Sex Differences in Age of Onset of Premature Cardiovascular Disease and Subtypes: The Coronary Artery Risk Development in Young Adults Study

Abstract

Background

Historical data indicate men develop coronary heart disease (CHD) 10 years before women. However, whether this sex gap persists in a contemporary sample amid changing cardiometabolic risk profiles, and whether differences exist for other cardiovascular disease (CVD) subtypes (ie, stroke, heart failure), is not known.

Methods

Data are from the CARDIA (Coronary Artery Risk Development in Young Adults) study, a prospective multicenter cohort study. US adults aged 18 to 30 years enrolled in 1985 to 1986 and were followed through August 2020. Sex differences in the cumulative incidence functions of premature CVD (onset <65 years), overall and for each subtype (CHD, heart failure, stroke), were compared using Gray’s test.

Results

Among 5112 participants (54.5% female, 51.6% Black) with a mean age of 24.8 years (SD: 3.7) at enrollment and a median follow‐up of 34.1 years (interquartile range, 33.8–35.7), men had a significantly higher cumulative incidence of CVD, CHD, and heart failure (P<0.05 for all), with no difference in stroke (P=0.63). Men reached 5% incidence of CVD 7.0 years earlier than women (50.5 versus 57.5 years, P<0.001). CHD was the most frequent CVD subtype, and men reached 2% incidence 10.1 years earlier than women (P<0.001). Men and women reached 2% stroke and 1% heart failure incidence at similar ages. Ten‐year CVD event rates diverged at an index age of 35.

Conclusions

Men developed CVD earlier than women, with the greatest difference observed for CHD. Sex differences in CVD risk emerged at age 35, persisted through midlife, and were not attenuated by accounting for cardiovascular health.

Nonstandard Abbreviation and Acronym

CARDIA

Coronary Artery Risk Development in Young Adults studynsion

Clinical Perspective.

What Is New?

• Sex differences in cardiovascular disease risk emerge by age 35, persist through midlife, and remain even after accounting for differences in cardiovascular health.

What Are the Clinical Implications?

• Cardiovascular disease risk differences emerging in the fourth decade of life supports initiating cardiovascular risk assessment and risk reduction strategies in young adulthood.

• Earlier onset of cardiovascular disease, and particularly coronary heart disease, highlights the need to initiate strategies to address atherosclerosis more intensively in men.

Cardiovascular disease (CVD) is the leading cause of death for both men and women in the United States, ¹ though the age of onset differs by sex. Historical estimates suggest men experience earlier onset of coronary heart disease (CHD) by about 10 years as compared with women. ² However, much of the prior literature has focused on CHD, and whether this difference persists for other types of CVD, including heart failure and stroke, is unknown.

Sex‐specific differences in CVD are attributed to multiple different pathways, including hormonal influences, differences in cardiovascular health behaviors and factors, and exposure to adverse social determinants of health. Historically, men had higher rates of smoking, diabetes, and hypertension. However, population shifts in cardiometabolic risk phenotypes have resulted in similar or higher rates of obesity, diabetes, and hypertension in women than men. ³ , ⁴ , ⁵ Additionally, the overall prevalence of smoking has decreased and is similar among men and women. ⁶ Thus changes in the distribution of CVD risk factors may have shifted the male–female age gap in incident CVD and its subtypes. Moreover, data suggest growing incidence of premature myocardial infarction among women. ⁷ As a result, sex‐based differences in CVD onset may have narrowed, but data on the magnitude of the narrowing and causes of potential narrowing in age of onset by sex are limited. ² , ⁸

To assess sex differences in timing of CVD incidence, we leverage a contemporary cohort, the biracial CARDIA (Coronary Artery Risk Development in Young Adults) study. The relatively young age of participants at enrollment allows us to investigate with precision when risk for CVD diverges. Our analysis has 3 key objectives. First, we describe sex differences in age of onset for premature CVD (defined as onset <65 years) and CVD subtypes. Second, we estimate 10‐year CVD event rates through midlife for men and women and quantify sex differences in rates. Third, we evaluate whether differences in cardiovascular health measures attenuate sex‐based differences in CVD risk.

METHODS

Study Sample

Data are from the CARDIA study. CARDIA is a prospective cohort study that recruited 5115 Black and White participants between the ages of 18 and 30 across 4 cities (Birmingham, AL; Chicago, IL; Minneapolis, MN; Oakland, CA) from March 1985 to June 1986. At each study site, enrollment was balanced by age, sex, race, and education. Study data, including health measures and surveys, were collected at baseline and participants completed subsequent examinations at study years 2, 5, 7, 10, 15, 20, 25, and 30. Retention rates among surviving participants at each examination were 91%, 86%, 81%, 79%, 74%, 72%, 72%, and 71%, respectively. Contact is maintained with participants via telephone, mail, or electronically every 6 months, with annual interim medical history ascertainment. Over the last 5 years, >90% of the surviving cohort members have been directly contacted, and follow‐up for vital status is virtually complete through related contacts and intermittent National Death Index searches.

CARDIA was approved by the institutional review board at each study site, and participants provided written informed consent. Additional details on the study design and methods have been reported elsewhere. ⁹ For the present analysis, 3 participants were excluded from the analysis: 1 individual who withdrew consent and 2 who reported change in gender identity.

Data Availability

CARDIA data are available upon reasonable request from the CARDIA Coordinating Center. CARDIA investigators are eager to collaborate with investigators interested in using CARDIA data. Please see the CARDIA website (http://www.cardia.dopm.uab.edu/publications‐2/publications‐documents) for publications policies and for a list of CARDIA Representatives. CARDIA data are also publicly available on the National Institutes of Health‐supported BioLINCC and dbGaP platforms.

Sex

Participants self‐reported sex (male, female) at study enrollment.

Cardiovascular Disease

Incident CVD was defined as fatal or nonfatal myocardial infarction, stroke, heart failure, hospitalized acute coronary syndrome not resulting in myocardial infarction, transient ischemic attack, coronary revascularization, carotid or peripheral arterial disease requiring intervention, or other fatal atherosclerotic or nonatherosclerotic heart disease. CVD was also categorized into 3 subtypes: CHD (includes myocardial infarction, hospitalized acute coronary syndrome not resulting in myocardial infarction, coronary revascularization, or other fatal atherosclerotic heart disease), stroke, and heart failure. CVD events are ascertained at each study visit and through annual contacts, where participants or their proxies are asked about hospital admissions and outpatient procedures. Deaths were identified through proxies, internet searches, and queries to the National Death Index. Following identification of potential CVD events and deaths, research staff collected medical records (admitting medical history, physical exam documentation, laboratory work, results of all diagnostic procedures, etc.), death certificates (reviewed by a nosologist to assign standardized International Classification of Diseases, Ninth Revision (ICD‐9) codes to causes of death), and other documents. These materials were independently reviewed by 2 physicians from the CARDIA Outcomes Surveillance and Adjudication Subcommittee to classify CVD events using standard definitions. ¹⁰ Additional details on outcome ascertainment are described in the CARDIA Endpoint Events Manual of Operations. ¹¹ Events ascertained through August 31, 2020 are included in the analysis. Given participants’ age at baseline through follow up, CVD events were considered premature based on onset <65 years.

Cardiovascular Health

We evaluated 7 health behaviors and health factors representing cardiovascular health. Health behaviors include diet, physical activity, and smoking status. Diet was assessed at baseline and study years 7 and 20 using a questionnaire on dietary intake over the previous 28 days, covering 1609 unique food items. ¹² , ¹³ The Dietary Approaches to Stop Hypertension diet score was calculated based on average intakes of 8 dietary groups: fruits, vegetables, whole grains, dairy, nuts and legumes, red and processed meats, sodium, and sweetened beverages. For each dietary group, participants were assigned scores from 1 to 5 based on sex‐specific quintiles, and these scores were summed to create a total score ranging from 8 (low adherence to Dietary Approaches to Stop Hypertension diet) to 40 (high adherence). Participants completed self‐administered questionnaires regarding smoking status at every examination. Physical activity was assessed at each examination based on the frequency of participation in 13 moderate‐ and vigorous‐intensity activities over the previous 12 months and summarized as exercise units. ¹⁴ , ¹⁵

Health factors include body mass index, blood pressure, blood lipids, and blood glucose. Body mass index was calculated based on height in cm, measured using a vertical ruler, and weight in kg, measured at every visit using a calibrated scale. Blood pressure was measured in triplicate at each examination and summarized based on the average of the last 2 measures. A random‐zero sphygmomanometer was used for study years 0 to 15 and an automated oscillometric blood pressure monitor was used for study years 20 to 30, with measures across examinations subsequently standardized to sphygmomanometer measures. Fasting lipids were measured from ethylene‐diaminetetraacetic acid plasma at every examination. High‐density lipoprotein cholesterol was quantified after precipitation with dextran sulfate‐magnesium chloride. ¹⁶ Non‐high‐density lipoprotein cholesterol was calculated as the difference between total cholesterol and high‐density lipoprotein cholesterol. Fasting serum glucose was measured using the hexokinase ultraviolet method at baseline and using hexokinase coupled to glucose‐6‐phosphate dehydrogenase for study years 7 to 30 (not measured at study years 2 and 5). Values were standardized across examinations.

All 7 metrics were scored based on the American Heart Association’s Life’s Essential 8. ¹⁷ Sleep, the eighth measure included in Life’s Essential 8, was not measured at baseline and therefore not included in the analysis. Scores for each measure range from 0 to 100, with higher scores indicating better cardiovascular health. In addition, composites were calculated to represent overall cardiovascular health, health behaviors, and health factors by taking the average of the relevant scored components.

Covariates

Additional covariates included baseline measures of age and sociodemographic factors, including race (self‐reported as Black or White and represents a social construct) and education (highest completed grade). Given the age of the sample at enrollment (18–30 years), we also considered maximum educational attainment, as reported across follow‐up study visits, as a better indicator of socioeconomic position.

Statistical Analysis

Sex Differences in the Cumulative Incidence of Premature CVD

Sex‐specific cumulative incidence functions for CVD and CVD subtypes were estimated with age as the time scale. Participants were censored at the time of the most recent contact and non‐CVD death before incident CVD was considered a competing risk. Sex differences in the overall cumulative incidence functions were compared using Gray’s test to account for the competing risk of non‐CVD death. We also evaluated and compared male–female differences in cumulative incidence at specified ages of 50 and 65 years, and we compared the age at which men and women reached specified incidences (5% for total CVD, 2% for CHD and stroke, and 1% for heart failure). Permutation testing was used to assess whether male–female differences in age at the specified incidence were statistically significant (n=10 000 permutations). Models were not adjusted for covariates. However, in a secondary analysis, we examined sex differences in cumulative incidence functions by social determinants of health, including maximum participant education (some college or less versus 4‐year college degree or more) as a proxy for socioeconomic position and race as a proxy for exposure to structural and social determinants of health.

Sex Differences in 10‐Year CVD Event Rates

To quantify sex‐specific differences in premature CVD event rates while allowing for age‐dependent change, we sequentially estimated 10‐year event rates using the methods previously described by Selvaraj et al. ¹⁸ In accordance with this approach, we first created a series of data sets representing 10‐year periods, starting from ages 25 (median age at the first examination) to 56 (median age at the most recent examination). Each data set includes only participants who were in the risk set at the start of the period, with participants censored at the end of the 10‐year period, at the age at their last study contact, or when a competing event occurred—whichever came first. For each 10‐year period, participants who experienced an event or were censored before the starting age of the period were excluded.

For example, the first data set in the sequence reflects participants from the ages 25 to 35, with censoring at 35. The last data set in the sequence reflects participants from the ages of 56 to 66, excluding those who experienced an event or were censored before age 56. For each data set in the sequence, we used Kaplan–Meier methods, stratified by sex, to estimate 10‐year CVD event rates for men and women and calculate corresponding male–female rate differences. This analysis was then repeated for CHD.

We also conducted 2 sensitivity analyses. First, recognizing that Kaplan–Meier methods can overestimate risk in the presence of competing events, ¹⁹ we extended this approach to estimate cumulative incidence functions, accounting for the competing risk of non‐CVD death. In the second sensitivity analysis, we estimated hazard ratios for men as compared with women using Cox proportional hazards models to evaluate whether observed patterns were consistent on a relative scale.

Cardiovascular Health and Sex Differences in CVD Risk

To evaluate whether differences in cardiovascular health explain male–female differences in CVD risk, we estimated a series of Cox proportional hazards models. The base model for estimating the hazard ratio (HR) comparing men to women included sex, age at study entry (to account for potential differences in age‐related CVD risk at baseline), and study site (to account for potential heterogeneity in participant characteristics and health care practices). ²⁰ In a subsequent round of models, we adjusted for each of the 7 cardiovascular health measures to evaluate whether sex differences in CVD risk persisted independent of differences in cardiovascular health. Models included time‐dependent covariates to account for changes in cardiovascular health across the follow‐up period, as measured across examinations. To quantify the extent to which each cardiovascular health component explained the observed male–female differences, we compared the coefficient reflecting the log of the HR contrasting men and women from the base model with the corresponding coefficients from the models including individual cardiovascular health components. Consistent with prior work in this sample assessing the contribution of clinical factors to racial differences in CVD risk, ²⁰ this comparison was quantified as the percent reduction in the estimate using the formula:

$$\left(\frac{{\widehat{\beta }}_{\mathit{\text{BASE}}}-{\widehat{\beta }}_{\mathit{NEW}}}{{\widehat{\beta }}_{\mathit{\text{BASE}}}}\right)\times 100$$

where ${\widehat{\beta }}_{\mathit{\text{BASE}}}$ reflects the estimate comparing the sexes from the base model and ${\widehat{\beta }}_{\mathit{NEW}}$ reflects the corresponding estimate from a model that additionally includes a component of cardiovascular health. A larger percentage of reduction indicates that adjustment for that cardiovascular health measure accounts for a greater share of the male–female difference in CVD risk.

Missing data were imputed using multiple imputation by chained equations with the mice package in R. ²¹ Data from all examinations were included such that each imputed cardiovascular health value is predicted based on data from both prior and future examinations. Ten imputed data sets were created, and regression results were combined using Rubin’s rules. ²² In a sensitivity analysis, we replicated the analysis using the complete‐case data set with measured values carried forward, and in a secondary analysis, we considered only baseline cardiovascular health measures rather than time‐varying measures. Statistical significance was determined based on an alpha‐level of 0.05 and all analyses were performed using R version 4.3. ²³

RESULTS

Study Sample

Of the 5112 participants included in the analysis, 2785 (54.5%) were women and 2327 (45.5%) were men. The mean age at the baseline was 24.9 years (SD: 3.7) for women and 24.8 years (SD: 3.6) for men (Table 1). Among women, 51.1% completed a 4‐year college degree or more in comparison to 45.3% among men. Over a median follow‐up of 34.1 years (25th–75th percentile: 33.8–35.7), there were 387 CVD events, with 160 among women and 227 among men.

Table 1.

Demographic Characteristics for the Study Sample and Stratified by Sex (n=5112)

No. (%) or mean±SD	Female participants (n=2785)	Male participants (n=2327)
Age, baseline, y	24.9±3.7	24.8±3.6
Race
Black	1480 (53.1%)	1157 (49.7%)
White	1305 (46.9%)	1170 (50.3%)
Education, maximum attained
Some college or less (<16 y)	1360 (48.9%)	1272 (54.7%)
4‐year degree or more (≥16 y)	1422 (51.1%)	1054 (45.3%)
Life’s Essential 8 score (out of 100), baseline	76.5±12.8	75.4±11.3
Body mass index, baseline, kg/m2	24.6±5.8	24.4±4.0
Systolic blood pressure, baseline, mm Hg	106.6±9.9	115.0±10.4
Moderate‐to‐vigorous physical activity, baseline, exercise units	334.9±250.5	522.2±323.3
Non‐high‐density lipoprotein cholesterol, baseline, mg/dL	121.7±32.8	125.8±35.6
Fasting glucose, baseline, mg/dL	81.1±18.1	84.4±13.0
Current smoker, baseline	816 (29.4%)	730 (31.7%)
Dietary Approaches to Stop Hypertension diet score, baseline	24.9±5.7	23.1±5.1

Sex Differences in Cumulative Incidence of Premature CVD

Sex‐specific cumulative incidence functions for premature total CVD and CVD subtypes are presented in Figure 1. Cumulative incidence functions for men and women were statistically significantly different for total CVD, with men reaching a cumulative incidence of 5% earlier than women by 7.0 years (P value <0.001; Table S1). By age 50, the cumulative incidence of CVD was 4.7% among men and 2.9% among women (P value=0.001; Table S2). For CHD, men reached a cumulative incidence of 2% earlier than women by 10.1 years (P value <0.001; Table S1), with a cumulative incidence of 2.5% by age 50 among men as compared with 0.9% among women (P‐value <0.001; Table S2). Sex‐specific patterns in cumulative incidence were consistent in subgroup analyses when stratified by education and race for both total CVD (Figure S1) and CHD (Figure S2).

Figure 1. Cumulative incidence for total cardiovascular disease, coronary heart disease, stroke, and heart failure.

Panels display cumulative incidence stratified by sex (n=2327 men and 2785 women). P values from Gray’s test.

Cumulative incidence for stroke did not differ between men and women (Figure 1), with men and women reaching a cumulative incidence of 2% within 0.5 years of each other (P value: 0.90; Table S1). Similarly, the cumulative incidence of stroke by age 50 was 1.2% among both men and women (P value: 0.94; Table S2). For heart failure, the cumulative incidence functions differed by sex (Figure 1), though the age at which men and women reached a 1% cumulative incidence was not significantly different (men: 48.7, women: 51.7, P value: 0.31; Table S1). Sex‐specific differences in heart failure emerged at later ages, with similar cumulative incidence at age 50 (men: 1.2%, women: 0.9%, P value: 0.36), and a higher incidence among men by age 65 (men: 3.0%, women: 1.7%, P value: 0.017; Table S2). Patterns were consistent when accounting for antecedent CHD (Figure S3).

Sex Differences in 10‐Year CVD Event Rates

For both men and women, 10‐year CVD event rates increased with increasing index age (Figure 2). Among those CVD free at age 50, the 10‐year CVD event rate was 6.0% (95% CI, 4.8–7.2) for men and 3.3% (95% CI, 2.5–4.2) for women. Ten‐year CVD event rates for men and women statistically significantly diverged at an index age of 35, with a rate difference of 1.1% (95% CI, 0.4–1.9; Figure 2, Figure S4), and somewhat larger differences at later index ages. Results were similar for CHD, with the male–female difference in 10‐year CHD event rate significantly diverging at an index age of 34 (rate difference, 0.5% [95% CI, 0.06–0.9]; Figure 2) and remaining significantly different thereafter.

Figure 2. Ten‐year event rates for (A) total cardiovascular disease and (B) coronary heart disease and corresponding male–female differences for (C) total cardiovascular disease and (D) coronary heart disease.

A and B, 10‐year event rates and 95% CIs for men and women at index ages 25 to 56. C and D, the corresponding male–female differences and 95% CIs.

Results were consistent in a sensitivity analysis estimating sex‐specific cumulative incidence functions to account for competing risks, where a statistically significant difference in the 10‐year event rate emerged at an index age of 35 (Figure S5). Results were also consistent in a sensitivity analysis estimating HRs to quantify sex differences on a relative, rather than absolute, scale. Men had a statistically significantly greater 10‐year risk of CVD compared with women starting at an index age of 35 (HR, 1.95 [95% CI, 1.26–3.02]; Figure S6). On an absolute scale, male–female differences increased with increasing age, but on a relative scale, differences were relatively stable after age 35, with a 10‐year HR of 2.08 (95% CI, 1.39–3.09) at age 55.

Cardiovascular Health and Sex Differences in CVD Risk

In a series of sequential Cox proportional hazards models adjusted for time‐varying individual components of cardiovascular health and summary measures, men consistently had significantly higher risk for total CVD than women (Table 2). The greatest reduction in risk was observed after adjustment for systolic blood pressure, with a 15.0% reduction in the coefficient comparing men and women (corresponding male–female HR, 1.64 [95% CI, 1.33–2.01]). There was minimal reduction (2.9%) in the model adjusted for overall cardiovascular health score. Results were consistent in a complete case analysis using nonimputed data and in an analysis using only baseline cardiovascular health measures (Table S3).

Table 2.

Hazard Ratios for Total Cardiovascular Disease Comparing Men With Women (Reference Group) From Cox Proportional Hazards Models With Time‐Varying Covariates (n=5112)

	Male–female HR*	95% CI	% $\widehat{\beta }$ reduction†
Individual CVH components‡
Base model	1.79	1.46–2.19	Reference
Body mass index	1.84	1.50–2.26	−4.7
Blood pressure	1.64	1.33–2.01	15.0
Non‐HDL cholesterol	1.68	1.37–2.06	10.9
Blood glucose	1.65	1.35–2.03	14.0
Physical activity	1.92	1.56–2.36	−12.0
Smoking	1.75	1.43–2.14	3.9
Diet quality	1.65	1.34–2.03	14.0
Overall CVH score§	1.76	1.44–2.16	2.9
Health factors||	1.64	1.34–2.01	15.0
Health behaviors¶	1.84	1.50–2.26	−4.7

CVH indicates cardiovascular health; HDL, high‐density lipoprotein; and HR, hazard ratio.

^*Hazard ratios compare men and women (reference group); all models adjusted for age at baseline and study site.

^† Percentage of reduction in the male–female log hazard ratio after adjusting for individual cardiovascular health components.

^‡ Individual CVH factors scored based on the American Heart Association’s Life’s Essential 8.

^§ Overall CVH score calculated as the average of body mass index, blood pressure, non‐HDL cholesterol, blood glucose, physical activity, smoking status, and diet quality, as scored based on Life’s Essential 8.

^|| Health factors calculated as the average of scored body mass index, blood pressure, non‐HDL cholesterol, and blood glucose, as scored based on Life’s Essential 8.

^¶ Health behaviors calculated as the average of scored physical activity, smoking status, and diet quality, as scored based on Life’s Essential 8.

DISCUSSION

In this observational biracial cohort study beginning in young adulthood, men developed CVD on average 7 years earlier than women, which was largely due to sex differences in CHD incidence, with earlier onset by 10 years. For stroke and heart failure, age of onset and 10‐year event rates were similar for men and women. We quantify, for the first time, that sex differences in 10‐year CVD event rates first emerge at age 35 and persist throughout middle adulthood, on both absolute and relative scales. Moreover, the greater risk of CVD among men was not attenuated by adjustment for cardiovascular health status. These findings highlight that established sex differences in CVD risk in a contemporary cohort are largely due to earlier onset of CHD, which may inform the need for strategies to address atherosclerosis more intensively in men. Further, these data demonstrate that the fourth decade of life is a critical life period where CVD rates begin to diverge between men and women and may warrant more intensive screening and detection of subclinical CVD to guide intensification of preventive measures.

Our finding that men reach a 2% cumulative incidence of CHD 10 years earlier than women is consistent with the well‐established and commonly stated age gap of 10 years for CHD. Although this gap has been reported for decades, our findings highlight that temporal trends in cardiometabolic risk factors or increases in premature CVD have not affected the sex difference in CHD. ²⁴ Although this gap in CHD risk is often attributed to hormonal protection among women and worse lipid profiles among men, ²⁵ our findings suggest that even after accounting for differences in lipids, sex‐based differences in CHD persisted. Moreover, we saw small to no sex differences in incidence of other CVD subtypes that are becoming more prevalent. There was no evidence of differences in stroke and differences in heart failure emerged only at older ages. Although some evidence suggests sex differences in very premature ischemic stroke (before 45 years), ²⁶ the incidence among CARDIA participants in this age range was too low to evaluate. Additionally, these gaps may become more pronounced in late adulthood as both stroke and heart failure more commonly affect older adults. Alternatively, the smaller gap for heart failure may be explained by increases in obesity and the cardiovascular‐kidney‐metabolic phenotype, which are more strongly associated with risk for heart failure. ²⁷ In addition to shifting risk profiles, our finding that differences in cardiovascular health do not fully attenuate sex‐based differences in CVD risk highlights the need to move beyond traditional risk factors and consider a broader range of biological and social factors to understand the pathobiology of sex differences in CVD that are principally due to CHD differences.

There is increasing recognition that young adulthood is a critical period for promoting cardiovascular health and reducing the burden of CVD. ²⁸ In line with this, the development of the American Heart Association Predicting Risk of Cardiovascular Disease Events risk equations now enable prediction of CVD risk starting from age 30, ²⁹ down from the previous calculator, the pooled cohort equations, which started at age 40. ³⁰ This increasing focus on young adulthood is supported by our finding that sex‐based differences in 10‐year CVD rates are apparent by age 35. Notably, this is a life stage where there are large differences in health care use by sex. Among adults aged 18 to 44, the rate of preventive care visits for women is >4 times that of men, largely due to receipt of gynecologic and obstetric care. ³¹ , ³² Promoting preventive care among young adult men may be a key opportunity to prioritize CVD prevention in this critical life period. However, this study does not intend to diminish the importance of CVD prevention in women, as CVD is the leading cause of mortality and premature death in women. ³³

STRENGTHS AND LIMITATIONS

Strengths of the analysis include the large, prospective biracial cohort with adjudicated CVD events. Cardiovascular health measures were also evaluated using standardized methods with repeated assessments. However, this analysis has a few limitations. Due to the relatively young age of study participants and examination of premature CVD events only, the absolute incidence of some CVD subtypes was low, and we were unable to evaluate fatal and nonfatal events separately, or further stratify events, such as ischemic stroke vs. intracerebral hemorrhage. Similarly, based on the age and follow‐up of the cohort to date, we were unable to evaluate whether sex differences persist in late adulthood, particularly after menopause when there may be a narrowing in sex differences in CVD risk. ³³ Sleep was not directly prospectively assessed until later exams in CARDIA and, therefore, was not included as a measure of cardiovascular health. Residual confounding may also be a concern in analyses adjusted for cardiovascular health status. Finally, CARDIA included only Black and White participants, and results may not generalize to other groups.

CONCLUSIONS

In a contemporary cohort of young adults enrolled before the onset of CVD risk and with >30 years of follow‐up, men had a significantly higher risk of CVD with significantly earlier onset of CVD. Differences by sex were largely attributable to earlier onset of CHD, with no differences in stroke. Sex differences in CVD event rates emerged in the fourth decade of life and persisted through middle adulthood. The limited attenuation of sex differences after adjustment for cardiovascular health suggests a broader need to understand onset of premature CVD in men and the need to initiate risk assessment, subclinical CVD detection, and risk reduction strategies in the critical young adult life period.

Sources of Funding

Alexa A. Freedman is supported by the National Heart, Lung, and Blood Institute (K01HL165038). CARDIA (Coronary Artery Risk Development in Young Adults Study) is conducted and supported by the National Heart, Lung, and Blood Institute in collaboration with the University of Alabama at Birmingham (75N92023D00002 & 75N92023D00005), Northwestern University (75N92023D00004), University of Minnesota (75N92023D00006), and Kaiser Foundation Research Institute (75N92023D00003). This article has been reviewed by CARDIA for scientific content. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Disclosures

None.

Supporting information

Tables S1–S3

Figures S1–S5

STROBE Checklist