Study population
Participants were residents of two communities in Beijing and were part of the Chinese Multi-provincial Cohort Study (CMCS). The CMCS is a multicenter, prospective, population-based cohort study on the determinants of CVD. Details of the study have been published previously [13]. The baseline survey began in 1992. A standardized questionnaire modified according to the WHO-MONICA protocol for risk factor survey was used to collect information on demographic characteristics, smoking habit, personal medical history, and medical therapy [14]. Participants signed the written informed consents, completed questionnaires, and underwent physical measurements as well as phlebotomy. The ethics committee of Beijing An Zhen Hospital approved the study.
A total of 2116 women aged 35–64 years who were free of CVD history at baseline were included in the present study. All participants had a follow-up visit every 1–2 years from the baseline examination through the first CVD event, death, or the end of follow-up (2018). After excluding women who were lost to follow-up, 2104 women were eligible for inclusion. Figure 1 shows the study profile.
Exposure
Women reported their natural menopause status (premenopausal or postmenopausal), and if postmenopausal, their age at natural menopause. All women were in postmenopausal status until the end of follow-up. Age at menopause was categorized as less than 45 years (early menopause), 45–49 years (relatively early menopause), 50–51 years (reference), 51 years or older (relatively late menopause). When conducting analyses of the combined effect of age at menopause and cardiovascular risk factors at baseline, age at menopause was also categorized into three different menopausal age groups (< 50 years, 50–51 years, and > 51 years).
Using an existing menopause staging system [15], baseline menopausal status was categorized as reproductive (> 3 years before menopause), menopausal transition (≤ 3 years before menopause and ≤ 0 years since menopause)/perimenopause (≤ 3 years before menopause and ≤ 1 years since menopause), early postmenopause (> 0/1 years and ≤ 6 years since menopause), and late postmenopause (> 6 years since menopause).
Outcome
The primary study endpoint was death or the occurrence of a first CVD event (including fatal CVD or non-fatal CVD), defined as a composite outcome of incident CHD or stroke (including ischemic stroke or hemorrhagic stroke). Additionally, a composite endpoint of incident CHD or ischemic stroke was also defined as ischemic CVD in the analysis. Acute CVD events were ascertained according to the diagnostic criteria of the World Health Organization-MONICA protocol [16]. Coronary events were diagnosed based on symptoms, electrocardiography recordings, serum myocardial enzymes, autopsy findings, and history of CHD. Stroke events were diagnosed as rapidly developing signs of focal (or global) disturbances in cerebral function lasting more than 24 h (unless interrupted by surgery or death) with no apparent non-vascular cause [17]. All suspected CVD events were reviewed by a team of physicians using data from the original medical records. All causes of death were registered based on the death certificate issued by the physician and were double-checked by researchers using the original medical or death records [18].
Covariates
The following factors were included in the analyses as covariates: age at baseline, family history of CVD, white blood cell count, time-varying menopause, use of oral estrogen due to menopause, smoking status (regular smoker, occasional smoker, never smoker, or quit smoking for > 1 year), body mass index (BMI), systolic blood pressure (SBP), diastolic blood pressure (DBP), fasting blood glucose (FG), total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C). Optimal risk factor levels were defined as never smoker or quit smoking for > 1 year, BMI ≤ 24 kg/m2, SBP < 130 mmHg, DBP < 80 mmHg, FG < 6.1 mmol/L, LDL-C < 3.37 mmol/L, TC < 5.17 mmol/L, TG < 1.70 mmol/L, and HDL-C > 1.04 mmol/L. Non-optimal risk factor levels were defined as occasional smoker, BMI 24–28 kg/m2, SBP 130–140 mmHg, DBP 80–90 mmHg, FG 6.1–7.0 mmol/L, LDL-C 3.37–4.14 mmol/L, TC 5.17–6.20 mmol/L, TG 1.70–2.26 mmol/L, and HDL-C 0.91–1.04 mmol/L. Elevated risk factor levels were defined as regular smoker, BMI > 28 kg/m2, SBP ≥ 140 mmHg, DBP ≥ 90 mmHg, FG ≥ 7.0 mmol/L, LDL-C ≥ 4.14 mmol/L, TC ≥ 6.20 mmol/L, TG ≥ 2.26 mmol/L, and HDL-C ≤ 0.91 mmol/L.
Statistical analysis
Baseline continuous variables were compared using one-way analysis of variance and categorical variables were compared using the χ2 test. Cumulative incidence of death and CVD events were estimated using the Kaplan–Meier method. Multivariate Cox regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for death and CVD events associated with menopause. For women who died or had first occurrence of a CVD event, follow-up time was calculated as their age at death or first event minus their entry age; for women without death and CVD endpoint events, follow-up time was identified as their age at last follow-up minus their entry age. Menopause at age 50–51 years was used as the reference category. In the fully adjusted models, HRs and 95% CIs were adjusted for: time-varying covariates including menopause, use of oral estrogen owing to menopause, smoking, BMI, SBP, DBP, FG, TC, TG, LDL-C, and HDL-C; as well as baseline covariates including age, family history of CVD, and white blood cell count. Sensitivity analysis was conducted after excluding cases of premature menopause (age < 40 years) to test the robustness of the findings.
To examine whether earlier menopause was associated with greater risk of death or first occurrence of a CVD event when combined with elevated risk factors compared with later menopause combined with optimal risk factor levels, we stratified the analyses using a conventional risk factor grade at baseline based on menopausal age groups: menopause at age < 50, 50–51, and > 51 years combined with risk factor grading at all risk factor levels (optimal, ≥ 1 non-optimal risk factor, and ≥ 1 elevated risk factor). Women with menopause at age 50–51 years and all risk factors at optimal levels were regarded as the reference group. In the combined effects models, HRs and 95% CIs were adjusted for: time-varying menopause and use of oral estrogen owing to menopause, as well as baseline age, family history of CVD, and white blood cell count. All statistical analyses were performed using IBM SPSS version 22.0 (IBM Corp., Armonk, NY, USA).