Cardiovascular disease (CVD) is the leading cause of mortality among women worldwide (1). The incidence of CVD increases significantly after menopause but is rarely seen in premenopausal women (2). Postmenopausal women are at increased risk of cardiovascular and metabolic diseases because of the deficiency of estradiol and the central distribution of body fat (3).

The mechanism of postmenopausal dyslipidemia leading to CVD has been clarified (1). Studies have confirmed that the levels of serum total cholesterol (TC), triglyceride (TG) and low density lipoprotein (LDL) increased in postmenopausal women, while the level of serum high density lipoprotein (HDL) decreased (4). Elevated level of serum LDL is often regarded as a sign of hyperlipidemia, which increases the risk of coronary heart disease. Moreover, the incidence of atherosclerosis increases with the level of LDL as a result of oxidative modification of LDL. Oxidized LDL deposits on the arterial wall of cardio-cerebral vascular areas and gradually forms atherosclerotic plaques. The atherosclerotic plaques block the corresponding blood vessels, lead to atherosclerosis and increase the risk of CVD (5). Studies have shown that even if the level of serum LDL is normal, partial oxidation which leads to atherosclerosis cannot be ruled out (6). In addition, elevated levels of serum TC and TG are also the factors of the deficiency in antioxidant mechanism, which contributes to subclinical atherosclerosis (7). Therefore, these changes are partly responsible for the increased incidence of CVD in postmenopausal women.

A number of studies have investigated the effect of earlier age at menopause on CVD outcomes (8–10). Compared to women with menopause at the age of 49~55 years, women with early menopause have an approximately 80% increased risk of mortality from CVD (11). Primary ovarian insufficiency (POI), which is a typical disease characterized by early menopause (before age 40years) in women, is defined as amenorrhea for four months and an elevated follicle stimulating hormone (FSH) level>25IU/L (measured twice at least four weeks apart) in women under 40 years old (12). Hence, POI, by virtue of early menopause and chronic estradiol deficiency, is a risk factor for CVD.

Knowledge of the potential associations of lipid profile with diminished ovarian response to FSH has public health significance for reducing the risk of cardiovascular morbidity and mortality in women. POI may provide important information to disentangle the effects of menopause and age on lipid profile. Thus, this systematic review and meta-analysis aimed to evaluate the levels of serum lipid profile in women with POI and healthy controls.

The funnel plot method was used to detect publication bias, regarding TC, HDL and LDL, the shape of the funnel plots was not obviously asymmetric ( Figure 3 ), indicated that there was no significant evidence of publication bias in this meta-analysis. The publication bias for TG was not assessed, as the Cochrane Handbook for Systematic Reviews of Interventions (www.cochranehandbook.org) stated that the test for publication bias yields unreliable results when <10 studies were included.

To our knowledge, this is the first meta-analysis to explore the relationship between lipid profile and POI. A total of 12 studies were selected in this meta-analysis, and 1805 women were represented. The subgroup analyses based on region, sample size, number of cases, mean BMI value of cases and mean age of cases were performed. This meta-analysis indicated that levels of serum lipid profile were remarkably altered in patients with primary ovarian insufficiency: levels of serum TC, TG and LDL in women with POI were significantly higher than that in healthy controls, however, no significant difference was observed regarding the level of serum HDL between women with POI and healthy controls.

Women with POI present several risk factors for the occurrence and development of CVD: autonomic and endothelial dysfunction, insulin dysfunction, abnormal lipid profile (33), genetic variants associated with early menopause (34, 35). Diminished ovarian response to FSH is thought to cause the occurrence and development of atherosclerosis based on non-human primate models (36) and studies of hypothalamic angiography (37). Moreover, women with POI have been demonstrated to have impaired endothelial function regarded as a precursor to atherosclerosis (31, 38). It has been reported that women during the menopause transition exhibit abnormal lipid profile, including increased serum levels of TC, LDL, TG and decreased level of HDL (4, 39). In the same way, early impairment of ovarian function result in earlier or worse lipid profile alterations which lead to CVD in women with POI. Futhermore, genome-wide association studies have identified genetic overlap between the age associated decline in reproduction, lipid regulation and glucose regulation (40, 41). Hence, early menopause is associated with the subsequent CVD, conversely, women who experience CVD events prematurely (before age 35 years) have POI due to the genetic factors and the compromised vasculature following CVD events (42).

The beneficial effects of intact ovarian function on lipid profile are generally ascribed to estradiol. Endogenous estradiol acts like an antioxidant, which has vasodilating and protective effects on cardiovascular system. It has also proven that estradiol promotes the degradation and excretion of blood fat, decreases the levels of serum TC and LDL, alters the distribution of TC in the body and reduces the accumulation of TC in the artery walls (43). Meanwhile, estradiol inhibits the activity of liver lipase, decelerates the degradation rate of HDL, accelerates the removal of LDL particles by liver and reduces the levels of serum TC and LDL by strengthening the activity of LDL receptors, thus improving lipid metabolism (44). In addition, estradiol deficiency leads to the activation of the renin-angiotensin system, the upregulation of the vasoconstrictor endothelin and the impairment of nitricoxide-mediated vasodilation. Endothelin and angiotensin II augment the oxidative stress, which may further contribute to atherosclerosis (45). Meanwhile, decreased estradiol reduces the release of nitric oxide by vascular endothelial cells by reducing the sensitivity of estrogen receptors, thereby increases the cardiac load (46). Thus, women with POI have a higher rate of CVD morbidity and mortality later in life due to estradiol deficiency for longer than women with natural menopause (38, 47, 48).

Hormone replacement therapy (HRT) has been recommended by international guidelines for women with POI for many years (12). HRT is an exogenous supplement, which can relieve clinical symptoms, prevent the occurrence and development of diseases, and reduce the long-term complications with high disability rate and fatality rate (e.g., cardiovascular disease), and improve living standards (12) of women with POI. It has been reported that HRT can regulate lipid metabolism, strengthen LDL clearance, reduce the levels of serum TC, TG, LDL, improve lipoprotein composition ratio, and reduce the risk of cardiovascular events (49). HRT and simvastatin have similar lipid-regulating effects (50). In addition, the second report of the National Cholesterol Education Program recommends HRT as a treatment for dyslipidemia in postmenopausal women. Importantly, a 1-mg/dl increase in HDL level may be associated with a 3%~5% decrease in CVD, and again a 1% reduction in TC level equals a 2% reduction in CVD, however, this could only be achieved with HRT (51). Unfortunately, the acceptance rate of HRT in women with POI is extremely low (52).

Subgroup analysis was conducted to further explore the source of heterogeneity and more accurately evaluate the correlation between lipid profile and POI. Heterogeneity in this meta-analysis of the correlation between TC and POI might come from region, sample size, mean BMI value of cases and mean age of cases. The correlation between TG and POI might come from region, sample size, number of cases, mean BMI value of cases and mean age of cases. The correlation between HDL and POI might come from region, mean BMI value of cases and mean age of cases. And the correlation between LDL and POI might come from region,sample size, number of cases, mean BMI value of cases and mean age of cases. (1) Because there is considerable overlap in Turkey of people with European ethnicity as well as west Asian (53), the regions in this meta-analysis were divided into Europe, Africa, Asia and Central Asia. People in different regions have different dietary habits, lifestyles and economic levels, which have different effects on levels of serum lipid profile; (2) all references included in this meta-analysis were designed as case-control studies. In case-control study, if the sample size or the number of cases is enormous, the research results will be affected by more confounders; (3) BMI≥25 kg/m² is diagnosed as overweight (54), overweight has significant impact on serum lipid profile, therefore, BMI value may be one of the sources of heterogeneity; and (4) ovarian reserve function is related to physiological age, in the reproductive domain, age>35 is considered as advanced age, and ovarian function significantly declines (55), thus physiological age is regarded as one of the sources of heterogeneity.

Part of the sources of heterogeneity was found by subgroup analysis, however, the heterogeneity was still obvious, hence, sensitivity analysis was conducted. In this meta-analysis,the combined results of the correlation between TG level and POI were sensitive to AbdulAzeez’s study (20), and the P values of the combined results were all less than 0.05 before and after the exclusion of the references. However, after excluding the references, the heterogeneity decreased (I² decreased from 72% to 19%), therefore, AbdulAzeez’s study (20) had significant impact on the heterogeneity of the combined results. After careful reading of the references, we found TG level in the study was estimated using glycerol-3 phosphate oxidase method, distinguishing from other studies using enzymatic and colorimetric methods. In addition, BMI value of the participants included in the study was not mentioned, and these confounders might cause heterogeneity. Sensitivity analysis revealed that the combined results of the correlation between TC, HDL, LDL levels and POI were stable and reliable. Moreover, no significant evidence of publication bias was found in this study, which also increased the credibility of this meta-analysis.

Nevertheless, a number of limitations in this meta-analysis should be acknowledged. Firstly, obvious heterogeneity existed among the original studies due to differences in sample size, background of subjects, diagnostic criteria of POI, and methods used to detect the levels of serum lipid profile. Secondly, the subgroup analyses of HRT using and mean age at diagnosis of POI were unable to be conducted due to the incomplete data. While the heterogeneity may relate to duration of POI rather than age per se, and the HDL level may be similar between POI and healthy controls due to the effect of HRT. Thirdly, the main confounders (e.g., diet or physical activity) influencing the lipid profile of patients in original studies were not adjusted. Finally, the references included in this study were all case-control studies, which would limit causal inference. Therefore, more cohort studies are needed to predict how the lipid profile in women with POI will develop over time. For these reasons, we recommend that our conclusions should be viewed conservatively.

In summary, this systematic review and meta-analysis demonstrated that the levels of serum TC, TG and LDL were significantly higher in subjects with POI. These mean that women who loss of ovarian function at a very young age will lead to some degree of changes in the lipid profile, increasing the risk of CVD. Therefore, we recommend that lipid profile should be evaluated and comprehensively studied to ensure that early medical intervention (especially HRT) and minimize the risk of CVD morbidity and mortality associated with dyslipidemia in patients with POI. Moreover, further studies are required to verify the relationship between serum HDL level and POI.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

LH, HW, MS, WK and MJ: literature search, screening, and data extraction. LH, HW, MS, WK: data analysis and results visualization. LH and HW: manuscript draft. MJ: manuscript modification. All authors reviewed the final version of the manuscript and approve it for publication. All authors contributed to the article and approved the submitted version.

This work was supported by grant from the National Natural Science Foundation of China (No. 81904240) and New Teacher Foundation of Beijing University of Chinese Medicine (No. 2021-JYB-XJSJJ-011).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.