The risk of long-term cardiometabolic disease in women with premature or early menopause: A systematic review and meta-analysis

Background Transition into menopause is associated with an increased risk of cardiovascular disease (CVD). However, it is unclear whether the association exists between premature menopause (defined as age at menopause 40 years) or early menopause (defined as age at menopause 40–45 years) and CVD or cardiovascular risk factors. The aim of this review was to comprehensively evaluate and meta-analyze the most reliable evidence about the relationship between menopausal age and the risk of long-term cardiometabolic disease. Methods A comprehensive literature search of the PubMed, Web of Science, and Embase databases from inception to October 1, 2022, for titles and abstracts with a restriction to English language papers led to the discovery of the studies. Data are expressed as the Hazard Ratio (HR) with 95% confidence intervals (CI). The degree of heterogeneity was measured using the I-square (I2) index. Results 921,517 participants from 20 cohort studies published between 1998 and 2022 were considered. Compared to women with menopause at age >45 years, women with premature menopause (PM) or early menopause (EM) had a higher risks of type 2 diabetes (RR: 1.32, 95% CI: 1.08–1.62; RR: 1.11, 95% CI: 0.91–1.36, respectively), hyperlipidemia (RR: 1.21, 95% CI: 1.05–1.39; RR: 1.17, 95% CI: 1.02–1.33, respectively), coronary heart disease (RR: 1.52, 95% CI: 1.22–1.91; RR: 1.19, 95% CI: 1.07–1.32, respectively), stroke (RR: 1.27, 95% CI: 1.02–1.58; RR: 1.13, 95% CI: 0.97–1.32, respectively) and total cardiovascular event (RR: 1.36, 95% CI: 1.16–1.60; RR: 1.14, 95% CI: 0.97–1.35, respectively). No difference was found for hypertension in PM or EM women (RR: 0.98, 95% CI: 0.89–1.07; RR: 0.97, 95% CI: 0.91–1.04, respectively). Additionally, we also found that PM women, but not EM women, were linked with an increased risk of ischemic and hemorrhagic stroke. However, this is not in line with the conclusion that both PM and EM had a higher risk of total stroke. Conclusion Women with PM or EM have a higher risk of developing long-term CVD, compared to women with menopause at age >45 years. Therefore, we recommend early lifestyle interventions (e.g., maintaining a healthy lifestyle) and medical treatments (e.g., timely initiation of menopausal hormone therapy) to decrease the risk of cardiometabolic disease in early or premature menopausal women. Systematic Review Registration PROSPERO, identifier CRD42022378750


Introduction
Menopause is chronically defined as the cessation of menstruation for 12 months as a result of follicular function loss (1). The average age of menopause has mostly been estimated to range from 49 to 52 years (2). However, about 5% of the female population enter menopause between the ages of 40 and 45, a condition termed "early" menopause. Approximately 1% of women enter menopause before the age of 40, a condition termed "premature" menopause (also named primary ovarian insufficiency if the amenorrhoea is spontaneous) (3,4). According to WHO statistics, by 2030, 1.2 billion women globally will be in perimenopause or postmenopause, with 4.7 million women experiencing menopause each year (5).
Except for the impact of vasomotor symptoms on women's quality of life, transition to menopause is also linked with a higher risk of osteoporosis, dementia and CVD (6). Although considerable evidence suggests that early menopause is linked with the incidence of cardiovascular morbidity and mortality, the relationship between age at menopause and cardiovascular risk factors or CVD remains relatively poorly established (7). Some meta-analyses regarding PM and cardiovascular risk factors were mainly extracted from cross-sectional studies, comparing endpoints, such as type 2 diabetes (T2D), hyperlipidemia (HL) and hypertension (HT), between women with menopause at age >45 years. For example, a recent systematic review including prospective and cross-sectional studies reported the association between reproductive lifespan characteristics and cardiovascular risk factors. However, half of the studies were cross-sectional and could not deduce etiological relations (cause-and-effect) (8). In addition, current studies have reported the association between age at menopause and CVD, but these studies had inconsistent exposure variables (mixing premature and early menopause for analysis) or only examined the association between PM and CVD but not EM (9,10). Women live roughly a third of their lives after menopause, and this time can extend further for earlier menopausal women. Under the circumstances, it is important to evaluate the relation between earlier menopause and the incidence of CVD (11).
The aim of this study, which examined all available evidence from cohort studies with long-term follow-up, was to assess whether PM or EM increases the risk of cardiovascular risk factors or CVD events compared to menopause at age >45 years.

Search strategy
This systematic review and meta-analysis was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines (12), and was prospectively registered on PROSPERO (registration number: CRD42022378750). The

Study selection
Cohort studies were included if they fulfilled the following criteria: (a) conducted in postmenopausal women (either naturally or surgically); (b) the primary outcome of interest was T2D, HT, HL, or CVD events. (c) reported the HR (RR or OR) and the corresponding 95% CI for outcomes associated with EM or PM vs. the reference category. Exclusion criteria were set as follows: (a) intervention studies, animal studies, or conference abstracts; (b) reviews or case reports; (c) follow-up for less than 3 years; (d) less than 10 subject cases. Moreover, if multiple similar articles were published in the same cohort, we selected only the one that contained the most complete information.

Data extraction
Two researchers (Liu J and Liu W) conducted data extraction and quality assessment respectively, and consulted the third researcher (Huang J) to resolve disagreements. The extracted data included the author's name, publication date, region of study, year baseline, population number, age at enrollment, age at menopause, follow-up duration, menopause category, primary outcomes, MHT/OC used, and multivariate-adjusted effect estimates (e.g., HR, RR, or OR) with corresponding 95%CI. The multivariate-adjusted effect estimate was used for further analysis in studies that reported both adjusted and multivariate-adjusted effect estimates.

Study quality assessment
The Newcastle-Ottawa Scale (NOS) for quality evaluation of systematic reviews was applied to long-term follow-up cohort studies (13). The scale consisted of eight items divided into three sections: participant selection, cohort comparability, and outcome assessment. The total score was 9. Summary scores of ≤4, 4-7, and ≥7 represent low, moderate, and high quality studies, respectively, and all included studies scored more than 7 points. Liu et al. 10.3389/fcvm.2023.1131251

Frontiers in Cardiovascular Medicine
Detailed scores for each study can be obtained in Supplementary  Table S2.

Statistical analysis
We identified reference category in each article, and the majority had menopause ages of 50 to 54 years. Multivariateadjusted HR or RR and 95% CI were extracted from the selected literature to measure the association between PM or EM and cardiovascular outcomes. The original HRs extracted from the articles were considered equivalent effects to RRs (14). I 2 statistic was used to assess the heterogeneity between studies, with ≤50% representing no heterogeneity, >50% representing moderate heterogeneity, and >70% representing high heterogeneity. When there was no heterogeneity between studies, we chose a fixed-effects model to conduct the analysis, if not, the random-effects model was selected. The randomeffects model was used to conduct meta-analysis of outcomes (T2D, HT, CHD, and stroke) since the included literatures might differ statistically and clinically to some extent. When there existed high heterogeneity (I 2 > 70%), sensitivity analysis were conducted excluding literatures with a large heterogeneity (outliers) (15). Furthermore, Egger's test was applied to assess potential publication bias. All analyses were performed with Stata software, version 15.1 (Stata Corp LP, Texas, United States).

Hypertension
Eight studies for PM and seven studies for EM reported HT. No significant effect was found for PM (RR: 0.98, 95% CI: 0.89-1.07, I 2 = 56%) or EM (RR: 0.97, 95% CI: 0.91-1.04, I 2 = 65%) in the meta-analysis of the estimations for risk of HT ( Figure 3). No significant publication bias was found in Egger's test (p = 0.395 for PM, p = 0.179 for EM).

Hyperlipidemia
Two studies for PM and EM reported HL, with no heterogeneity between the studies (I 2 = 0% for PM women, I 2 = 23% for EM).

Total cardiovascular events
Five studies for PM and four studies for EM reported total cardiovascular events, with no heterogeneity between the studies (I 2 = 0%). Meta-analysis showed that both PM (pooled RR: 1.36, 95% CI: 1.16-1.60) and EM (pooled RR: 1.14, 95% CI: 0.97-1.35) had an increased risk of total cardiovascular events ( Figure 9). No significant publication bias was found in Egger's test (p = 0.357 for PM, p = 0.333 for EM).

Discussion
In this meta-analysis, based on 20 cohort studies, we found that premature or early menopausal women had an increased risk of cardiovascular risk factors, such as T2D and HL, but not of HT in comparison to women in the reference category. Both PM and EM were linked to an increased risk of CHD, stroke, as well as total CVD events. The results for stroke were inconsistent In this study, we also found that PM women, but not EM women, had an elevated risk of hemorrhagic and ischemic stroke. This is not in line with the conclusion that both PM and EM women had a higher risk of total stroke. This may be caused by the following reasons: Four studies (20, 24, 31, 34) reported ischemic and hemorrhagic stroke separately and one study (17) reported ischemic stroke, most of them observed a limited number of events (17,24,31,34). Furthermore, due to the late application of modern medical instruments such as computed tomography (CT), only a small number (18.5%) of the previously published cohort were specifically recorded as ischemic or hemorrhagic stroke (24). Finally, four of the included studies (20, 24, 31, 34) did not distinguish the type of stroke in some patients, it may be that hospital medical staff failed to provide specific information on the type of stroke or that the code of the disease was not validated. Therefore, the rate of ischemic or hemorrhagic stroke in the included studies may not have been as high as it should have been, and our findings regarding the associations between ischemic or hemorrhagic stroke and PM or EM may not be as reliable. However, in current systematic reviews, the relationship between PM or EM and the increased risk of stroke was not consistent. A recent meta-analysis of 78 studies reported that an earlier age at menopause is significantly associated with an increased risk of hemorrhagic stroke but not significantly total and ischemic stroke (35). And a meta-analysis included 10 prospective cohorts found total stroke, ischemic stroke, and hemorrhagic stroke were not increased in POI women (9). These results suggest that the relationship between menopausal age and the risk of stroke remains controversial, and further well-designed prospective cohorts with adequate samples are required to clarify this association.

Mechanisms of CVD in early menopausal women
The higher cardiometabolic risk in earlier menopausal women may involve the following mechanisms: Estrogen exerts a positive protective effect on the cardiovascular system, and a longer reproductive lifespan (representing longer duration of estrogen exposure) is associated with a lower risk of CVD events (36). Early estrogen depletion may exacerbate the expression of inflammatory chemicals and decrease vascular function (37). In a cross-sectional study comparing early vs. late menopause, higher estrogen levels were linked to smaller carotid interadventitic diameter (reflecting less carotid remodeling) and increased flowmediated dilatation of the brachial artery (suggestive of better endothelial function) (38). Additionally, declining estrogen levels are linked to higher levels of a wide variety of cardiovascular risk factors (39, 40). PM or EM is accompanied by abnormal glucose FIGURE 2 Forest plot presents the association between PM or EM and the risk of T2D in cohort study.  Forest plot presents the association between PM or EM and the risk of hypertension in cohort study. FIGURE 4 Forest plot presents the association between PM or EM and the risk of hyperlipidaemia in cohort study. and lipid metabolism, which are independent and important cardiovascular risk factors. A meta-analysis included 21 studies reported that earlier menopause was linked to significantly higher lipid and glucose levels (41). Also, hyperandrogenism and low sex hormone-binding globulin are involved in the development of CVDs (42).

The influence of confounders
The cohort studies examined in this analysis adjusted for diverse confounding factors including MHT. There were not enough studies to demonstrate whether the results were independent of MHT, a major confounder in the association between earlier menopause and CVD (16). This issue may be better addressed by a meta-analysis included detailed information on each participant. According to a meta-analysis, the impact of estrogen on CVD was related to the age of menopause and the age at which MHT began. The risk of CHD in women who started MHT soon after menopause was significantly reduced, with a RR of 0.72 (95% CI: 0.56-0.92) when using estrogen plus progestin. The RR of estrogen-progestin administration more than 10 years after menopause was 0.90 (95% CI: 0.62-1.29). The RR of CHD was 1.07 (95% CI: 0.65-1.78) in women older than 60 years who started hormone therapy (43). Observational studies have shown that MHT can provide CVD benefits, and this benefit may be more evident in early menopausal women (26, 44). The International Menopause Society (IMS) published a recommendation on MHT in 2016: starting and long-term use of estrogen has cardiovascular protective effects in perimenopausal and postmenopausal women under the age of 60, and it is not recommended to use MHT to prevent CHD in women over the age of 60 (45). However, for women with established CVD, there is currently no strong evidence that hormones improve cardiac outcomes. The HERS study included 2,763 women with CHD at a mean age of 66.7 years. At a mean follow-up of 4.1 years, the treatment group's relative risk for myocardial infarction or death from CHD was 0.99 (95% CI: 0.80-1.22) compared to the control group, but the treatment group's risk of CHD was higher in the first year. While the risk decreased in 3-5 years, to determine whether the reduction in risk was sustained, a subsequent follow-up of 2.7 years was conducted, and the risk of CHD did not decrease (46). The results suggest that MHT should not be utilized to lower the risk of recurrent CHD events in women with CVD.
The form of menopause was also a confounder. During natural menopause, hormone levels decline gradually, and for many years after menopause, large amounts of testosterone are still produced Forest plot presents the association between PM or EM and the risk of coronary heart disease in cohort study. Liu et al. 10.3389/fcvm.2023.1131251 Frontiers in Cardiovascular Medicine and peripherally converted to estrogen. With bilateral oophorectomy, hormone levels are abruptly decreased, resulting in acute hypoestrogenic and hypoandrogenic conditions (47). Different types of menopause may lead to inconsistent postmenopausal hormone levels, which are associated with different cardiovascular risks. However, three studies included in this analysis separately analyzed and reported the effects of natural and surgical menopause on cardiovascular outcomes, and no significant associations were observed (17,20,34). This is consistent with another cohort study that specifically examined the relationship between natural or surgical menopause and CVD (48). According to this study, the cardiovascular risk profile before surgery already determines the future cardiovascular risk of women undergoing oophorectomy and has little to do with the surgery itself.

Strengths and limitations
To our knowledge, this is the first systematic assessment that reports on all significant cardiometabolic endpoints based on longitudinal cohort studies. The strength of this study is that the literature included was of high quality, scoring over 7 points according to the NOS. Furthermore, we conducted subgroup analysis to examine the relationship of PM and EM with cardiovascular risk factors and CVD events separately, which may allow for risk stratification based on age at menopause and better patient management. Moreover, sufficient numbers of American, Asian, and European populations were included in our study, which allowed us to perform subgroup analysis according to ethnicity. We used multiple-adjusted effect sizes to avoid the effect of confounders on the final results, unfortunately, the confounders varied across studies. Only analyzing the data for each individual would possibly address this question.
Our study has certain limitations, mainly in terms of clinical heterogeneity. The reference age for normal menopause was not identical between studies, and the outcome measures were affected by differences in the reporting methods, some from selfreports and some from device diagnoses. The clinical heterogeneity of the literature may influence how generalizable our findings are to specific clinical contexts using different reference categories. This is the primary limitation of our study and other studies have examined the association between earlier menopause and the increased risk of CVD. Moderate heterogeneity (I 2 > 50%) existed in the systematic review for T2D, HT, CHD (<40), stroke (40)(41)(42)(43)(44)(45), ischemic stroke (40)(41)(42)(43)(44)(45). Considerable heterogeneity (I 2 > 70%) existed for stroke (<40), FIGURE 6 Forest plot presents the association between PM or EM and the risk of stroke in cohort study.  Forest plot presents the association between PM or EM and the risk of ischemic stroke in cohort study. FIGURE 8 Forest plot presents the association between PM or EM and the risk of hemorrhagic stroke in cohort study. Unfortunately, most studies only included natural menopausal women or treated surgical menopause as a confounding factor without comparing the cardiovascular risks of natural and surgical menopause. Considering the few included studies separately analyzing natural and surgical menopause, we are not available to perform subgroup analysis based on the pattern of menopause. However, a very recent observational study had shown that, when compared to natural menopause, surgical menopause did not significantly increase the risk of total stroke (34). Third, postmenopausal hormone use may have an impact on the relationship between earlier age at menopause and CVD (49). We could not assess the effect of MHT use, a potential major confounder, on CVD in early or premature menopausal women. However, studies have shown that about 80% of women use hormone treatment for at least 6 years after menopause (50). Therefore, we can assume that the bias due to hormone use is very small (16). Fourth, the self-reported age of menopause might lead to recall bias. However, some articles have reported that the majority of women accurately described the state of menopause, indicating that self-reported menopausal age is reasonably reliable (51). Besides, considering that each article included standard questionnaires and detailed questions, we assume that the effect of self-reported age at menopause on the final outcome is limited.

Public health value
Our results may have useful public health value. First, compared with smoking, hypertension, and diabetes (OR: 2.86, 95% CI: 2.36-3.48; OR: 2.95, 95% CI: 2.57-3.39; OR: 4.26, 95% CI: 3.51-5.18, respectively) (52), PM or EM is a moderate risk factor for CVD. Although modest, the increased CVD risk is significant. Therefore, maintaining a healthy lifestyle, such as quitting smoking and exercising on a regular basis (53), is the cornerstone of CVD prevention in early menopausal women. Second, the timing of MHT initiation affects its cardiovascular effects, namely the "timing hypothesis" (45). Timely initiating MHT in early menopausal women not only improves symptoms and quality of life but also has cardiovascular benefits that can be used as primary CVD prevention. Third, data for secondary prevention of CVD and improved survival for women are still lacking, with further research needed to develop evidence-based recommendations specifically for women (54). Forest plot presents the association between PM or EM and the risk of fatal cardiovascular events in cohort study.

Conclusion
In summary, this study included 20 cohorts showed that women with early menopausal age (< 40 years or 40-45 years) had an increased risk of developing CVD. This needs to be taken into consideration when developing prognostic models for the early diagnosis of CVD in women, particularly in those at higher risk, which requires timely lifestyle intervention and possibly hormonal treatment. Prospective cohort studies with larger sample sizes and well-designed intervention studies may help to address this issue.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.