Infertility and Risk of Cardiovascular Disease: A Prospective Cohort Study

Enrollment for the NHSII began in 1989; 116 429 registered nurses, who were between the ages of 25 to 42 years, were enrolled if they returned a mailed questionnaire. Every 2 years, follow‐up questionnaires were sent to participants to collect detailed information on a variety of health conditions as well as risk factors. Questionnaire completion implied consent. The study was approved by the Institutional Review Board of Brigham and Women's Hospital and Michigan State University. Because of the sensitive nature of the data collected for this study, requests to access the data set from qualified researchers trained in human subject confidentiality protocols may be sent to the Channing Division of Network Medicine at [email protected] . Participants self‐reported whether they had “tried to become pregnant for >1 year without success” on every questionnaire between 1989 and 2001 and then every other questionnaire until 2009 (when women were between the ages of 45 and 62). Participants were defined as having a history of infertility if they indicated they had tried to conceive for ≥1 year without success at any point in time, including women who concurrently or subsequently reported becoming pregnant. Infertility diagnoses were updated until age 45. 29 Women were also asked about the cause of their infertility and could choose from the following: tubal blockage, ovulatory disorder, endometriosis, cervical mucus factors, male factor infertility, not investigated, not found, and other. Women could report multiple causes for infertility. In sensitivity analyses, ovulatory infertility was further classified by symptoms associated with PCOS, based on the self‐reported presentation of menstrual cycle irregularity and excess androgens 30 , 31 Specifically, participants who reported ovulatory infertility were further stratified by menstrual cycle length (<32, ≥32 days), regularity (very regular or regular versus irregular or no period), and hirsutism (yes/no). Self‐reported recall of ovulatory infertility has been previously validated using a supplemental questionnaire (93% concurrence) and medical records (95% concurrence) and was found to have high validity, 32 as was self‐reported endometriosis (97% concordance with medical records). 33 In each follow‐up questionnaire, participants were asked to report new diagnoses of myocardial infarction (MI), stroke (cerebrovascular accident or transient ischemic attack), and coronary artery bypass grafting/angioplasty/stent. We included cases from enrollment in 1989 until the 2017 questionnaire cycle (which ended in May 2017). Following a self‐reported MI or stroke, participants or next of kin were asked for permission to obtain and review medical records or death certificates, which were then reviewed by study clinicians blinded to the questionnaire information. Confirmed MIs met the criteria of the World Health Organization: symptoms and either diagnostic electrocardiographic changes or raised cardiac enzymes 34 ( International Classification of Diseases, Eighth Revision [ ICD‐8 ] codes 410 and 412). Stroke was classified as ischemic or hemorrhagic by the National Survey of Stroke criteria (atypical neurological deficit of rapid or sudden onset lasting ≥24 hours or until death attributable to a vascular cause 35 ) ( ICD‐8 codes 430–437). For reports of stroke for which medical records were not available/permitted, strokes were corroborated by nurse participant information. Information on physician‐diagnosed coronary artery bypass grafting, angioplasty, and stent was based on self‐report. Time‐varying Cox proportional hazard models were jointly stratified by age (months) and calendar time (months) and used to calculate the hazard ratio (HR) of CVD. The proportional hazard assumptions were tested and met. Person‐months at risk were calculated from age at return date of the questionnaire in which the participant reported either first report of infertility or first pregnancy until incident CVD incidence; death; or June 30, 2017. Participants who never reported a pregnancy attempt lasting >12 months or who never reported a pregnancy were excluded (n=10 455) ( FigureFigure ). Participants who reported having experienced MI or stroke before enrollment into the NHSII in 1989 were excluded at baseline (n=875), as were participants who never returned subsequent questionnaires (n=1370). Our primary analysis compared women with a history of any type of infertility (exposed) with gravid women with no history of infertility (unexposed). In secondary analyses, we investigated differences by different self‐reported causes of infertility. Multivariable model 1 adjusted for potential confounding factors, including age at menarche (<12, 12, 13, ≥14 years), White race (yes, no), marital status (ever/currently married, never), BMI at age 18 years (<19, 20.5–21.9, 22–24.9, 25–29.9, ≥30 kg/m 2 ), time‐varying daily aspirin use (yes, no), time‐varying oral contraceptive use (never, past, current), time‐varying gravidity (≤1, 2, 3, ≥4), and time‐varying breastfeeding duration (12 months). Model 2 additionally adjusted for cardiovascular risk factors, including time‐varying covariates (updated every 2–4 years) of BMI (<24.9, 25–29.9, 30–34.9, or ≥35 kg/m 2 ), smoking status (never, former, current 1–34 cigarettes/d, or current ≥35 cigarettes/d), physical activity (0, 0.1–1.0, 1.1–2.4, 2.5–5.9, or ≥6 h/wk), and Alternative Healthy Eating Index 2010 diet quality score (quintiles). For covariates with missing data, a missing indicator variable was created. Cumulative missingness ranged from <1% for BMI at age 18 to 10% for diet. We additionally investigated the impact of timing of infertility experience on CVD risk by classifying infertility history on the basis of age at reporting first infertility (≤25, 26–30, >30 years), as well as investigating the association with specific infertility diagnoses. We examined effect modification by smoking status (never, ever/current), time‐varying BMI (<25, ≥25 kg/m 2 ), nulliparity (yes, no), history of hypertensive disorders of pregnancy (yes, no), and BMI at age 18 years (<25, ≥25 kg/m 2 ) by testing for interaction using likelihood ratio tests. In sensitivity analyses, we stratified by whether infertility was primary (occurring before gravidity) or secondary (occurring after proven gravidity). Additionally, in sensitivity analyses, we restricted our infertility definition to women experiencing infertility before age 37 36 and excluded women reporting infertility only attributable to male partners. Finally, we conducted mediation analyses to quantify the proportion of the association between infertility and coronary heart disease (CHD) mediated by specific covariates 37 that may be on the causal pathway between infertility and CHD, including type 2 diabetes, hypertension, and hypercholesterolemia. 38 , 39

In our study sample, 28 611 women reported having ever experienced infertility (27.6%) (Table  1 ), and there were 782 incident MIs, 984 incident revascularization events, and 762 incident stroke events (210 ischemic). Compared with gravid women without a history of infertility, women with a history of infertility were less likely to be gravid or parous at cohort baseline in 1989, but were more likely to have experienced gestational diabetes or hypertensive disorders of pregnancy, to be obese, to have a parent with a history of myocardial infarction or stroke, and to be never smokers. Among those who had their infertility investigated, the cause of infertility was attributed to ovulatory disorders (39.1%), spouse/male factor infertility (25.7%), other cause (21.8%), endometriosis (21.5%), or cause not found (30.5%). Age‐Standardized Baseline (1989) Characteristics Among 103 729 Women in the Nurses' Health Study II, by Infertility History Across Follow‐Up * Values are given as means (SDs) or percentages and are standardized to the age distribution of the study population. Value is not age‐adjusted. Aspirin or aspirin‐containing products used at least once per week in the past 2 years. Among parous participants. Attributed causes of infertility, which were not mutually exclusive, were gathered at baseline and during follow‐up among those who reported having tried to conceive unsuccessfully for at least 12 months up to age 40. N=1011 (3.5%) did not report if they had had an investigation or not or what the cause was attributed to if investigated. We observed that women with a history of infertility were at greater risk of CHD in age‐adjusted models (HR, 1.16 [95% CI, 1.05–1.29]) and multivariable adjusted models (risk ratio [RR], 1.13 [95% CI, 1.01–1.26]) (Table  2 ). This association was predominantly driven by an association with the end point of coronary revascularization (HR, 1.18 [95% CI, 1.03–1.35]), with an attenuated, not statistically significant association between infertility and risk of MI (HR, 1.08 [95% CI, 0.93–1.27]). We observed no association between history of infertility and risk of stroke (HR, 0.91 [95% CI, 0.77–1.07]). Hazard Ratios and 95% CIs for the Risk of CVD According to History of Infertility Among 103 729 Women (NHSII, 1989–2017) BMI indicates body mass index; CHD, coronary heart disease; CVD, cardiovascular disease; ICD‐8, International Classification of Diseases, Eighth Revision ; MI, myocardial infarction; and NHSII, Nurses’ Health Study II. Models were jointly stratified by age (months) and calendar time (months). Model 1 additionally adjusted for age at menarche (<12, 12, 13, ≥14 years of age), White Non‐Hispanic race and ethnicity (yes, no), marital status (ever married, never), daily aspirin use (yes, no), oral contraceptive use (never, past, current), gravidity (≤1, 2, 3, or ≥4 pregnancies), BMI at age 18 years (<19, 20.5–21.9, 22–24.9, 25–29.9, or ≥30 kg/m 2 ), and breastfeeding duration (12 months). Model 2 is additionally adjusted for time‐varying BMI (<24.9, 25–29.9, 30–34.9, or ≥35 kg/m 2 ), cigarette smoking status and current dose (never, past, current 1–34 cigarettes/d, current ≥35 cigarettes/d), physical activity (0, 0.1–1.0, 1.1–2.4, 2.5–5.9, or ≥6 h/wk), and Alternative Healthy Eating Index 2010 diet quality score (quintiles). Fatal and nonfatal MI and fatal coronary heart disease ( ICD‐8 : 410, 412). Self‐reported coronary artery bypass grafting/angioplasty/stent. Fatal and nonfatal cerebrovascular accident or transient ischemic attack ( ICD‐8 : 430–437). The association between infertility and risk of CHD was strongest among women who reported experiencing infertility at younger ages (Table  3 ). Women who were ≤25 years old when they first experienced infertility had a 26% greater risk of CHD compared with gravid women who never experienced infertility (RR, 1.26 [95% CI, 1.09–1.46]), while women who first experienced infertility after age 30 years were not at greater risk of CHD (RR, 0.91 [95% CI, 0.70–1.19]). Hazard Ratios and 95% CIs for the Risk of CHD and Stroke According to Age at First Report of Infertility Among 103 729 Women (NHSII, 1989–2017) BMI indicates body mass index; CHD, coronary heart disease; CVD, cardiovascular disease; ICD‐8, International Classification of Diseases, Eighth Revision ; MI, myocardial infarction; and NHSII, Nurses’ Health Study II. Models were jointly stratified by age (months) and calendar time (months). Model 2 additionally adjusted for age at menarche (<12, 12, 13, ≥14 years of age), White Non‐Hispanic race and ethnicity (yes, no), marital status (ever married, never), daily aspirin use (yes, no), oral contraceptive use (never, past, current), gravidity (≤1, 2, 3, or ≥4 pregnancies), BMI at age 18 years (<19, 20.5–21.9, 22–24.9, 25–29.9, or ≥30 kg/m 2 ), and breastfeeding duration (12 months). Model 3 is additionally adjusted for current BMI (<24.9, 25–29.9, 30–34.9, or ≥35 kg/m 2 ), cigarette smoking status and current dose (never, past, current 1–34 cigarettes/d, current ≥35 cigarettes/d), physical activity (0, 0.1–1.0, 1.1–2.4, 2.5–5.9, or ≥6 h/wk), and Alternative Healthy Eating Index 2010 diet quality score (quintiles). Tests for linear trend were conducted by modeling categories of age at first report of infertility as an ordinal variable assigning integer values. CHD includes fatal and nonfatal MI, fatal coronary heart disease ( ICD‐8 : 410, 412), and self‐reported coronary artery bypass grafting/angioplasty/stent. CVD includes CHD and fatal and nonfatal cerebrovascular accident or transient ischemic attack ( ICD‐8 : 430–437). We observed heterogeneity in the association with CHD by the underlying cause of infertility (Table  4 ). Women whose infertility was attributed to ovulatory disorders had a 28% greater risk of CHD (RR, 1.28 [95% CI, 1.05–1.55]), and women whose infertility was attributed to endometriosis had a 42% greater risk of CHD (RR, 1.42 [95% CI, 1.09–1.85]) compared with gravid women with no history of infertility. No other infertility diagnoses were associated with the risk of CHD. Hazard Ratios and 95% CIs for the Risk of CHD According to Conditions That Were Reported to be the Cause of Infertility (NHSII, 1989–2017) CHD includes fatal and nonfatal MI ( ICD‐8 : 410), fatal coronary heart disease ( ICD‐8 : 412), and self‐reported coronary artery bypass grafting/angioplasty/stent. BMI indicates body mass index; CHD, coronary heart disease; ICD‐8, International Classification of Diseases, Eighth Revision ; MI, myocardial infarction; and NHSII, Nurses’ Health Study II. Models were jointly stratified by age (months) and calendar time (months). Model 2 additionally adjusted for age at menarche (<12, 12, 13, ≥14 years of age), White Non‐Hispanic race and ethnicity (yes, no), marital status (ever married, never), daily aspirin use (yes, no), oral contraceptive use (never, past, current), gravidity (≤1, 2, 3, or ≥4 pregnancies), BMI at age 18 years (<19, 20.5–21.9, 22–24.9, 25–29.9, or ≥30 kg/m 2 ), and breastfeeding duration (12 months). Model 3 is additionally adjusted for current BMI (<24.9, 25–29.9, 30–34.9, or ≥35 kg/m 2 ), cigarette smoking status and current dose (never, past, current 1–34 cigarettes/d, current ≥35 cigarettes/d), physical activity (0, 0.1–1.0, 1.1–2.4, 2.5–5.9, or ≥6 h/wk), Alternative Healthy Eating Index 2010 diet quality score (quintiles), and mutually adjusted for other infertility diagnoses. Comparison group is gravid women without a history of any kind of infertility. Attributed causes of infertility are not mutually exclusive. We observed no statistically significant effect modification by smoking status, parity, hypertensive disorders of pregnancy, or BMI at age 18 years (Table  S1 ). However, we did observe differences in the association between infertility and CHD by updating adult BMI ( P value, test for heterogeneity=0.005). Among women with BMI ≥25 kg/m 2 , infertility was associated with an increased risk of CHD (RR, 1.24 [95% CI, 1.09–1.41]). No association was observed between infertility and CHD among women with a BMI <25 kg/m 2 (RR, 0.88 [95% CI, 0.70–1.11]). In sensitivity analyses, we observed no meaningful differences between individuals with primary compared with secondary infertility (Table  S2 ). Our findings also did not appreciably change in analyses in which infertility was not updated after age 37 years or when we excluded participants with only infertility attributed to male factor (Table  S3 ). We observed that 16.3% (95% CI, 3.6%–50.2%) of the association between infertility and risk of CHD was mediated through development of type 2 diabetes; however, CIs were wide, while the history of hypertension and hypercholesterolemia did not meaningfully mediate the association. When we investigated infertility diagnoses and risk of stroke (Table  S4 ), endometriosis was associated with a 27% increased risk of stroke; however, it did not reach the threshold of statistical significance, nor did any other infertility diagnoses. In sensitivity analyses restricted to women with ovulatory infertility, we did not observe meaningful differences between women who did and did not express common PCOS symptoms and risk of CHD (Table  S5 ).

This work was supported by grants HD096033 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development; and U01 CA176726 and U01 HL145386 from the National Cancer Institute.

Overall, we observed that women with infertility were at increased risk of CHD, and this risk was strongest among women who experienced infertility at a younger age. When we investigated differences by cause‐specific infertility, we observed that the greatest risk of CHD was observed among women whose infertility was attributed to ovulatory disorders or endometriosis. The association between infertility and CHD was observed only among women who were overweight or obese and not among women who were underweight or normal weight. Prior research has suggested that reproductive factors may be associated with the risk of CVD. For example, nulliparous women and women who had 1 birth had a modestly increased risk of cardiovascular disease compared with women with 2 births (0 births: RR, 1.11; 1 birth: RR, 1.10). 4 Prior research from our group observed that women who experienced pregnancy loss had an increased risk of CVD (HR, 1.21) 8 and CVD risk factors. 40 Research from the Women's Health Initiative observed a similar pattern between pregnancy loss, CVD risk (HR, 1.11), and CVD risk factors (hypertension, type 2 diabetes) after 16 years of follow‐up 6 . However, there has been limited prior research into the relationship between overall infertility and the risk of cardiovascular disease. 29 One of the largest studies to investigate this question used data from the Swedish Birth Registry. 17 They observed that years of trying to conceive without success, potentially a marker for infertility severity, was associated with a risk of CVD. Parous women who had experienced ≥5 years of subfertility had a 19% greater risk of CVD in models adjusted for birth year, age at first delivery, education level, income, country of birth, hypertension, diabetes, preterm birth, small for gestational age birth, smoking, and parity compared with other parous women. A similar pattern of association has been observed in smaller cross‐sectional analyses using the National Health and Nutrition Examination Survey (OR, 1.71), 19 in the Women's Health Initiative (HR, 1.13), 11 Trøndelag Health Study (HR, 1.10), 41 and among Norwegian women who experienced a longer time to conception (HR, 1.14). 18 In the current study, we observed that women with infertility were at a modestly increased risk of CHD (HR, 1.13) and that this association was strongest among women who reported having experienced infertility early in life (HR for women experiencing infertility ≤25 years, 1.26). Infertility at a younger age may be a marker of infertility severity, which is consistent with findings from the Swedish Birth Registry that observed elevated CVD risk among women with infertility lasting >5 years. 17 Notably, we did not observe an association between infertility and risk of stroke (HR, 0.91 [95% CI, 0.77–1.07]). A recent meta‐analysis that included 5 prior studies on infertility and risk of stroke concluded that existing research was inconsistent and inconclusive (meta‐analysis HR, 1.07 [95% CI, 0.87–1.32]). 42 Indeed, subsequent analyses have also produced mixed findings. The InterLACE pooling project observed a 14% increased risk of stroke, 43 while a subsequent analysis by the SWAN (Study of Women's Health Across the Nation) observed no association for their composite CVD measurement (stroke, MI, angina: HR, 0.79 [95% CI, 0.52–1.21]). 44 Differences in findings may be influenced by variability in exposure definitions (eg, experiencing infertility versus using treatment for infertility, which is also influenced by access to care), outcome definitions (ischemic versus hemorrhagic stroke, fatal versus nonfatal stroke), and differences in the demographics of the population sampled. Indeed, focusing on ischemic stroke only, prior research among women enrolled in the Taiwan National Health Insurance program observed a protective effect for women using infertility treatment (HR, 0.82) 45 ; however, a harmful effect was observed for women with a history of infertility in the InterLACE pooling project (HR, 1.15). 43 Research from the Framingham Heart Study observed that women with infertility had increased risk factors for CVD, such as higher BMI, waist circumference, and triglyceride levels, and lower levels of high‐density lipoprotein cholesterol. 15 While we were able to a priori adjust for adult BMI and BMI at age 18 years, as well as take into account high cholesterol in mediation analyses, we observed that the association between infertility and risk of CHD varied by BMI ( P value, test for heterogeneity=0.005). Among women with an overweight or obese BMI, women with a history of infertility were at greater risk of CHD (HR, 1.24 [95% CI, 1.09–1.41]); however, we observed no association among women with a BMI <25 kg/m 2 , suggesting that BMI may be an important modifiable risk factor for women with infertility. These data also suggest that history of infertility among women with large body size may be a marker of insulin insensitivity and could be a target for early therapeutics or intervention. When investigating specific infertility diagnoses, we observed that the risk of CHD was greatest among women whose infertility was attributed to ovulatory disorders (HR, 1.28) or endometriosis (HR, 1.42). The diagnosis of ovulatory infertility includes the diagnosis of PCOS, which is known to be associated with cardiometabolic outcomes. A recent meta‐analysis investigating the association between PCOS and cardiovascular disease observed that women with PCOS had a greater risk of nonfatal cerebrovascular disease (8 studies; RR, 1.41 [95% CI, 1.02–1.94]) and nonfatal coronary events (6 studies; HR, 1.78 [95% CI, 0.99–3.23]). 28 Women with PCOS were also observed to have a greater risk of hypertension (RR, 1.75), type 2 diabetes (RR, 3.00), and adverse lipid profiles (higher serum concentrations of total cholesterol and lower serum concentrations of high‐density lipoprotein). 28 PCOS may also be an extreme manifestation of the more common symptom of menstrual cycle irregularity that has been associated with CVD risk and CVD mortality. 46 Women with endometriosis have also been shown in prior research to have an elevated risk of hypertension, 26 hypercholesterolemia, 26 , 47 , 48 , 49 and subclinical atherosclerosis 50 compared with women without endometriosis, suggesting an adverse vascular profile among these women. 51 , 52 Additionally, data from our research team observed a 62% greater risk of CHD 27 and a 34% greater risk of stroke 53 among women with laparoscopically confirmed endometriosis independent of history of infertility. Diagnoses associated with infertility may therefore confer risk of CHD through different mechanisms; PCOS manifests through ovulatory dysfunction and hyperandrogenic milieu, 30 , 54 while endometriosis has been associated with higher levels of systemic inflammation and changes in endogenous hormones, 55 , 56 which may influence CVD risk. Future research into infertility and CHD should further investigate informative heterogeneity by infertility diagnoses and infertility severity. A recent scientific statement from the American Heart Association 57 underscores the importance of knowing patient history of adverse pregnancy outcomes and early menopause when counseling patients for prevention of CVD events and appropriate therapeutic interventions for CVD. Clinically, our findings suggest the need to pay attention to female‐specific risk factors when studying CHD. If our findings are robustly replicated, then information on reproductive and gynecologic history should be discussed with medical providers to better counsel patients on future CHD risk. Although this study has many strengths, including its large sample size, longitudinal follow‐up, and detailed information on infertility symptoms, we must also recognize its limitations. Information on infertility and infertility diagnoses and confounding variables (eg, BMI) were based on self‐report and therefore may be prone to misclassification; however, prior research has suggested that women are able to recall these with high accuracy. 32 , 33 Therefore, we would expect any misclassification to be nondifferential with respect to our outcome and therefore attenuate any associations. The population under study was relatively homogenous in terms of racial or ethnic background and education level; therefore, these results may not be generalizable to other groups who may have different risks for cardiometabolic outcomes. However, this may increase internal validity and reduce the risk of residual confounding by these covariates. For covariates with missing data, a missing indicator variable was used, which will bias the results if data are missing not at random. We lacked sufficient sample size to be able to differentiate between ischemic and hemorrhagic stroke, which may have different mechanisms of association with infertility. In summary, we observed that women with a history of infertility had greater risk of CHD, and this risk was greatest among women with infertility attributed to ovulatory problems or endometriosis. These findings are consistent with other studies that have suggested that women with severe infertility, PCOS, and endometriosis have poorer cardiometabolic health. Future research should continue to investigate informative heterogeneity by infertility symptom presentation, as this may lead to a better understanding of mechanisms of association and identify groups that may benefit from targeted screening or interventions. Additionally, future research should investigate the contribution of knowing infertility history on top of established CVD risk factors and disentangling the mechanism of association between infertility and CVD.

None.

Tables S1–S5 Click here for additional data file.