Pregnancy and Reproductive Risk Factors for Cardiovascular Disease in Women

Menarche is the onset of the first menstrual period and is often considered the central event of female puberty. The average age of menarche is ≈12 years in the U.S. Early menarche is typically defined as occurring before age 12 years, but some researchers have defined it as occurring before 10 to 11 years of age. Although genetic factors play a role, potentially modifiable factors such as lower birthweight, increased weight fluctuation during childhood, and especially higher childhood body mass index (BMI) may increase the likelihood of early menarche.⁸ In parallel with rising rates of obesity in children, the mean age of menarche in the United States has decreased over the past 20 years from 12.1 to 11.9 years.⁹ Black and Hispanic women are disproportionately likely to undergo early menarche.¹⁰

Early menarche is associated with future obesity and metabolic syndrome in adulthood. Whether this is independent of premenarchal and adolescent BMI values is unclear. In the United Kingdom European Prospective Investigation into Cancer-Norfolk study of 15 807 women, history of early menarche was associated with greater odds of having hypertension, diabetes, obesity (BMI≥30 kg/m²), and central obesity (waist circumference ≥80 cm) later in life.¹¹ Other studies have linked early menarche with greater risk of future metabolic syndrome and its individual components, including dyslipidemia and hypertension.¹²

Early menarche has been associated with a 15% to 30% higher risk of future CVD^11,13,14 independent of sociodemographic factors.^11,14 Data further suggest the relationship between age of menarche and future CVD may be U-shaped, with increased risk also noted for late menarche (after age 17 years).^13–15 The association of early menarche with future CVD may be stronger in higher-risk women. Among women with suspected ischemia undergoing coronary angiography in the Women’s Ischemic Syndrome Evaluation study (mean baseline age 58 years), a history of menarche ≤10 years of age was associated with >4-fold risk of a subsequent major adverse cardiovascular event over 6-year follow-up (hazard ratio [HR], 4.53 [95% CI, 2.13–9.63]) in adjusted analysis; menarche ≥15 years of age was also associated with CVD risk (HR, 2.58 [95% CI, 1.28–5.21]).¹⁵

Mechanisms linking early menarche to increased CVD risk later in life are incompletely understood. Given the strong association between elevated childhood BMI and early menarche, early menarche may reflect genetic or lifestyle factors, including excess calorie consumption and reduced physical activity. Higher leptin levels, which are associated with increased adiposity, are associated with earlier onset of menarche.¹⁶ This suggests dysregulation of adipokines during puberty may reflect negative metabolic influence that increases a woman’s CVD risk later in life.¹⁶ Early versus normal or late menarche is associated with unfavorable levels of insulin, glucose, blood pressure (BP), and body fat among adolescent girls.¹⁷

Polycystic ovary syndrome (PCOS), which affects 5% to 13% of women, is characterized by amenorrhea or oligomenorrhea, hyperandrogenism and its associated clinical features (eg, hirsutism, acne), and polycystic ovaries.¹⁸ PCOS is associated with an adverse cardiometabolic profile. Women with PCOS have a higher prevalence of elevated BMI, dyslipidemia, hypertension, and insulin resistance and deficiencies in insulin secretion¹⁹ compared with similarly-aged women without PCOS.^18,20 The most common lipid abnormalities noted in PCOS are elevated LDL-C (low-density lipoprotein cholesterol), elevated triglycerides, and decreased HDL-C (high-density lipoprotein cholesterol) levels.²¹ Among women with PCOS, insulin resistance is seen in approximately two-thirds,²² and 7.5% to 10% have established type 2 diabetes (T2D).²³ Women with PCOS have a higher risk of developing T2D independent of BMI.²⁴ In addition, a recent meta-analysis confirmed an association between PCOS and incident hypertension (pooled risk ratio, 1.70 [95% CI, 1.43–2.07]), although risk attributable to PCOS was greater among women of reproductive versus postmenopausal age.²⁵ PCOS has also been associated with higher levels of inflammatory markers such as CRP (C-reactive protein) as well as thrombotic markers such as homocysteine.^26,27

Women with PCOS have greater carotid intima-media thickness (CIMT) and coronary artery calcium (CAC), even after adjusting or matching for BMI versus women without PCOS.^28–30 The relationship between PCOS and clinical CVD is well established, but data are conflicting regarding whether excess cardiovascular risk conferred by PCOS is independent of other conventional risk factors. Some studies,^31,32 but not all,^33,34 have reported increased CVD risk among women with PCOS after accounting for adiposity. A recent meta-analysis found that PCOS was associated with a 30% higher risk of overall CVD (odds ratio [OR], 1.30 [95% CI, 1.09–1.56]), including both coronary heart disease (CHD; OR, 1.44 [95% CI, 1.13–1.84]) and stroke (OR, 1.36 [95% CI, 1.09–1.70]).³⁵ A recent Danish registry found 19% increased risk of incident CVD (adjusted HR, 1.19 [95% CI, 1.07–1.33]) for women with PCOS, but this association was not seen in women after age 50 years.³⁶ Women with PCOS are also ≈3-fold more likely to develop APOs, including gestational diabetes mellitus (GDM) and preeclampsia, which in turn predict increased CVD risk.³⁷

Multiple effective contraceptive options are available, including long-acting reversible contraception (levonorgestrel-releasing and copper intrauterine devices [IUDs] and subdermal implants), injectable depot medroxyprogesterone acetate, combined oral contraceptive pills (OCPs) that contain a combination of estrogen and progestins, progestin-only pills, the transdermal contraceptive patch and the vaginal ring (NuvaRing; Figure 2). Of these, long-acting reversible contraceptions are preferred given high efficacy and safety, especially in women with established CVD or high CVD risk, where avoidance of unplanned pregnancy is key.^38,39 As copper IUDs are associated with increased menstrual bleeding, hormonal intrauterine devices may be preferred in women taking antiplatelet therapy or anticoagulation.³⁹

OCPs are the most prescribed contraceptive method in the United States and are used by ≈25% of women of reproductive age. Safety concerns that arise from their use are largely related to estrogen content.^38,39 The metabolism of the synthetic estrogens can cause BP to rise, and normotensive women taking OCPs can have up to a 7 to 8 mm Hg increase in BP.⁴⁰ Within the first year of use, risk of venous thromboembolism (VTE) is increased among OCP users (3–9/10 000 woman-years) compared with nonusers (1–5/10 000 woman-years).⁴¹ Similar to elevations in BP, this increased VTE risk is attributed to the synthetic estrogens within the pills. Progestins in OCPs have evolved over time to reduce, but not erase, this VTE risk. The risk of myocardial infarction (MI) or ischemic stroke also appears to be higher in women using OCPs, especially those formulations with higher levels (≥50 µg) of estrogen.⁴² The hormonal makeup of these pills has evolved, with newer generations of OCPs containing less synthetic ethinyl estradiol (≤30 µg) than older generation pills, reducing the CVD risk associated with newer pills; older data may, therefore, overestimate contemporary OCP-associated risks.

Cardiovascular risk factors should be considered when prescribing OCPs. Women with uncontrolled hypertension (BP ≥160/100 mm Hg) should not use OCPs, and they are also not recommended in women with systolic BP 140 to 159 mm Hg and diastolic BP 90 to 99 mm Hg or in women over the age of 35 years with controlled hypertension. Smoking is a relative contraindication with OCPs given that concurrent use is associated with a 10-fold increased risk of MI and 3-fold increased risk of stroke.⁴³ OCPs are not recommended in women at increased risk of VTE, which includes but is not limited to women in the early postpartum period, ≥35 years old who use tobacco, and those with thrombophilia.³⁹ The US Medical Eligibility Criteria for Contraceptive Use, in addition, cautions against the use of OCPs in other groups of women, including women with a history of ischemic heart disease, migraine with aura, or complicated valvular heart disease, or certain congenital heart disease conditions.³⁸ Long-acting reversible contraceptions and progestin-only options are recommended for these patients.^38,39 Although barrier methods are safe from a cardiovascular perspective, use of barrier methods in isolation is discouraged given their poor efficacy. An update to the US Medical Eligibility Criteria is expected in 2022.

High-quality data on the relationship between infertility (defined by the World Health Organization as inability to achieve pregnancy after ≥12 months of unprotected intercourse), fertility treatment, and cardiovascular risk are limited. Women with infertility, even before conception, have a higher prevalence of cardiovascular risk factors,^44–46 suggesting that shared risk factors may underlie both infertility and CVD. Higher rates of diabetes and chronic hypertension⁴⁵ and obesity⁴⁶ are seen in both women with infertility and CVD as well as shared comorbidities, for example, PCOS as discussed above.

Use of assisted reproductive technology (ART) itself is also associated with increased CVD risk, although causality is unclear. High-quality studies to clarify these relationships are challenging due partly to marked disparities in access to ART and variability in ART practice patterns. ART use is associated with significantly higher rates of APOs such as hypertensive disorders of pregnancy (HDPs) and GDM,⁴⁴ conditions which in turn are associated with increased future CVD. A systematic review of 47 studies showed that in vitro fertilization in particular is associated with HDPs.⁴⁷ Data regarding the longer-term associations of ART use with long-term cardiovascular health, however, are both limited and mixed.^44,48 One cohort study of >6000 women found no increased risk after successful ART, suggesting ART itself may not be a causal risk factor.⁴⁴ Further research is needed to clarify whether ART is a marker for subclinical risk at the time of fertility treatment or whether it contributes causally to CVD risk.

GDM results when pancreatic beta-cells incompletely respond to the physiological and placental-mediated insulin resistance that characterizes pregnancy.⁸⁷ Approximately one-third of women who develop GDM will develop T2D in the future, corresponding to a 7-fold greater risk compared to women without GDM.⁸⁸ International diabetes associations recommend glucose screening starting within the first 1 to 6 months postpartum and lifelong screening every one to three years in women who develop GDM.^89–91 In women who develop GDM, subclinical risk factors (eg, suboptimal glycemic control, dyslipidemia, CRP, adiponectin) may be evident before pregnancy and early in the first trimester.⁹²

Women who develop GDM have a 2-fold risk of CAC at midlife independent of progression to prediabetes or T2D⁹³ and 1.5- to 2-fold risk of cardiovascular events⁹⁴; CVD event risk appears greater if there is progression to T2D.^82,94 Similar to the relationship between impaired gestational glucose tolerance and risk of subsequent T2D, there appears to be a dose-dependent relationship between degree of glucose impairment during pregnancy and risk of subsequent CVD.⁹⁵ These relationships have been consistently demonstrated across databases, countries, and ethnic groups.⁹⁵

The association between preterm delivery (PTD) and future maternal CVD risk is well established.^96,97 The underlying mechanisms, however, remain unclear. PTD is broadly defined as a live birth before 37 weeks gestation and further delineated based on gestational age (late preterm [34–36 weeks], moderately preterm [32–36 weeks], and very preterm [<32 weeks]). In the United States, Black women experience the highest rate of PTD at 29% higher than the national average (Figure 4) and have nearly 3-fold risk of delivering a very preterm baby (3.1% versus 1.2% in White women).⁹⁸

Whether PTD occurs spontaneously or due to medical necessity influences its relationship with future CVD risk.⁹⁹ Compared with spontaneous term delivery, medically indicated PTD is associated with a higher risk of cardiovascular mortality (HR, 3.7) than spontaneous PTD (HR, 1.7).¹⁰⁰ Women who experience medically indicated moderate to very preterm delivery due to HDPs appear to be the highest risk group. Preeclampsia, however, only explains ≈25% of the association between PTD and CVD.¹⁰¹ Although spontaneous and medically indicated PTD share many ASCVD risk factors, the most notable difference is in vascular function. The augmentation index, which indirectly measures smooth muscle vascular function, appears to be reduced in women with spontaneous PTD versus medically indicated PTD. These findings have been reproduced in both the early antepartum as well as the early postpartum phases.^102,103

Infant birth weight has been independently associated with maternal CVD risk. In the Women’s Health Initiative (WHI), delivery of a low-birth-weight infant was independently associated with increased maternal ASCVD risk after adjustment for conventional risk factors and other APOs (adjusted OR, 1.25).¹⁰⁴ Infant birth weight has also been shown to predict maternal life span.¹⁰⁵

Many studies have examined the association between small-for-gestational-age and intrauterine growth restriction, that is, small-for-gestational-age associated with pathological growth restriction, with CVD.¹⁰⁶ Fetal growth has been correlated with subclinical markers of CVD including arterial stiffness and endothelial dysfunction.^107,108 Bonamy et al¹⁰⁹ observed a multiplicative interaction between gestational age and birth weight, with a 3-fold maternal CVD risk in women with preterm and small-for-gestational-age infants even after accounting for socioeconomic factors, smoking, and pregnancy-related complications. Growth restriction at term is also associated with increased risk of maternal CVD,¹¹⁰ and, to a lesser extent, with paternal CVD.¹⁰⁵ This suggests that environmental and behavioral risk factors may influence both parents and infants.

Most cases of intrauterine growth restriction are thought to result from uteroplacental insufficiency due to poor implantation of the spiral arteries and subsequent increased placental vascular resistance, which has similarities to the early pathophysiology of preeclampsia. Preexisting hypertension and increasing maternal age have also been associated with growth restriction and may result from endothelial dysfunction that affects placental implantation. Low levels of insulin-like growth factor 1 are found in pregnancies complicated by intrauterine growth restriction and are also associated with increased risk of CVD suggesting a potential shared mechanism.¹¹¹ Further research is needed to understand the complex interplay between environmental, behavioral and maternal vascular health and the association between birth weight, fetal growth, and CVD risk.

Placental abruption has been associated with cardiovascular risk factors and CVD.^112,113 In a retrospective Dutch study, abruption was strongly associated with other concomitant APOs as well as with higher BMI, total cholesterol, and fasting blood glucose in women with isolated abruption without other APOs.¹¹² A large retrospective cohort study similarly found a 1.7-fold risk of CVD (adjusted HR, 1.7 [95% CI, 1.3–2.2]) in women with placental abruption or infarction.¹¹³ Pregnancies with large-for-gestational-age fetuses are also associated with increased CVD risk, likely mediated by associations with elevated BMI and diabetes.^114,115

Although there are numerous shared risk factors between miscarriage and CVD, including obesity and cigarette smoking, data show that miscarriage is independently associated with future CHD and MI after adjustment for conventional risk factors.^116,117 Greater number of miscarriages appears to associate with progressively higher CVD risk,^117,118 with one cohort study observing a >3-fold CVD risk associated with ≥3 miscarriages (HR, 3.18 [95% CI, 1.76–5.78]).¹¹⁸ Mechanisms linking miscarriage to CVD remain unclear, although shared vascular, genetic, immunologic factors have been proposed.¹¹⁹

Maternal breastfeeding is associated with a lower risk of later-life cardiometabolic and CVD independent of socioeconomic and lifestyle factors. A recent meta-analysis indicated that achieving ≥1 year of cumulative lactation was associated with 13% lower risk of chronic hypertension,¹²⁰ with a dose-responsive relationship between duration of lactation and degree of risk reduction.¹²¹ An even stronger dose-dependent protective association has been observed between lactation and incident T2D.^120,122 Notably, associations were not explained by postpartum weight differences by breastfeeding status.^121,122

Similar associations have been observed for subclinical atherosclerosis and clinical CVD events.³⁵ In the SWAN (Study of Women’s Health Across the Nation) Heart Study, women who did not breastfeed had elevated risks of aortic calcification (adjusted OR, 3.85 [95% CI, 1.47–10.00]) and CAC (adjusted OR, 2.78 [95% CI, 1.05–7.14]) compared to women who breastfed each child for ≥3 months, even after adjustment for demographic and lifestyle factors.¹²³ In the Nurses’ Health Study, women with ≥2 years of cumulative breastfeeding versus no breastfeeding had 23% (95% CI, 6%–38%) lower adjusted risk of CHD events.¹²⁴ A recent analysis observed similar findings for cardiovascular hospitalization and mortality.¹²⁵

A smaller number of studies have examined lactation and cardiometabolic risk factors specifically in women with APOs with mixed findings. In a study of 622 women with various APOs seen 6 months postpartum at a dedicated postpartum transitional clinic in Ontario, Canada, increased lactation duration was associated with higher HDL-C and lower fasting glucose, triglycerides, BMI, and risk of metabolic syndrome after adjustment for demographic factors and prepregnancy confounders.¹²⁶ Finally, in the Coronary Artery Risk Development in Young Adults study (CARDIA), longer duration of lactation was associated with lower risk of incident metabolic syndrome, with greater protection observed in women with versus without GDM.¹²⁷

Gestation is associated with accumulation of maternal visceral fat and insulin resistance.¹²⁸ Given the observed associations of lactation with cardiometabolic traits in human and animal studies, lactation has been hypothesized to reset maternal metabolism postpartum. Specific protective mechanisms, however, are incompletely understood. A recent analysis found that lactation duration was inversely associated with computed tomography imaging-derived measurements of pericardial adipose tissue and visceral adipose tissue volumes assessed at age 50 years and that longitudinal weight changes mediated only ≈20% of this relationship.¹²⁹ Mouse studies suggest lactation-related hormones may preserve pancreatic beta cell function.¹³⁰

Most large cohort studies suggest a J-shaped relationship between parity and CVD, in which the highest risk is among nulliparous and grand multiparous (≥5 births) women.¹³¹ The Atherosclerosis Risk in Communities Study found that a history of ≥5 live births was associated with CHD risk and hospitalization for MI, independent of breastfeeding history.¹³² These studies also found that having a prior pregnancy and no live birth was associated with greater CHD and heart failure risk. Parikh et al,¹³³ in the largest study to date of 1.3 million Swedish women, observed a J-shaped association between parity and CVD with a nadir of CVD risk in women with 2 births, HR 1.09 (95% CI, 1.03–1.15) for women with 1 birth, and HR 1.47 (95% CI, 1.37–1.57) in women with ≥5 births.

Parity may be a proxy for socioeconomic and behavioral factors.¹³⁴ In the British Women’s Heart and Health Study and British Regional Heart Study, a J-shaped association has been observed between number of children and CHD in both sexes, with the lowest prevalence among those with 2 children, suggesting a role of social and lifestyle factors. For those with ≥2 children, each additional child increased the age-adjusted odds of CHD by 30% in women versus 12% for men, however, suggesting the possibility of additional biological effects in women.

Several putative biological mechanisms have also been proposed to explain this relationship. Repeated pregnancies may result in multiple exposures to the metabolic changes of pregnancy, including insulin resistance, impaired lipid metabolism, weight gain, inflammation, and oxidative stress that result in enduring vascular changes.¹³⁵ Further research is needed to fully understand the biological and social mechanisms responsible for these associations.

Sex-specific risk factors represent an important component of comprehensive CVD risk assessment in women. Women with PCOS, for instance, should be screened for glucose intolerance, dyslipidemia, and hypertension.^195,196 As noted above, a history of premature menopause and APOs are now considered risk-enhancing factors to refine ASCVD risk assessment and guide statin prescription among women aged 40 to 75 years with borderline (5%–<7.5%) or intermediate (7.5%–<20%) calculated 10-year ASCVD risk using conventional risk calculators, for example, the Pooled Cohort Equations.^4,5 Consensus statements recommend close monitoring for CVD risk factors in women during the first year postpartum following APOs.¹⁹⁷ Although guidelines recommend aspirin use for the prevention of preeclampsia in persons at risk,¹⁹⁸ data are unclear regarding the long-term benefit of aspirin use for cardiovascular prevention in women with a history of HDP.¹⁹⁸ A recent American Heart Association statement on menopause and CVD recommends an aggressive prevention-based approach for women during the MT to decrease the probability of a future CVD event.¹³⁷ For women whose CVD risk may be uncertain (eg, borderline/intermediate calculated 10-year CVD risk), monitoring for subclinical development of CVD including CAC measurement can be considered to identify those who might benefit from preventive pharmacotherapies.^199,200 Overall, a history of pregnancy-associated and other reproductive risk factors for CVD represent an opportunity for primordial and primary prevention.

Appropriate risk assessment and prevention of CVD in women remains a challenge.^201,202 One study found that only 42% of cardiologists felt extremely well prepared to assess cardiovascular risk in their female patients, and only 22% reported using guideline-directed prevention guidelines in treating female patients.²⁰³ This is an especially important issue in the United States given the considerable burden of CVD risk factors among women, with pronounced racial and ethnic disparities.^1,204,205

Additional work is needed to better educate women about their risk of CVD such that they can become more active in their own care. Recent data show that only 45% of United States women recognize that CVD is the leading cause of death among women,²⁰³ and fewer than half of Canadian women recognized common symptoms of and risk factors for CVD.²⁰⁶ Minority women in particular continue to significantly underrecognize their risk of CVD.²⁰³ Awareness of CVD risk factors and symptoms has been shown to improve patient engagement with healthful behaviors and medication adherence.^207,208

With many elements of cardiovascular health to be discussed in a time-limited clinical encounter, efficiently incorporating reproductive history in clinical practice can be challenging. Formal clinical tools have been developed to help with this integration. A toolkit recommended by the American College of Obstetrics and Gynecology provides an algorithm to help in screening and identifying women who may be at increased CVD risk based on prepregnancy or pregnancy history.^7,209 Enhanced CVD risk factor screening can also be facilitated by standardized previsit patient questionnaires and patient note templates to prompt reflection on and discussion of risk factors.²⁰² Written educational materials about cardiovascular health may efficiently supplement clinical encounters.

Postpartum transitional clinics for women with APOs have emerged as a promising strategy for identifying women at elevated risk for CVD and facilitating the transition to longitudinal primary care.^210,211 Many structural barriers exist to universal access to such clinics, however, including inconsistent access to healthcare insurance particularly among high-risk women, lack of social support for maternity leave, and logistical challenges to accessing medical appointments with a young infant. In addition, bundled payments for obstetric care represent a potential financial disincentive for systems to introduce transitional clinics. Many patients are now turning to wearable consumer medical technologies, such as mobile phone applications (apps) and remote BP monitoring, to become more actively engaged in their own care; these results can be integrated into clinic visits to help facilitate patient-centered lifestyle recommendations.^202,212,213 Reliance on wearables, however, risks exacerbating existing racial and socioeconomic disparities among those without access to these new technologies. Finally, all fellows-in-training should receive education regarding the importance of reproductive history-taking as part of their core educational curriculum.^214,215 Women’s cardiovascular health topics should be featured more prominently at professional society conferences such that all providers may feel better informed about the unique aspects of caring for women with CVD.

Despite considerable advances in our understanding of sex-specific risk factors for CVD, many unanswered questions remain (Table 3). Ongoing research is needed to elucidate the pathophysiology of sex-specific CVD risk factors and their mechanistic links to later-life CVD to inform precision prevention strategies. Whether prevention strategies tailored to specific subsets of women can alter long-term cardiometabolic risk trajectories will be a critically important area of future clinical investigation. Including more women, and in particular women of child-bearing age who may be either pregnant or breastfeeding, as well as women transitioning through menopause,¹³⁷ in clinical trials remains an important component of efforts to address these knowledge gaps, as does shifting the priorities of trial leaders, pharmaceutical companies, and regulatory bodies to recognize the value of this inclusion. Capturing reproductive history in clinical trials is critical to generate evidence about the effects of therapies on specific subpopulations of women. To date, incorporation of sex-specific risk factors to refine CVD risk assessment has not demonstrated clinically meaningful risk reclassification over the intermediate-term in relatively low-risk populations.^216,217 However, ongoing efforts to develop sex-specific risk calculators over longer time horizons (eg, 30-year or lifetime risk) and in diverse populations remain important, as traditional risk models underestimate CVD risk in women with reproductive risk factors.²¹⁶ Increased clinician and patient awareness of sex-specific CVD risk factors and implementation of evidence-based prevention strategies are critical to improving the cardiovascular health of women.