. Phan, E. Hawk, M. Javle, J.S. Lee, H.A. Torres, A. Rashid, R. Lenzi, H.M. Hassabo, Y. Abaza, A.S. Shalaby, S. Lacin, J. Morris, Y.Z. Patt, C.I. Amos, S.A. Khaderi, J.A. Goss, P.K. Jalal, A.O. Kaseb, Estrogen replacement reduces risk and increases survival times of women with hepatocellular carcinoma, Clin. Gastroenterol. Hepatol. 15 (11) (2017) 1791–1799. [3] A.L. Shafrir, A. Babic, R.M. Tamimi, B.A. Rosner, S.S. Tworoger, K.L. Terry, Reproductive and hormonal factors in relation to survival and platinum resistance among ovarian cancer cases, Br. J. Cancer 115 (11) (2016) 1391–1399. [4] K. Tamakoshi, H. Yatsuya, A. Tamakoshi, Early age at menarche associated with increased all-cause mortality, Eur. J. Epidemiol. 26 (10) (2011) 771–778. [5] E.K. Bjelland, S. Hofvind, L. Byberg, A. Eskild, The relation of age at menarche with age at natural menopause: a population study of 336 788 women in Norway, Hum. Reprod. 33 (6) (2018) 1149–1157. [6] D. Canoy, V. Beral, A. Balkwill, F.L. Wright, M.E. Kroll, G.K. Reeves, J. Green, B.J. Cairns, Age at menarche and risks of coronary heart and other vascular diseases in a large UK cohort, Circulation 131 (3) (2015) 237–244. [7] B.K. Jacobsen, K. Oda, S.F. Knutsen, G.E. Fraser, Age at menarche, total mortality and mortality from ischaemic heart disease and stroke: the Adventist Health Study, 1976–88, Int. J. Epidemiol. 38 (1) (2009) 245–252. [8] M.A. Merritt, E. Riboli, N. Murphy, M. Kadi, A. Tjonneland, A. Olsen, et al., Reproductive factors and risk of mortality in the European Prospective Investigation into Cancer and Nutrition; a cohort study, BMC Med. 13 (2015) 252.
56
Contents lists available at ScienceDirect
Maturitas journal homepage: www.elsevier.com/locate/maturitas
Ages at menarche and menopause, and mortality among postmenopausal women
T
Xi Zhanga, Luyang Liub, Fengju Songb, Yiqing Songc, Hongji Daib,⁎ a
Clinical Research Unit, Xin Hua Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, PR China Department of Epidemiology and Biostatistics, National Clinical Research Center for Cancer, Key Laboratory of Molecular Cancer Epidemiology of Tianjin, Tianjin Medical University Cancer Institute and Hospital, Tianjin Medical University, Tianjin, PR China c Department of Epidemiology, Richard M. Fairbanks School of Public Health, Indiana University, Indianapolis, IN, USA b
ARTICLE INFO
ABSTRACT
Keywords: Menarche Menopause Mortality Cause-specific mortality
Objectives: Although both age at menarche and age at menopause may independently affect the risk of cardiovascular diseases and all-cause mortality, their joint association with mortality is less clear. The objectives of this study were to address the relationship between ages at menarche and at menopause with mortality among postmenopausal women. Study design: The study included 75,359 U.S. postmenopausal women aged 50–78 years from the Prostate, Lung, Colorectal, and Ovarian (PLCO) cohort. Information on ages at menarche and menopause was self-reported and collected at baseline, by questionnaire. Main outcome measures: All-cause, cardiovascular and cancer mortality. Results: After a median follow-up of 13 years, we identified 7826 deaths among 75,359 women in the PLCO cohort. Compared with women with an age at menarche of 12–13 years and an age at menopause of 45–54 years, the adjusted hazard ratios (95% confidence interval) for all-cause mortality for women with early menarche (≤11 years) and menopause (≤44 years) and those with late menarche (≥14 years) and menopause (≥55 years) were 1.20 (1.09, 1.32) and 0.82 (0.71, 0.96), respectively. This association remained significant in a sensitivity analysis that excluded women who did not undergo natural menopause. The indexes for the additive effect of the combined association showed no excess risk due to an interaction. Conclusions: Early menarche and early menopause seemed to have an exactly additive effect on all-cause mortality. The findings suggest that it is important to evaluate ages at both menarche and menopause rather than to consider either variable on its own in assessing the risk of mortality.
1. Introduction Menarche and menopause are two reproductive milestones in a woman’s life and strongly reflect endogenous hormonal changes in puberty and adulthood. Early onset of menarche or pre-mature menopause has been reported linked to risks of cardiovascular diseases (CVD) and several cancers [1–3]. The U-shaped relation between age at menarche and risk of coronary disease was observed in the previous study [4]. Most [5–11], but not all [12], relevant studies reported that an earlier age at menarche was associated with a higher risk for all-cause mortality or selected cause-specific mortalities. Meanwhile, higher risks of major cardiovascular diseases and all-cause mortality have been observed among women with early menopause [10,13–15]. Although previous studies shown early menarche or menopause may be independently associated with mortality, their joint association ⁎
with mortality was less clear. It is largely unknown which menstrual factor could be a key indicator for mortality among post-menopausal women. Two previous studies have shown conflicting information regarding the association between reproductive life span (defined as the duration from menarche to menopause) and risk of mortality. The Rotterdam Study including 4076 postmenopausal women reported longer reproductive life span was significantly harmful for all-cause mortality [16]. A larger study of Women’s Health Initiative including 16,251 postmenopausal women shown that longer reproductive lifespan was significantly associated with increased longevity [17]. Of note, it remains controversial whether the reproductive lifespan used in these studies sufficiently reflected the combined effect of these two menstrual factors, especially for populations with short follow-up period or limited number of postmenopausal women. Ages at onset of menarche and menopause are heritable and under
Corresponding author at: Tianjin Medical University Cancer Institute and Hospital, Tianjin Medical University, Tianjin, 300060, PR China. E-mail address: [email protected] (H. Dai).
https://doi.org/10.1016/j.maturitas.2019.10.009 Received 8 August 2019; Received in revised form 19 September 2019; Accepted 13 October 2019 0378-5122/ © 2019 Elsevier B.V. All rights reserved.
Maturitas 130 (2019) 50–56
X. Zhang, et al.
race, body mass index (BMI) at baseline (in kg/m2), current smoking status (never, current, and past), current alcohol consumption (never, current, and past), marital status (married or living as married, widowed, divorced or separated, and never married), education level (12 years or less, post high school, and college graduate or postgraduate), hormone replacement therapy use (never, current, former, and unknown use), CVD history, number of live births (0, 1–2, and ≥3), age at first birth (< 16, 16–19, 20–24, 25–29, 30–34, and 35–39 years) and randomized arms (intervention and control groups), reason for menstrual period stopped (natural, unnatural and unknown reason). Information about physical activity, including current physical activity status at baseline (none or < 1 day/week, 2–3 days/week, 4–5 days/ week, and 6–7 days/week) and changes in physical activity in past 10 years (more active, about the same, and less active), were obtained from supplemental questionnaires which contains additional data not collected at baseline.
genetic and nutritional influences [18–20]. Changes in both, the average ages at menarche and menopause, have implications on the total length of exposure to high levels of circulating estrogens during reproductive years [21]. Evaluating the combination of age at menarche and age at menopause in relation to all-cause and cause-specific mortality risks may provide further insights into how reproductive ages influence the general life-course health of women. Therefore, in this study, using data from a large-scale and long-term prospective study, the Prostate, Lung, Colorectal, and Ovarian (PLCO) cohort, we aimed to comprehensively evaluate the combined associations of ages at menarche and menopause with risks of all-cause and cause-specific mortality among postmenopausal women. 2. Methods 2.1. Study population The PLCO study is a randomized, controlled trial of screening tests for prostate, lung, colorectal and ovarian cancers in the United States. Detailed information about the PLCO study design and procedures has been reported previously [22]. Briefly, all eligible participants enrolled and randomly assigned to either usual care arm or screening arm between November 1993 and July 2001. A total of 78,215 women completed a baseline questionnaire which contained information on demographics, social economics, lifestyle, and medical history. Participants with missing information on race (n = 27), age at menarche or age at menopause (n = 2829) were excluded from the analyses. Finally, a total of 75,359 women were included into the analyses.
2.5. Statistical analysis Person-years of follow-up for each participant were calculated from the date of the baseline interview through the date of death, last contact or December 31, 2009. We divided women into three categories by menstrual variables based on their frequency distributions in the study population. We chose ages of 12–13 years and 45–54 years as the reference groups for age at menarche and age at menopause, respectively. The reference groups were chosen as they cover the medians of each menstrual variable. We evaluated the associations of menarche (early menarche, age ≤11; late menarche, age ≥14 years) and menopause (early menopause, age ≤44; late menopause, age ≥55 years) with mortality risk using multivariable adjusted Cox proportional hazard regression models and estimated the hazard ratios (HRs) and 95% confidence intervals (CIs). To assess the combined associations, we estimated adjusted HRs with 95% CIs for the association between 9 categories of age at menarche and age at menopause with risk of mortality and used the Gail-Simon method to test the interactions. We also calculated the indexes for additive effect, including synergy index (SI), relative excess risk due to interaction (RERI), and the attributable proportion due to interaction (AP) and their 95% CIs to evaluate the interaction in additive scale [25]. In order to assess the magnitude of associations among women with similar menarche ages, we further did stratified analyses for combined associations of age at menarche and age at menopause with all-cause and cause-specific mortality. Covariates adjusted in all multivariable analyses included age, race/ethnicities, BMI, smoking status, alcohol consumption, physical activity, education levels, hormone replacement therapy use, live births, CVD history, reasons for menstrual period stopped, and intervention arms. As a sensitivity analysis, we excluded women who were not naturally menopaused (e.g. by surgery, radiation or drug therapy) and evaluated the combined association of menarche and natural menopause with mortality risk. All statistical analyses were conducted using Statistical Analysis System software (SAS, version 9.4; SAS Institute, Cary, NC). All statistical tests were two tailed, and the significance level was set at P < 0.05.
2.2. Assessment of exposure Information on ages of menarche and menopause was self-reported and collected at baseline questionnaire. Age of menarche was defined as the age at first menstrual period. Age at menopause was defined as the age at final menstrual period, because of any reasons (natural and unnatural). Both of these two variables were originally categorized into five groups (age at menarche: < 10, 10–11, 12–13, 14–15, and ≥16 years; age at menopause: < 40, 40–44, 45–49, 50–54, and ≥55 years). 2.3. Assessment of outcome All participants were asked to complete follow-up questionnaires to report cancer diagnosis. Follow-up was censored at study withdrawal, last annual questionnaire interview, the 13th year of follow-up, or end of study through December 31, 2009, whichever came first. Deaths were ascertained primarily according to mail annual study update forms, newsletters and birthday cards; additionally, if there was no response to repeated attempts to contact, these participants would be included for an annually National Death Index search. All deaths were reassessed and their medical records were reviewed according to the Death Review Process [23,24]. Cause-specific deaths including deaths from CVD and cancers, which were selected according to the Ninth Version of the International Classification of Diseases (ICD-9). Cardiovascular mortality included deaths from ischemic heart disease (410414), heart rhythm disturbance (426, 427), heart failure (428), and cerebrovascular disease/stroke (430-438). Cancers included the ICD-9 codes 140-208. Women specific cancers included cancers of breast (174), cervix (180), ovary (183), vagina (184.0) and endometrium (179,182).
3. Results A total of 7826 deaths were reported during a median follow-up of 12.5 years (range: 5.25–13.0 years). The total mortality rate is 9.17 per 1000 person-years. Baseline characteristics of study participants are described in Table 1. Mean age of participated women was 62.6 ± 5.4 years. The mean BMI was 27.1 ± 5.6 kg/m2. Among a total of 75,359 women, a majority of them had their first menstruation at 12–13 years of age (54%) and their last time of menstruation at 45–54 ages (60%). About 20% women experienced early menarche at age ≤11 years, and 29% women experienced early menopause at age ≤44 years. When
2.4. Potential confounders Socioeconomic and environmental condition may have potential biological associations between menstrual characteristics and mortality risk. Therefore, we obtained the following information as covariates in the statistical models from the baseline questionnaire: baseline age, 51
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Table 1 Characteristics of participants according to age at menarche and age at menopause. Age at menarche, years
≤11 (n = 15,345)
12-13 (n = 40,491)
≥14 (n = 19,523)
Total (n = 75,359)
Age, years Non-Hispanic Whites, % Baseline BMI, km/m2 Current smokers, % Current alcohol drinking, % Educational years > 12, % Physical activity ≥4–5 days/week, % Number of live births ≥3 Current hormone therapy users, % Age at the first birth ≥35 years, % Age at menopause ≥55 years, % History of CVD, % Natural menopause, %
61.9 (5.27) 13,482 (87.9) 28.6 (6.19) 1551 (10.1) 8067 (66.8) 10,347 (67.5) 2215 (14.37) 8954 (58.4) 7582 (49.5) 236 (1.71) 1594 (10.4) 6251 (40.7) 8854 (58.2)
62.5 (5.38) 36,228 (89.4) 27.0 (5.41) 3826 (9.45) 22,299 (68.8) 26,843 (66.3) 5736 (14.21) 23,856 (59.0) 20,013 (49.5) 723 (1.98) 4420 (10.9) 14,536 (35.9) 25,290 (62.9)
63.0 (5.48) 17,083 (87.5) 26.3 (5.09) 1947 (9.97) 10,489 (68.2) 12,382 (64.4)) 2783 (14.21) 11,889 (61.0) 9261 (47.5) 375 (2.11) 2370 (12.1) 6632 (34.0) 12,435 (64.1)
62.5 (5.40) 66,793 (88.6) 27.1 (5.55) 7324 (9.72) 40,855 (68.3) 49,572 (65.8) 10,724 (14.2) 44,699 (59.4) 36,856 (49.0) 1334 (1.96) 8384 (11.1) 27,419 (36.4) 46,579 (62.2)
Age at menopause, years
≤44 (n = 21,537)
45–54 (n = 45,438)
≥55 (n = 8384)
Total (n = 75,359)
Age, years Non-Hispanic Whites, % Baseline BMI, km/m2 Current smokers, % Current alcohol drinking, % Educational years > 12, % Physical activity ≥4–5 days/week, % Number of live births ≥3, % Current hormone therapy users, % Age at the first birth ≥35 years, % Age at menarche ≤11 years, % History of CVD, % Natural menopause, %
62.4 (5.25) 18,777 (87.2) 27.6 (5.71) 2658 (12.3) 10,927 (64.8) 13,315 (61.9) 2786 (12.93) 12,802 (59.6) 11,836 (55.0) 224 (1.15) 5102 (23.7) 8678 (40.3) 4771 (22.2)
62.7 (5.48) 40,469 (89.1) 26.9 (5.46) 4178 (9.20) 25,152 (69.5) 30,356 (66.8) 6582 (14.48) 26,843 (59.2) 20,491 (45.1) 964 (2.35) 8649 (19.0) 15,802 (34.8) 35,511 (78.2)
61.8 (5.23) 7547 (90.0) 27.2 (5.56) 488 (5.82) 4776 (70.4) 5901 (70.4) 1356 (16.19) 5054 (60.4) 4529 (54.1) 146 (1.90) 1594 (19.0) 2939 (35.1) 6297 (79.8)
62.5 (5.40) 66,793 (88.6) 27.1 (5.55) 7324 (9.72) 40,855 (68.3) 49,572 (65.8) 10,724 (14.2) 44,699 (59.4) 36,856 (49.0) 1334 (20.4) 15,345 (11.1) 27,419 (36.4) 45,679 (62.2)
considered in combination, about 6.8% women experienced both early menarche and early menopause. Women with late menarche, but not late menopause, tended to be older and have lower BMI. Around 10% of women reported to be current smokers while 68% be current alcohol user at baseline survey. More than half (59%) of the participants had more than 2 live births. Nearly half (49%) of the participants were hormone therapy users. Age at menarche was significantly associated with risks of all-cause and cancer mortality, but not CVD or women specific cancer mortality
(Table 2). Compared with women with an age at menarche of 12–13 years, women with early menarche had 11% higher risks of all-cause (HR = 1.11, 95% CI 1.05–1.18) and cancer mortality (HR = 1.11, 95% CI 1.01–1.22), and women with late menarche tended to have 7% lower risk of all cause (HR = 0.93, 95% CI 0.88–0.99) and 12% lower risk of cancer mortality (HR = 0.88; 95% CI 0.80–0.96). There was a trend towards decreased risks of all-cause, cancer, and women specific cancer mortalities as increasing age at menarche (all P-values for trend < 0.05).
Table 2 Associations between ages at menarche and all-cause, CVD, cancer, and women specific cancer mortality. Age at menarche
No. of women (mortality ratea)
All-cause mortality ≤11 years 15,345 (9.91) 12–13 years 40,491 (8.89) ≥14 years 19,523 (9.18) P for trend 75,359 (9.17) CVD mortality ≤11 years 15,345 (1.95) 12–13 years 40,491 (1.74) ≥14 years 19,523 (2.04) P for trend 75,359 (1.86) Cancer mortality ≤11 years 15,345 (3.90) 12–13 years 40,491 (3.56) ≥14 years 19,523 (3.41) P for trend 75,359 (3.59) Women specific cancer mortality ≤11 years 15,345 (0.84) 12–13 years 40,491 (0.71) ≥14 years 19,523 (0.70) P for trend 75,359 (0.73)
No. of deaths
HRb
95% CI
HRc
95% CI
HRd
95% CI
1707 4083 2036 7826
1.20 1.00 0.97 < 0.0001
1.13, 1.27 Ref. 0.92, 1.03
1.12 1.00 0.93 < 0.0001
1.06, 1.18 Ref. 0.88, 0.98
1.11 1.00 0.93 < 0.0001
1.05, 1.18 Ref. 0.88, 0.99
335 797 452 1584
1.24 1.00 1.08 0.13
1.09, 1.40 Ref. 0.96, 1.21
1.10 1.00 1.04 0.54
0.96, 1.25 Ref. 0.92, 1.17
1.09 1.00 1.04 0.60
0.96, 1.24 Ref. 0.92, 1.17
671 1633 756 3060
1.15 1.00 0.92 < 0.0001
1.05, 1.26 Ref. 0.85, 1.01
1.11 1.00 0.88 < 0.0001
1.01, 1.22 Ref. 0.8, 0.96
1.11 1.00 0.88 < 0.0001
1.01, 1.22 Ref. 0.80, 0.96
145 326 155 626
1.24 1.00 0.95 0.03
1.02, 1.50 Ref. 0.79, 1.15
1.16 1.00 0.92 0.06
0.95, 1.42 Ref. 0.76, 1.12
1.17 1.00 0.92 0.05
0.96, 1.43 Ref. 0.76, 1.12
a
Deaths per 1000 person per year. Adjusted for age, race, and arm. c Adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone therapy uses, live birth, history of CVD, reasons for menstrual period stopped, and intervention arms. d Additionally adjusted for age at natural menopause. b
52
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Table 3 Associations between ages at menopause and all-cause, CVD, cancer, and women specific cancer mortality. Age at menopause
No. of women (Mortality ratea)
All-cause mortality ≤44 years 21,537 (10.6) 45–54 years 45,438 (8.93) ≥55 years 8384 (7.04) P for trend 75,359 (9.17) CVD mortality ≤44 years 21,537 (2.20) 45–54 years 45,438 (1.79) ≥55 years 8384 (1.35) P for trend 75,359 (1.86) Cancer mortality ≤44 years 21,537 (3.86) 45–54 years 45,438 (3.61) ≥55 years 8384 (2.81) P for trend 75,359 (3.59) Women specific cancer mortality ≤44 years 21,537 (0.63) 45–54 years 45,438 (0.77) ≥55 years 8384 (0.77) P for trend 75,359 (0.73)
No. of deaths
HRb
95% CI
HRc
95% CI
HRd
95% CI
2538 4613 675 7826
1.24 1.00 0.87 < 0.0001
1.18, 1.30 Ref. 0.80, 0.94
1.12 1.00 0.92 < 0.0001
1.05, 1.18 Ref. 0.85, 1.00
1.11 1.00 0.92 < 0.0001
1.05, 1.18 Ref. 0.85, 1.00
530 924 130 1584
1.32 1.00 0.87 < 0.0001
1.18, 1.47 Ref. 0.72, 1.04
1.14 1.00 0.89 0.008
1.01, 1.30 Ref. 0.73, 1.07
1.14 1.00 0.89 0.009
1.01, 1.30 Ref. 0.73, 1.07
927 1863 270 3060
1.10 1.00 0.83 < 0.0001
1.02, 1.19 Ref. 0.73, 0.94
1.03 1.00 0.89 0.10
0.93, 1.13 Ref. 0.78, 1.02
1.02 1.00 0.90 0.14
0.93, 1.12 Ref. 0.79, 1.02
152 400 74 626
0.83 1.00 1.05 0.05
0.69, 1.01 Ref. 0.82, 1.34
0.77 1.00 1.07 0.02
0.62, 0.96 Ref. 0.83, 1.37
0.76 1.00 1.07 0.02
0.61, 0.95 Ref. 0.83, 1.37
a
Deaths per 1000 person per year. Adjusted for age, race, and arm. c Adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone therapy uses, live birth, history of CVD, reasons for menstrual period stopped, and intervention arms. d Additionally adjusted for age at natural menarche. b
Age at menopause was significantly associated with risks of allcause, CVD, and women specific cancer mortality, but not with total cancer mortality (Table 3). Women with early menopause had 11% and 14% higher risk of all-cause (HR = 1.11, 95% CI 1.05–1.18) and CVD mortality (HR = 1.14, 95% CI 1.01–1.30), respectively, compared with women with menopausal age of 45–54 years. There was a trend towards decreased risk of all-cause and CVD mortality with increasing age at menopause (all P-values for trend < 0.05). For women specific cancers, early menopause tended to have a lower mortality risk (HR = 0.76, 95% CI 0.61–0.95); while late menopause likely to have a higher mortality risk (HR = 1.07, 95% CI 0.83–1.37) (P for trend = 0.02). A combination of early menarche and early menopause was significantly associated with higher risks of all-cause and CVD mortality (Table 4 and Fig. 1); while a combination of late menarche and menopause seemed to have lower risks. Compared with women with an
age at menarche of 12–13 years and an age at menopause of 45–54 years, women with both early menarche and early menopause were among those with the highest risk of all-cause mortality (HR = 1.20, 95% CI 1.09–1.32) and CVD mortality (HR = 1.25, 95% CI 1.02–1.54), while those with late menarche and late menopause were at the lowest risk of all-cause mortality (HR = 0.82, 95% CI, 0.71–0.96) and CVD mortality (HR = 0.86, 95% CI 0.62–1.21). The results from the sensitivity analyses by excluding unnaturally menopaused women did not substantially change with regard to significance, direction, and size of the effect. (Supplemental Table 1). The indexes for additive effect, including SI, RERI and AP, showed no excess risk due to interaction. All 95% CIs of synergy index included 1 and 95% CIs of AP and RERI covered the null value; and the interactions were not significant. The combined effects of ages at menarche and menopause seems to be “exactly additivity” on moralities. There were no significant
Table 4 Combined associations of age at menarche and age at menopause with all-cause, CVD, cancer, and women specific cancer mortality. Age at menarche
All-cause mortality ≤11 years 12–13 years ≥14 years CVD mortality ≤11 years 12–13 years ≥14 years Cancer mortality ≤11 years 12–13 years ≥14 years Women specific cancer mortality ≤11 years 12–13 years ≥14 years
Age at menopause, years
P
≤44
45–54
≥55
1.20 (1.09, 1.32) 1.12 (1.04, 1.21) 1.06 (0.97, 1.17)
1.14 (1.06, 1.23) 1.00 (Ref.) 0.94 (0.87, 1.00)
1.07 (0.89, 1.27) 0.94 (0.84, 1.05) 0.82 (0.71, 0.96)
1.25 (1.02, 1.54) 1.12 (0.95, 1.33) 1.21 (0.99, 1.48)
1.10 (0.92, 1.31) 1.00 (Ref.) 1.03 (0.89, 1.21)
0.87 (0.56, 1.36) 0.95 (0.74, 1.23) 0.86 (0.62, 1.21)
1.06 (0.91, 1.24) 1.04 (0.93, 1.18) 0.94 (0.81, 1.10)
1.16 (1.03, 1.30) 1.00 (Ref.) 0.88 (0.78, 0.98)
1.07 (0.82, 1.40) 0.91 (0.76, 1.09) 0.74 (0.58, 0.95)
0.82 (0.56, 1.19) 0.84 (0.64, 1.12) 0.73 (0.50, 1.06)
1.34 (1.05, 1.72) 1.00 (Ref.) 0.94 (0.74, 1.20)
0.99 (0.55, 1.77) 1.19 (0.84, 1.67) 1.08 (0.69, 1.70)
interaction
0.74
0.93
0.61
0.52
The reference groups were chosen as they cover the medians of each menstrual variable. Cox model adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone replacement therapy use, CVD history, number of live births, reasons for menstrual period stopped, and randomized arms. 53
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Fig. 1. Adjusted hazard ratios for combined association of ages at menarche and menopause with all-cause (a) and cause-specific mortality (b–d). Analyses were adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone therapy use, live birth, history of CVD, and intervention arms.
cancer mortality. This finding could be explained by the opposite role of premenopausal sex hormones in CVD [27] and cancer development in women [3]. In the combined analysis, the association between menarche and menopause seems to have an exactly additive effect on all-cause mortality and CVD mortality. On the contrary, for cancer mortality, especially women-specific cancers, the risk of mortality was significantly reduced among women with early menarche and early menopause. Breast and ovarian cancers were the majority cancer types in womenspecific cancers. These two commonly diagnosed cancers in women can be connected by similar genetic mutations and subjected to similar hormonal therapy. It is reported that a later menopause increased the risk of fatal uterine or ovarian cancer [3,28]. Previous study indicated the effects of menarche and menopause on breast cancer risk might not be acting merely by lengthening women's total number of reproductive years [1]. Potential mechanisms include the interplay between endogenous hormones and environmental exposure which may lead to heterogeneous cancer outcome. We found the risks of all-cause mortality and CVD mortality associated with early menarche might be slightly attenuated by a late menopause. This finding could have important clinical implications. Hormone replacement therapy (HRT) or other drugs could be used as effective tools for delaying the menopausal age. Women with early menarche could be benefit from using HRT in order to reduce their CVD mortality risk and even all-cause mortality risk. Some studies suggested that shorter reproductive life span was associated with specific mortality risk [29]. However, using reproductive years as an indicator has an obvious pitfall. In that case, the reproductive years of a woman with early menarche (e.g. 8 years) and early menopause (e.g. 45 years) was equal to woman with normal menarche (e.g. 13 years) and normal menopause (e.g. 50 years). Women with same reproductive years could have different mortality risks according to our findings. Therefore, we suggest that it is important to evaluate both ages at menarche and menopause rather than to consider each variable separately in order to assess the mortality risk. Our study findings are strengthened by our ability to adjust for most known confounders, e.g. hormone replacement therapy and unnatural
interactions between ages at menarche and menopause with morality risk (Table 4). In stratified analyses by ages at menarche and menopause, the risks of all-cause mortality and CVD mortality associated with early menarche might be slightly attenuated by the effect of late menopause. On the other hand, women with early menopause and late menarche still had a higher risk of all-cause mortality and cancer mortality (Table 5). Additionally, the risk of women specific cancer mortality reduced by half (HR = 0.49, 95% CI 0.32–0.77) among women with early menarche and early menopause, compared to the reference group. 4. Discussion In this large-scale cohort study, we found that early menarche was associated with higher total mortality and cancer mortality. Early menopause was associated with higher risk of mortality from all causes and CVD, but with lower risk of mortality from women specific cancers. Early menarche and early menopause seemed to have an exactly additivity effect on all-cause mortality. Our findings suggest that both age at menarche and age at menopause could be key indicators for all-cause and CVD mortality among women, although further investigations in mechanistic studies and well-designed prospective studies are needed. To our knowledge, the present study was one of the largest studies to date including the largest population of U.S. postmenopausal women. Consistent with most of previous studies [6,7,11,26], we found a significant association between early age at menarche and increased total mortality risk. As for CVD and cancer mortality, the associations were reported to be significant only among specific populations, e.g. younger women and women without major CVD risk factors [11]. In this study, we found a significant association with CVD mortality only in age, race, and intervention arm-adjusted model. However, this association became non-significant when further adjusting with smoking and alcohol drinking status, hormone therapy uses and other covariates. This implicates that age at menarche could be a robust indicator for risk of total mortality rather than CVD or cancer mortality. We also observed a negative association of ages at menopause with CVD and allcause mortality, while a positive association with women specific 54
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Table 5 Stratified analysis of combined associations of age at menarche and age at menopause with all-cause and cause-specific mortalities. Age at menopause, years
No. of women (Mortality ratea)
No. of deaths
HR (95% CI)b
P
≤44 45–54 ≥55 ≤44 45–54 ≥55 ≤44 45–54 ≥55
5102 (11.06) 8649 (9.72) 1594 (7.39) 11214 (10.23) 24857 (8.63) 4420 (7.05) 5221 (10.78) 11932 (8.98) 2370 (6.78)
625 948 134 1284 2443 356 629 1222 185
1.06 1.00 0.93 1.14 1.00 0.94 1.11 1.00 0.89
(0.94, (Ref.) (0.77, (1.05, (Ref.) (0.84, (0.99, (Ref.) (0.76,
1.20)
0.18
1.12) 1.24)
0.0004
1.05) 1.24)
0.009
≤44 45–54 ≥55 ≤44 45–54 ≥55 ≤44 45–54 ≥55
5102 (2.32) 8649 (1.88) 1594 (1.16) 11214 (2.01) 24857 (1.68) 4420 (1.37) 5221 (2.52) 11932 (1.95) 2370 (1.47)
131 183 21 252 476 69 147 265 40
1.20 1.00 0.78 1.09 1.00 0.96 1.20 1.00 0.84
(0.92, (Ref.) (0.49, (0.91, (Ref.) (0.74, (0.95, (Ref.) (0.59,
≤44 45–54 ≥55 12–13 years ≤44 45–54 ≥55 ≥14 years ≤44 45–54 ≥55 Women specific cancer mortality ≤11 years ≤44 45–54 ≥55 12–13 years ≤44 45–54 ≥55 ≥14 years ≤44 45–54 ≥55
5102 (3.93) 8649 (4.02) 1594 (3.14) 11214 (3.86) 24857 (3.55) 4420 (2.87) 5221 (3.79) 11932 (3.43) 2370 (2.49)
222 392 57 484 1004 145 221 467 68
0.85 1.00 0.94 1.10 1.00 0.91 1.03 1.00 0.85
(0.69, (Ref.) (0.71, (0.97, (Ref.) (0.76, (0.85, (Ref.) (0.66,
5102 (0.67) 8649 (0.97) 1594 (0.66) 11214 (0.64) 24857 (0.72) 4420 (0.81) 5221 (0.58) 11932 (0.73) 2370 (0.77)
38 95 12 80 205 41 34 100 21
0.49 1.00 0.70 0.99 1.00 1.19 0.66 1.00 1.18
(0.32, (Ref.) (0.38, (0.74, (Ref.) (0.85, (0.43, (Ref.) (0.73,
Age at menarche All-cause mortality ≤11 years 12–13 years ≥14 years CVD mortality ≤11 years 12–13 years ≥14 years Cancer mortality ≤11 years
trend
1.04) 1.58)
0.06
1.24) 1.31)
0.31
1.24) 1.52)
0.05
1.18) 1.03)
0.25
1.24) 1.26)
0.05
1.08) 1.24)
0.26
1.10) 0.77)
0.07
1.28) 1.35)
0.45
1.67) 1.04)
0.05
1.90)
a
Deaths per 1000 person per year. Cox model adjusted for age, race, BMI, current smoking status, alcohol drinking status, physical exercise, education, hormone replacement therapy use, CVD history, number of live births, reasons for menstrual period stopped, and randomized arms. b
menopause. In addition, due to the large sample size in the PLCO cohort, majority of our subgroups classified by age categories of menarche and menopause still has sufficient power to test the significant of associations. Our study has several limitations. First, we couldn’t examine the relationship separately by subgroups as defined by each CVD or other cancer subtypes, due to limited number of specific-deaths. Second, because the exposures (age at menarche and age at menopause) were collected retrospectively at baseline, there could be potential biases due to possible misclassification. Third, although we adjusted for many potential confounders, we cannot rule out residual confounding by factors that have not been measured. Finally, the original data for age at menarche and menopause were categorical variables. We have no sufficient statistic power to evaluate the possible non-linearity.
Contributors Xi Zhang designed and conducted the research, contributed to data analysis, drafted the manuscript, interpretation of the results, and is responsible for the final content of the manuscript. Luyang Liu contributed to data collection, and interpretation, and revision of the manuscript. Fengju Song contributed to data collection, and interpretation, and revision of the manuscript. Yiqing Song contributed to study design, interpretation and revision of the manuscript. Hongji Dai designed and conducted the research, contributed to data analysis, drafted the manuscript, interpretation of the results, and is responsible for the final content of the manuscript. All authors read and approved the final version.
5. Conclusions In summary, the current study provides prospective evidence of the combined association between ages at menarche and menopause with all-cause mortality risk in a large population of U.S. postmenopausal women. Better understanding the interplay between endogenous hormone exposure and CVD and cancer development may elucidate novel mechanistic pathways in survival. We suggest evaluate both ages at menarche and menopause rather than to consider each variable separately in order to assess the mortality risk.
Funding This work was supported by Natural Science Foundation of China (81703238) and Tianjin Science & Technology Foundation for Selected Overseas Scholar (2016016).
55
Maturitas 130 (2019) 50–56
X. Zhang, et al.
Ethical statement
[9] N.T. Mueller, A.O. Odegaard, M.D. Gross, W.P. Koh, J.M. Yuan, M.A. Pereira, Age at menarche and cardiovascular disease mortality in Singaporean Chinese women: the Singapore Chinese Health Study, Ann. Epidemiol. 22 (10) (2012) 717–722. [10] K. Murakami, H. Metoki, M. Satoh, K. Asayama, M. Hosaka, A. Matsuda, R. Inoue, M. Tsubota-Utsugi, T. Murakami, K. Nomura, M. Kikuya, Y. Imai, T. Ohkubo, Menstrual factors and stroke incidence in Japanese postmenopausal women: the Ohasama Study, Neuroepidemiology 47 (2) (2016) 109–116. [11] L. Yang, L. Li, I.Y. Millwood, S.A.E. Peters, Y. Chen, Y. Guo, Z. Bian, X. Chen, L. Chen, S. Feng, S. Lv, Z. Pang, M. Woodward, Z. Chen, Age at menarche and risk of major cardiovascular diseases: evidence of birth cohort effects from a prospective study of 300,000 Chinese women, Int. J. Cardiol. 227 (2017) 497–502. [12] M.W. Lundblad, B.K. Jacobsen, Is age at menarche associated with total mortality? The Tromso Study, Int. J. Womens Health 10 (2018) 203–209. [13] D. Gong, J. Sun, Y. Zhou, C. Zou, Y. Fan, Early age at natural menopause and risk of cardiovascular and all-cause mortality: a meta-analysis of prospective observational studies, Int. J. Cardiol. 203 (2016) 115–119. [14] B.K. Jacobsen, I. Heuch, G. Kvale, Age at natural menopause and all-cause mortality: a 37-year follow-up of 19,731 Norwegian women, Am. J. Epidemiol. 157 (10) (2003) 923–929. [15] T. Muka, C. Oliver-Williams, S. Kunutsor, J.S. Laven, B.C. Fauser, R. Chowdhury, M. Kavousi, O.H. Franco, Association of age at onset of menopause and time since onset of menopause with cardiovascular outcomes, intermediate vascular traits, and all-cause mortality: a systematic review and meta-analysis, JAMA Cardiol. 1 (7) (2016) 767–776. [16] L. Jaspers, M. Kavousi, N.S. Erler, A. Hofman, J.S. Laven, O.H. Franco, Fertile lifespan characteristics and all-cause and cause-specific mortality among postmenopausal women: the Rotterdam Study, Fertil. Steril. 107 (2) (2017) 448–456 e1. [17] A.H. Shadyab, C.A. Macera, R.A. Shaffer, S. Jain, L.C. Gallo, M.L. Gass, M.E. Waring, M.L. Stefanick, A.Z. LaCroix, Ages at menarche and menopause and reproductive lifespan as predictors of exceptional longevity in women: the Women’s Health Initiative, Menopause 24 (1) (2017) 35–44. [18] F.R. Day, D.J. Thompson, H. Helgason, D.I. Chasman, H. Finucane, P. Sulem, et al., Genomic analyses identify hundreds of variants associated with age at menarche and support a role for puberty timing in cancer risk, Nat. Genet. 49 (6) (2017) 834–841. [19] C. He, P. Kraft, C. Chen, J.E. Buring, G. Pare, S.E. Hankinson, S.J. Chanock, P.M. Ridker, D.J. Hunter, D.I. Chasman, Genome-wide association studies identify loci associated with age at menarche and age at natural menopause, Nat. Genet. 41 (6) (2009) 724–728. [20] L. Stolk, J.R. Perry, D.I. Chasman, C. He, M. Mangino, P. Sulem, et al., Meta-analyses identify 13 loci associated with age at menopause and highlight DNA repair and immune pathways, Nat. Genet. 44 (3) (2012) 260–268. [21] H.B. Nichols, A. Trentham-Dietz, J.M. Hampton, L. Titus-Ernstoff, K.M. Egan, W.C. Willett, P.A. Newcomb, From menarche to menopause: trends among US Women born from 1912 to 1969, Am. J. Epidemiol. 164 (10) (2006) 1003–1011. [22] J.K. Gohagan, P.C. Prorok, P. Greenwald, B.S. Kramer, The PLCO cancer screening trial: background, goals, organization, operations, results, Rev. Recent Clin. Trials 10 (3) (2015) 173–180. [23] A.B. Miller, R. Feld, R. Fontana, J.K. Gohagan, I. Jatoi, W. Lawrence Jr., A. Miller, K.P. Proro, A. Rajput, M. Sherman, G. Welch, P. Wright, S. Yurgalevitch, P. Albertsen, Changes in and impact of the death review process in the prostate, lung, colorectal and ovarian (PLCO) cancer screening trial, Rev. Recent Clin. Trials 10 (3) (2015) 206–211. [24] A.B. Miller, S. Yurgalevitch, J.L. Weissfeld, L.C. Prostate, T. Ovarian Cancer Screening trial project, death review process in the prostate, lung, colorectal and ovarian (PLCO) cancer screening trial, Control. Clin. Trials 21 (Suppl. 6) (2000) 400S–406S. [25] M.J. Knol, T.J. VanderWeele, R.H. Groenwold, O.H. Klungel, M.M. Rovers, D.E. Grobbee, Estimating measures of interaction on an additive scale for preventive exposures, Eur. J. Epidemiol. 26 (6) (2011) 433–438. [26] R. Lakshman, N.G. Forouhi, S.J. Sharp, R. Luben, S.A. Bingham, K.T. Khaw, N.J. Wareham, K.K. Ong, Early age at menarche associated with cardiovascular disease and mortality, J. Clin. Endocrinol. Metab. 94 (12) (2009) 4953–4960. [27] J.E. Manson, T.K. Woodruff, Reproductive health as a marker of subsequent cardiovascular disease: the role of estrogen, JAMA Cardiol. 1 (7) (2016) 776–777. [28] M.E. Ossewaarde, M.L. Bots, A.L. Verbeek, P.H. Peeters, Y. van der Graaf, D.E. Grobbee, Y.T. van der Schouw, Age at menopause, cause-specific mortality and total life expectancy, Epidemiology 16 (4) (2005) 556–562. [29] X. Wu, H. Cai, A. Kallianpur, Y.T. Gao, G. Yang, W.H. Chow, H.L. Li, W. Zheng, X.O. Shu, Age at menarche and natural menopause and number of reproductive years in association with mortality: results from a median follow-up of 11.2 years among 31,955 naturally menopausal Chinese women, PLoS One 9 (8) (2014) e103673.
We confirm that the study was approved by the Institutional Review Board of the National Cancer Institute and the participating centers (https://www.ncicirb.org/). Each participant provided written consent upon enrollment. Provenance and peer review This article has undergone peer review. Research data (data sharing and collaboration) The data set is available via https://biometry.nci.nih.gov/cdas/ datasets/plco/. Declaration of Competing Interest The authors declare that they have no conflict of interest. Acknowledgements We would like to thank the National Cancer Institute (NCI) for access to data collected by the PLCO study group. We also thank Jerome Mabie of Information Management Services, Inc. for data management and support. The statements contained herein are solely those of the authors and do not represent or imply concurrence or endorsement by NCI. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.maturitas.2019.10. 009. References [1] 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) (2012) 1141–1151. [2] M.M. Hassan, G. Botrus, R. Abdel-Wahab, R.A. Wolff, D. Li, D. Tweardy, A.T. Phan, E. Hawk, M. Javle, J.S. Lee, H.A. Torres, A. Rashid, R. Lenzi, H.M. Hassabo, Y. Abaza, A.S. Shalaby, S. Lacin, J. Morris, Y.Z. Patt, C.I. Amos, S.A. Khaderi, J.A. Goss, P.K. Jalal, A.O. Kaseb, Estrogen replacement reduces risk and increases survival times of women with hepatocellular carcinoma, Clin. Gastroenterol. Hepatol. 15 (11) (2017) 1791–1799. [3] A.L. Shafrir, A. Babic, R.M. Tamimi, B.A. Rosner, S.S. Tworoger, K.L. Terry, Reproductive and hormonal factors in relation to survival and platinum resistance among ovarian cancer cases, Br. J. Cancer 115 (11) (2016) 1391–1399. [4] K. Tamakoshi, H. Yatsuya, A. Tamakoshi, Early age at menarche associated with increased all-cause mortality, Eur. J. Epidemiol. 26 (10) (2011) 771–778. [5] E.K. Bjelland, S. Hofvind, L. Byberg, A. Eskild, The relation of age at menarche with age at natural menopause: a population study of 336 788 women in Norway, Hum. Reprod. 33 (6) (2018) 1149–1157. [6] D. Canoy, V. Beral, A. Balkwill, F.L. Wright, M.E. Kroll, G.K. Reeves, J. Green, B.J. Cairns, Age at menarche and risks of coronary heart and other vascular diseases in a large UK cohort, Circulation 131 (3) (2015) 237–244. [7] B.K. Jacobsen, K. Oda, S.F. Knutsen, G.E. Fraser, Age at menarche, total mortality and mortality from ischaemic heart disease and stroke: the Adventist Health Study, 1976–88, Int. J. Epidemiol. 38 (1) (2009) 245–252. [8] M.A. Merritt, E. Riboli, N. Murphy, M. Kadi, A. Tjonneland, A. Olsen, et al., Reproductive factors and risk of mortality in the European Prospective Investigation into Cancer and Nutrition; a cohort study, BMC Med. 13 (2015) 252.
56