Article Text
PDF
Abstract
Objective This study aimed to examine the association between lifetime oestrogen exposure and ischaemic heart disease (IHD), based on the hypothesis that higher lifetime oestrogen exposure is linked to lower cardiovascular risk.
Methods In 2004–2008, lifetime cumulative exposure to reproductive factors was assessed among postmenopausal females from the China Kadoorie Biobank using reproductive lifespan (RLS), endogenous oestrogen exposure (EEE) and total oestrogen exposure (TEE). EEE was calculated by subtracting pregnancy-related and contraceptive use duration from RLS, while TEE by adding up the same components except for lactation. Incident IHD during follow-up (2004–2015) was identified. Stratified Cox proportional hazards models estimated the HRs and 95% CIs of IHD for RLS, EEE and TEE.
Results Among 118 855 postmenopausal females, 13 162 (11.1%) developed IHD during a median follow-up of 8.9 years. The IHD incidence rates were 13.0, 12.1, 12.5, 13.8 per 1000 person-years for RLS Q1–Q4, 15.8, 12.6, 11.3, 12.1 per 1000 person-years for EEE Q1–Q4 and 13.7, 12.3, 12.2, 13.4 per 1000 person-years for TEE Q1–Q4. The highest quartile (Q4) of RLS and TEE were associated with lower risks of IHD (adjusted HR (aHR) 0.95, 95% CI 0.91 to 1.00 and 0.92, 95% CI 0.88 to 0.97, respectively) compared with the lowest quartile (Q1). Longer EEE showed progressively lower risks of incident IHD (aHR 0.93, 95% CI 0.88 to 0.97; 0.88, 95% CI 0.84 to 0.93; 0.87, 95% CI 0.83 to 0.92 for Q2–Q4 vs Q1).
Conclusions Longer RLS, TEE and EEE were associated with lower risks of IHD among Chinese postmenopausal females.
- coronary artery disease
- epidemiology
Data availability statement
Data are available upon reasonable request. This research has been conducted using the China Kadoorie Biobank (CKB) resource (www.ckbiobank.org). Publication of results does not require or imply approval by the membership of the CKB Collaborative Group. The raw CKB data underlying this article can be accessed via https://www.ckbiobank.org/CKBDataAccess, following the institution’s data-access policies. Preliminary event adjudication data are not publicly available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Female-specific risk factors for ischaemic heart disease (IHD), such as reproductive events and oestrogen exposure across females’ reproductive lifespan (RLS), were largely understudied.
Very few studies have combined a series of reproductive factors to reflect lifetime oestrogen levels, let alone to explore their impact on the development of IHD.
By far, little evidence is available linking cumulative lifetime reproductive factors to incident IHD among postmenopausal females.
WHAT THIS STUDY ADDS
In this prospective cohort study including 118 855 postmenopausal females, we found that longer RLS, endogenous oestrogen exposure and total oestrogen exposure were associated with lower risks of IHD.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
These findings present valuable opportunities to understand the lifetime cumulative effects of reproductive factors on IHD in postmenopausal females, and how these may contribute to female-specific risk prediction and further prompt actionable recommendations for primary and secondary IHD prevention for females.
Introduction
Ischaemic heart disease (IHD) is the most prevalent cardiovascular disease (CVD), as there were an estimated 197 million IHD cases in 2019, and the number of IHD cases is continually rising globally.1 2 For another, health inequalities can manifest between females and males in the prevalence and adverse health outcomes of IHD and have attracted much attention in research and clinical practice.3–5 Some sex-specific risk factors for CVDs, specifically for IHD, such as reproductive events and oestrogen exposure across females’ reproductive lifespan (RLS) were largely understudied.1 4 There is evidence that females experience higher incidence and mortality of IHD after menopause because of the postmenopausal decline in endogenous oestrogen, which has been identified as a protective factor for IHD.6–8
Measuring the cumulative effects of various reproductive factors might allow for a robust and precise understanding of oestrogen-related health issues among postmenopausal females. Previous studies have widely used the RLS (the difference between age at menarche and menopause) to assess oestrogen exposure throughout a female’s life.4 9 10 Some researchers have also adopted composite indicators such as endogenous oestrogen exposure (EEE) and total oestrogen exposure (TEE), which were used to estimate cumulative exposure to oestrogen over females’ lifetime.9 11 According to a recent systematic review and meta-analysis by Mishra et al, only four studies adopted EEE as a lifetime reproductive factor and only two used TEE to explore the association of such cumulative reproductive factors with all-cause and cardiovascular mortality.9
Are higher levels of lifetime oestrogen exposure associated with lower CVD risk? This study investigated the lifetime cumulative effect of reproductive factors, as measured by RLS, EEE and TEE, on incident IHD among postmenopausal females from the China Kadoorie Biobank (CKB) study (figure 1).
Visual overview of key study findings.
Methods
Study design and data collection
Details of the CKB study have been previously described.12–14 Between 25 June 2004 and 15 July 2008 (baseline), 512 715 individuals aged 30–79 years without any significant physical or mental conditions were recruited from five rural and five urban survey sites in China. Data on demographic and socio-economic factors, health-related behaviours, pre-existing diseases, related medications and various other factors were collected using door-to-door open-ended interviews by the trained local survey team. Data on reproductive history on age at menarche and menopause, number of live births, number of stillbirths, number of miscarriages or terminations, lactation duration and contraceptive pill (CP) use were also collected.
Composite indicators of lifetime cumulative exposure to reproductive factors
In the study, RLS was defined as the age at menarche subtracted from the age at menopause:
The equations for EEE and TEE were then constructed based on previous studies,11 15 16 by collectively considering the fluctuations of exogenous and endogenous oestrogen levels across a female’s RLS.11 In this study, we constructed EEE equation based on the assumption that pregnancy, stillbirth, miscarriage and CP use reduce endogenous oestrogen generation, while lactation duration increases endogenous oestrogen generation transiently but then results in a period of low oestrogen levels. Therefore, EEE was calculated by subtracting the following from RLS: 9 months for each live birth, 7 months for each stillbirth, lifetime lactation duration, 3 months for each miscarriage or termination and the CP use duration:
The calculation for TEE was built based on the assumption that the number of pregnancies, terminations and CP use contribute to a higher level of oestrogen, whereas breast feeding is excluded from TEE because it represents a period when oestrogen levels are low.11 Accordingly, TEE calculates a woman’s total exposure to oestrogen by adding up the estimated increases in oestrogen due to pregnancy, stillbirth, miscarriage and CP use.
Ascertainment of IHD
IHD was confirmed through electronic linkage with disease registries and the new national health insurance claim databases during follow-up (2004–2015). The IHD event was coded based on the 10th International Classification of Diseases (ICD-10) by trained medical staff, who were blinded to participants’ baseline information. The procedure of IHD diagnosis (ICD-10: I20-I25) and treatment of participants were also available for further review and confirmed by a clinical research fellow.
Assessment of covariates
Sociodemographic factors included age at baseline, marital status, residence, education, occupation and household income. Regarding lifestyle factors, smoking status, secondhand smoking status and alcohol intake were categorised dichotomously. Body mass index was calculated as weight (kg) divided by the square of standing height (m) and classified into four levels: underweight (<18.5 kg/m2), normal weight (18.5–23.9 kg/m2), overweight (24.0–27.9 kg/m2) and obesity (≥28 kg/m2). Normal waist circumference was defined as <85 cm for females. Physical activity in metabolic equivalents of task was recorded as continuous variables. As for medical factors, anticoagulation therapy, hypolipidemic therapy, history of stroke, diabetes and hypertension were recorded as binary variable. Except for age and physical activity, the rest of the covariates were categorical in analysis.
In this study, postmenopausal females with an age at menarche of 9–18 years, and with an age at menopause of 40 years or over were included. Those with prior IHD or coronary heart disease (CHD) or a history of cancer, hysterectomy or removal of any breast lump and ovary at baseline were excluded. Females who had missing data on covariates were excluded. Moreover, females with an abnormal EEE value (EEE ≤0) were also excluded. Overall, 118 855 females were included in the final analysis (figure 2).
Flow chart of selected participants. CHD, coronary heart disease; EEE, endogenous oestrogen exposure; IHD, ischaemic heart disease.
Statistical analysis
All continuous variables in the analyses of baseline characteristics were highly skewed in tests for normality of distribution and were reported as medians and IQRs. A Wilcoxon rank-sum test was used to calculate differences accordingly. Numbers and percentages were presented for categorical variables, and the differences were captured using χ2 tests. The incidence rate of IHD was calculated by dividing the number of new cases within the specified time period by the total number of person-years at risk for the population, and was expressed as the number of new cases per 1000 person-years at risk.
A multivariable Cox proportional hazards model was used to estimate the HRs and 95% CIs of incident IHD associated with indicators for lifetime cumulative exposure to reproductive factors, namely RLS, EEE and TEE (in quartiles). The proportional hazards assumption was examined by Schoenfeld test and Schoenfeld residual plots. The survival time was determined as the period from the baseline interview to the date of the IHD event, loss to follow-up or 31 December 2015 (the end point), whichever came first. Four multivariable models were hierarchically established to illustrate possible confounders: model 1 was adjusted for age; model 2 was adjusted for sociodemographic factors; model 3 was further adjusted for lifestyle factors; model 4 was further adjusted for medical factors based on model 3. Age-stratified analyses were further performed using a multivariable Cox regression model. Sensitivity analyses were conducted by (1) excluding participants taking aspirin and statins and (2) excluding participants with related chronic diseases including stroke, diabetes and hypertension. Associations of individual reproductive factors, including age at menarche, age at menopause, number of live births, number of stillbirths, number of miscarriages or terminations, lifetime lactation duration, CP use duration, with the risk of incident IHD were also assessed using multivariable Cox regression models.
This study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology recommendations for reporting cohort studies. All statistical analyses were conducted using SAS V.9.4 and R V.4.2.1. P values were two-tailed and considered nominally statistically significant at p<0.05.
Results
Baseline characteristics of postmenopausal females by incident IHD status are presented in table 1. Of the 118 855 postmenopausal females included in the study, 13 162 (11.1%) developed IHD during the follow-up and 105 693 (88.9%) did not. The median age at baseline for all participants was 58.2 (IQR 53.9–64.8) years. With regard to reproductive characteristics, the median number of pregnancies and live births for all participants were 4.0 (IQR 3.0–5.0) and 3.0 (IQR 2.0–4.0), respectively. The median RLS for females with incident IHD was 34.0 (IQR 31.0–36.0) years, whereas for non-IHD females, it was 33.0 (IQR 31.0–36.0) years. Compared with those who had no IHD, females with IHD were more likely to have shorter EEE (median EEE for those with IHD 26.6 years, IQR 22.5–30.3 vs median EEE for those without IHD 27.3 years, IQR 23.6–30.7), with a p value of Wilcoxon rank-sum test <0.05. The median TEE was 32.3 years (IQR 29.0–35.5) for IHD group, while 32.4 years (IQR 29.3–35.3) for non-IHD group. More baseline characteristics of participants analysed by quartiles of RLS, EEE and TEE are shown in online supplemental eTables 1–3.
Supplemental material
Baseline and reproductive characteristics of included 118 855 postmenopausal females by incident IHD
During a median follow-up period of 8.9 (IQR 8.0–10.1) years, the incidence rate of total IHD was 12.9 per 1000 person-years (figure 2 and online supplemental eTable 5). In general, higher IHD incidence rates were observed among females who had the longest RLS, and the shortest EEE and TEE. Postmenopausal females exposed to RLS Q2 (31–32.9 years), EEE Q3 (27.3–30.5 years) and TEE Q3 (32.4–35.2 years) had the lowest IHD incidence rates, which were 12.1, 11.3 and 12.2 per 1000 person-years, respectively.
The association of lifetime cumulative reproductive indicators, namely RLS, EEE and TEE, with the risk of IHD among postmenopausal females is presented in figure 3 and online supplemental eTable 4. In the fully adjusted model, compared with RLS Q1, RLS Q4 had a slightly lower risk of IHD (adjusted HR (aHR) 0.95, 95% CI 0.91 to 1.00). A similar association was found between TEE and the risk of incident IHD, where postmenopausal females who experienced the highest quartile of TEE had a decreased risk of IHD (aHR 0.92, 95% CI 0.88 to 0.97). Compared with postmenopausal females with the shortest EEE (Q1), those with a longer EEE showed lower risks of incident IHD (EEE Q2: aHR 0.93, 95% CI 0.88 to 0.97; Q3: aHR 0.88, 95% CI 0.84 to 0.93; Q4: aHR 0.87, 95% CI 0.83 to 0.92).
RLS, EEE, TEE and the risk of IHD among postmenopausal females: multivariable Cox regression. *P<0.05 with a negative association. Incidence rate was expressed in 1000 person-years. Model was adjusted for sociodemographic characteristics (age at baseline, marital status, residence, education, occupation and household income), lifestyle factors (body mass index, waist circumference, smoking status, secondhand smoking, alcohol intake and physical activity in metabolic equivalent hours/day) and medical covariates (history of anticoagulation therapy, hypolipidemic therapy, stroke, diabetes and hypertension). EEE, endogenous oestrogen exposure; IHD, ischaemic heart disease; RLS, reproductive lifespan; TEE, total oestrogen exposure.
Associations of each reproductive factor with the risk of incident IHD are also mentioned in table 2. In the fully adjusted model, age at menopause (aHR 0.95, 95% CI 0.91 to 1.00) and CP use duration (aHR 0.97, 95% CI 0.95 to 0.99) were weakly correlated with IHD risk, while age at menarche (aHR 0.96, 95% CI 0.86 to 1.07) showed no significance. Increased risk of IHD appeared with other reproductive factors, including the number of live births (aHR 1.04, 95% CI 1.02 to 1.06), number of stillbirths (aHR 1.14, 95% CI 1.10 to 1.17), number of miscarriages or terminations (aHR 1.02, 95% CI 1.00 to 1.03) and lifetime lactation duration (aHR 1.02, 95% CI 1.01 to 1.03).
Association of reproductive factors and the risk of ischaemic heart disease among postmenopausal females: multivariable Cox regression
Sensitivity analyses were performed with (1) exclusion of individuals taking aspirin and statins (n=117 215) and (2) exclusion of individuals with chronic diseases including diabetes, stroke and hypertension (n=60 988) (online supplemental eTable 6).
Further results on age-stratified analyses assessing associations between lifetime cumulative exposure to reproductive factors and risk of incident IHD are presented in online supplemental eTable 7.
Discussion
This study found that longer RLS (≥36 years), EEE (≥23.5 years) and TEE (≥35.3 years) were associated with a lower risk of IHD, supporting the hypothesis that higher lifetime cumulative oestrogen levels may indicate a lower risk of IHD among postmenopausal females. Our findings were generally consistent with the biological mechanism of oestrogen effect on CVDs in postmenopausal females.17 18 Theoretically, oestrogen deficiency is associated with significant shifts in slower lipid metabolism, higher visceral fat storage, inflammation and atherogenic dyslipidaemia, potentially causing a higher risk of CVDs.18 Moreover, loss of oestrogen may directly lead to physiological dysfunction of vasodilators, vasoconstrictors, cellular adhesion molecules and antioxidant mechanisms, which have been relative to cardiovascular protection.19
RLS has been widely used, and a shorter RLS was found to be associated with higher risks of CVDs. Our study showed only postmenopausal females with the highest quartile of RLS (≥36 years) were associated with a lower risk of IHD.20 However, such association was not statistically significant and almost diminished after stratification by age. This weaker association of RLS Q4 with IHD could be explained by a relatively shorter RLS in Chinese postmenopausal females compared with those in high-income countries.10 20–22 Another possible reason for the shorter RLS and weaker association of RLS on IHD yielded in our study might be due to the inconsistent measurements of age at menopause adopted. Although mostly self-report has been adopted for females to recall the time of menopause, some studies accepted the date of menses having stopped after 12 months, while other studies used the age at the final menstrual period to define age at menopause.23 24 In the CKB questionnaire, participants were asked about age of completion of menopause without further explanation, potentially resulting in a relatively shorter RLS in our study. Therefore, future studies are encouraged to make menopause well-defined and adopt a consistent measurement technique to ensure that the outcomes are comparable.
A recent meta-analysis reported that whether longer EEE is cardioprotective has been inconclusive, given only four studies on the association of EEE with CVD, IHD and stroke mortality were conducted.9 However, in our study, a longer EEE showed a protective role on incident IHD compared with EEE Q1. Furthermore, bringing the number of stillbirths into our EEE equation has taken a step further to construct a comprehensive indicator, which was rarely seen in previous research and this innovation might also lead to different findings.
Our study also observed that a longer TEE (≥35.3 years) was associated with a lower risk of incident IHD. This result seemed aligned with the Women’s Ischemia Syndrome Evaluation study, which showed longer TEE might decrease the risk of CAD.16 However, they stated that such association might be driven by menopausal hormone therapy (MHT) use because it disappeared when removing MHT from TEE calculation.16 Given the lack of information on MHT in CKB, TEE calculation in our study only considered cumulative oestrogen before menopause. But its association with incident IHD was maintained, indicating that whether MHT was the main driver for IHD protection is unknown. Although this difference remains unclear, compared with the study of Merz et al (n=646), a much larger sample size in our study (n=118 855) might be able to minimise the degree of variability and stabilise our results.16 Additionally, compared with postmenopausal females in some high-income countries, females in China are more likely to adopt traditional Chinese medicine instead of MHT to reduce menopause-related symptoms, resulting in nearly no contribution of MHT to our TEE calculation.25 It will be worth for future research to include the duration of culturally appropriate treatment or therapy for menopause in TEE algorithm and explore its association with incident IHD or other CVDs among Chinese postmenopausal females.
Our study also explored the association between each reproductive factor and incident IHD. Aligned with such biological plausibility and previous epidemiological studies, we found the risks of incident IHD were related to early age at menopause, and pregnancy-related events including number of live births, stillbirths and miscarriage or terminations, with the magnitude of risk increasing with the number of such reproductive events.26 27 Although optimal lactation duration with reduced risk of maternal CVDs was historically reported, longer lifetime lactation duration was slightly associated with increased risk of IHD in our study. It might be possible that continuous variable of lifetime lactation duration adopted in our study potentially concealed the optimal cut-off effect.28 Additionally, our study showed increasing CP use was associated with decreased risk of IHD. Compared with previous research in the Netherlands, our study showed a much shorter CP use time among Chinese postmenopausal females.29 This might be due to the fact that CP was not one of the popular contraceptives used in China, potentially resulting from global disparities in knowledge and access to contraceptives, and various family planning policies.9 11 30 Therefore, our results on the protection that EEE or TEE conferred against the incident IHD could be driven by any or combination of above reproductive factors. Further research should investigate how each reproductive factor interacts with IHD through lifetime oestrogen fluctuation, considering potential localised EEE or TEE components.
This study has some important strengths. First, this is the first study that explored the associations of EEE and TEE with incident IHD among Chinese postmenopausal females. These findings provide a significant contribution to the limited literature on lifetime cumulative oestrogen with IHD. Second, the prospective CKB study possesses a large sample size from 10 geographical areas across China, ensuring reliable and precise findings in our results. Third, relatively long-term follow-up data including 13 162 new-onset IHD cases validated based on medical records enable precious outcome measurement and the avoidance of recall bias.
Our findings also need to be interpreted in light of limitations. First, lacking information on certain reproductive events such as MHT suggests the future improvement of TEE construction. Second, reproductive factors were collected based on self-report, which might have introduced recall bias. Moreover, using oestrogen-related reproductive variables to ascertain EEE and TEE may not always be accurate due to the many assumptions involved in estimating the duration of lactation, pregnancy, stillbirths and miscarriages. Third, the study was not able to adjust for all life course impacts that might be related to IHD. Despite all limitations mentioned above, this study presents valuable opportunities to understand the lifetime cumulative effects of reproductive factors on IHD in postmenopausal females, and how these may contribute to sex-specific risk prediction and further prompt actionable recommendations for primary and secondary IHD prevention for females.
Conclusion
In conclusion, this study found that longer RLS, TEE and EEE were associated with lower risks of IHD. These findings suggested that the lifetime cumulative effect of reproductive factors might serve to be valuable sex-specific indicators for IHD prevention and intervention programmes for females.
Data availability statement
Data are available upon reasonable request. This research has been conducted using the China Kadoorie Biobank (CKB) resource (www.ckbiobank.org). Publication of results does not require or imply approval by the membership of the CKB Collaborative Group. The raw CKB data underlying this article can be accessed via https://www.ckbiobank.org/CKBDataAccess, following the institution’s data-access policies. Preliminary event adjudication data are not publicly available.
Ethics statements
Patient consent for publication
Ethics approval
Ethical approval was obtained from the Oxford University Tropical Research Ethics Committee (approval number: 025-04, 3 February 2005), the Chinese Academy of Medical Sciences Ethical Review Committee, the Ethical Review Committee of the Chinese Centre for Disease Control and Prevention (Beijing, China, approval number: 005/2004, 9 July 2004) and the institutional research boards at the local Centre for Disease Control and Prevention in each of the 10 survey sites. All participants provided written informed consent.
Acknowledgments
We thank all China Kadoorie Biobank (CKB) participants, project staff, the China National Centre for Disease Control and Prevention and its regional offices (for access to death and disease registries) and the Chinese National Health Insurance scheme (for linkage to hospitalisation data). We appreciate CKB collaborative group providing data for this study (application number: DAR-2020-00212).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
LH and WL are joint first authors.
LH and WL contributed equally.
Contributors PS, as the guarantor, designed the study and was responsible for the overall content. LH managed and analysed the data. WL prepared the first draft. LH, WL, WS and JC reviewed and edited the manuscript, with comments from PS, SS, YF, YZ, CY and XL. All authors were involved in revising the paper, had full access to the data and gave final approval of the submitted versions.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the 'Methods' section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.