Incorporating longitudinal history of risk factors into atherosclerotic cardiovascular disease risk prediction using deep learning.

It is increasingly clear that longitudinal risk factor levels and trajectories are related to risk for atherosclerotic cardiovascular disease (ASCVD) above and beyond single measures. Currently used in clinical care, the Pooled Cohort Equations (PCE) are based on regression methods that predict ASCVD risk based on cross-sectional risk factor levels. Deep learning (DL) models have been developed to incorporate longitudinal data for risk prediction but its benefit for ASCVD risk prediction relative to the traditional Pooled Cohort Equations (PCE) remain unknown. Our study included 15,565 participants from four cardiovascular disease cohorts free of baseline ASCVD who were followed for adjudicated ASCVD. Ten-year ASCVD risk was calculated in the training set using our benchmark, the PCE, and a longitudinal DL model, Dynamic-DeepHit. Predictors included those incorporated in the PCE: sex, race, age, total cholesterol, high density lipid cholesterol, systolic and diastolic blood pressure, diabetes, hypertension treatment and smoking. The discrimination and calibration performance of the two models were evaluated in an overall hold-out testing dataset. Of the 15,565 participants in our dataset, 2170 (13.9%) developed ASCVD. The performance of the longitudinal DL model that incorporated 8 years of longitudinal risk factor data improved upon that of the PCE [AUROC: 0.815 (CI 0.782-0.844) vs 0.792 (CI 0.760-0.825)] and the net reclassification index was 0.385. The brier score for the DL model was 0.0514 compared with 0.0542 in the PCE. Incorporating longitudinal risk factors in ASCVD risk prediction using DL can improve model discrimination and calibration.

Incorporating longitudinal history of risk factors into atherosclerotic cardiovascular disease risk prediction using deep learning.

Background: It is increasingly clear that longitudinal risk factor levels and trajectories are related to risk for atherosclerotic cardiovascular disease (ASCVD) above and beyond single measures. Currently used in clinical care, the Pooled Cohort Equations (PCE) are based on regression methods that predict ASCVD risk based on cross-sectional risk factor levels. Deep learning (DL) models have been developed to incorporate longitudinal data for risk prediction but its benefit for ASCVD risk prediction relative to the traditional Pooled Cohort Equations (PCE) remain unknown. Objective: To develop a ASCVD risk prediction model that incorporates longitudinal risk factors using deep learning. Methods: Our study included 15,565 participants from four cardiovascular disease cohorts free of baseline ASCVD who were followed for adjudicated ASCVD. Ten-year ASCVD risk was calculated in the training set using our benchmark, the PCE, and a longitudinal DL model, Dynamic-DeepHit. Predictors included those incorporated in the PCE: sex, race, age, total cholesterol, high density lipid cholesterol, systolic and diastolic blood pressure, diabetes, hypertension treatment and smoking. The discrimination and calibration performance of the two models were evaluated in an overall hold-out testing dataset. Results: Of the 15,565 participants in our dataset, 2,170 (13.9%) developed ASCVD. The performance of the longitudinal DL model that incorporated 8 years of longitudinal risk factor data improved upon that of the PCE [AUROC: 0.815 (CI: 0.782-0.844) vs 0.792 (CI: 0.760-0.825)] and the net reclassification index was 0.385. The brier score for the DL model was 0.0514 compared with 0.0542 in the PCE. Conclusion: Incorporating longitudinal risk factors in ASCVD risk prediction using DL can improve model discrimination and calibration.

Social and Psychosocial Determinants of Racial and Ethnic Differences in Cardiovascular Health in the United States Population.

BACKGROUND: Social and psychosocial factors are associated with cardiovascular health (CVH). Our objective was to examine the contributions of individual-level social and psychosocial factors to racial and ethnic differences in population CVH in the NHANES (National Health and Nutrition Examination Surveys) 2011 to 2018, to inform strategies to mitigate CVH inequities. METHODS: In NHANES participants ages ≥20 years, Kitagawa-Blinder-Oaxaca decomposition estimated the statistical contribution of individual-level factors (education, income, food security, marital status, health insurance, place of birth, depression) to racial and ethnic differences in population mean CVH score (range, 0-14, accounting for diet, smoking, physical activity, body mass index, blood pressure, cholesterol, blood glucose) among Hispanic, non-Hispanic Asian, or non-Hispanic Black adults compared with non-Hispanic White adults. RESULTS: Among 16 172 participants (representing 255 million US adults), 24% were Hispanic, 12% non-Hispanic Asian, 23% non-Hispanic Black, and 41% non-Hispanic White. Among men, mean (SE) CVH score was 7.45 (2.3) in Hispanic, 8.71 (2.2) in non-Hispanic Asian, 7.48 (2.4) in non-Hispanic Black, and 7.58 (2.3) in non-Hispanic White adults. In Kitagawa-Blinder-Oaxaca decomposition, education explained the largest component of CVH differences among men (if distribution of education were similar to non-Hispanic White participants, CVH score would be 0.36 [0.04] points higher in Hispanic, 0.24 [0.04] points lower in non-Hispanic Asian, and 0.23 [0.03] points higher in non-Hispanic Black participants; P<0.05). Among women, mean (SE) CVH score was 8.03 (2.4) in Hispanic, 9.34 (2.1) in non-Hispanic Asian, 7.43 (2.3) in non-Hispanic Black, and 8.00 (2.5) in non-Hispanic White adults. Education explained the largest component of CVH difference in non-Hispanic Black women (if distribution of education were similar to non-Hispanic White participants, CVH score would be 0.17 [0.03] points higher in non-Hispanic Black participants; P<0.05). Place of birth (born in the United States versus born outside the United States) explained the largest component of CVH difference in Hispanic and non-Hispanic Asian women (if distribution of place of birth were similar to non-Hispanic White participants, CVH score would be 0.36 [0.07] points lower and 0.49 [0.16] points lower, respectively; P<0.05). CONCLUSIONS: Education and place of birth confer the largest statistical contributions to the racial and ethnic differences in mean CVH score among US adults.

Analysis of apoB Concentrations Across Early Adulthood and Predictors for Rates of Change Using CARDIA Study Data.

The cumulative exposure to apolipoprotein B (apoB)-containing lipoproteins in the blood during early adult life is a central determinant of atherosclerotic cardiovascular disease risk. To date, the patterns and rates of change in apoB through early adult life have not been described. Here, we used NMR to measure apoB concentrations in up to 3055 Coronary Artery Risk Development in Young Adults (CARDIA) Study participants who attended the years 2 (Y2), 7 (Y7), 15 (Y15), 20 (Y20), and 30 (Y30) exams. We examined individual-level spaghetti plots of apoB change, and we calculated average annualized rate of apoB concentration change during follow-up. We used multivariable linear regression models to assess the associations between CARDIA participant characteristics and annualized rates of apoB change. Male sex, higher measures of adiposity, lower HDL-C, lower Healthy Eating Index, and higher blood pressures were observed more commonly in individuals with higher apoB level at Y2 and Y20. Inter- and intra-individual variation in apoB concentration over time was substantial-while the mean (SD) rate of change was 0.52 (1.0) mg/dl/year, the range of annualized rates of change was -6.26 to +9.21 mg/dl/year. At baseline, lower first apoB measurement, female sex, White race, lower BMI, and current tobacco use were associated with apoB increase. We conclude that the significant variance in apoB level over time and the modest association between baseline measures and rates of apoB change suggest that the ability to predict an individual's future apoB serum concentrations, and thus their cumulative apoB exposure, after a one-time assessment in young adulthood is low.

Status of Cardiovascular Health in US Adults and Children Using the American Heart Association's New "Life's Essential 8" Metrics: Prevalence Estimates From the National Health and Nutrition Examination Survey (NHANES), 2013 Through 2018.

BACKGROUND: The American Heart Association recently published an updated algorithm for quantifying cardiovascular health (CVH)-the Life's Essential 8 score. We quantified US levels of CVH using the new score. METHODS: We included individuals ages 2 through 79 years (not pregnant or institutionalized) who were free of cardiovascular disease from the National Health and Nutrition Examination Surveys in 2013 through 2018. For all participants, we calculated the overall CVH score (range, 0 [lowest] to 100 [highest]), as well as the score for each component of diet, physical activity, nicotine exposure, sleep duration, body mass index, blood lipids, blood glucose, and blood pressure, using published American Heart Association definitions. Sample weights and design were incorporated in calculating prevalence estimates and standard errors using standard survey procedures. CVH scores were assessed across strata of age, sex, race and ethnicity, family income, and depression. RESULTS: There were 23 409 participants, representing 201 728 000 adults and 74 435 000 children. The overall mean CVH score was 64.7 (95% CI, 63.9-65.6) among adults using all 8 metrics and 65.5 (95% CI, 64.4-66.6) for the 3 metrics available (diet, physical activity, and body mass index) among children and adolescents ages 2 through 19 years. For adults, there were significant differences in mean overall CVH scores by sex (women, 67.0; men, 62.5), age (range of mean values, 62.2-68.7), and racial and ethnic group (range, 59.7-68.5). Mean scores were lowest for diet, physical activity, and body mass index metrics. There were large differences in mean scores across demographic groups for diet (range, 23.8-47.7), nicotine exposure (range, 63.1-85.0), blood glucose (range, 65.7-88.1), and blood pressure (range, 49.5-84.0). In children, diet scores were low (mean 40.6) and were progressively lower in higher age groups (from 61.1 at ages 2 through 5 to 28.5 at ages 12 through 19); large differences were also noted in mean physical activity (range, 63.1-88.3) and body mass index (range, 74.4-89.4) scores by sociodemographic group. CONCLUSIONS: The new Life's Essential 8 score helps identify large group and individual differences in CVH. Overall CVH in the US population remains well below optimal levels and there are both broad and targeted opportunities to monitor, preserve, and improve CVH across the life course in individuals and the population.

Development and Validation of a Long-Term Incident Heart Failure Risk Model.

BACKGROUND: Average lifetime risk for heart failure (HF) is high but differs significantly across and within sex-race groups. No models for estimating long-term risk for HF exist, which would allow for earlier identification and interventions in high-risk subsets. The authors aim to derive 30-year HF risk equations. METHODS: Adults between the ages of 20 to 59 years and free of cardiovascular disease at baseline from 5 population-based cohorts were included. Among 24 838 participants (55% women, 25% Black based on self-report), follow-up consisted of 599 551 person-years. Sex- and race-specific 30-year HF risk equations were derived and validated accounting for competing risk of non-HF death. HF was based on a clinical diagnosis. Model discrimination and calibration were assessed using 10-fold cross-validation. Finally, the model was applied to varying risk factor patterns for systematic examination. RESULTS: The rate of incident HF was 4.0 per 1000 person-years. Harrell C statistics were 0.82 (0.80-0.83) and 0.84 (0.82-0.85) in White and Black men and 0.84 (0.82-0.85) and 0.85 (0.83-0.87) in White and Black women, respectively. Hosmer-Lemeshow calibration was acceptable, with χ2 <30 in all subgroups. Risk estimation varied across sex-race groups: for example, in an average 40-year-old nonsmoker with an untreated systolic blood pressure of 140 mm Hg and body mass index of 30 kg/m2, risk was estimated to be 22.8% in a Black man, 13.7% in a White man, 13.0% in a Black woman, and 12.1% in a White woman. CONCLUSIONS: Sex- and race-specific equations for prediction of long-term risk of HF demonstrated high discrimination and adequate calibration.

Cigarette Smoking and Competing Risks for Fatal and Nonfatal Cardiovascular Disease Subtypes Across the Life Course.

Background Cigarette smoking is significantly associated with premature death related and not related to cardiovascular disease (CVD). Whether risk associated with smoking is similar across CVD subtypes and how this translates into years of life lost is not known. Methods and Results We pooled and harmonized individual-level data from 9 population-based cohorts in the United States. All participants were free of clinical CVD at baseline with available data on current smoking status, covariates, and CVD outcomes. We examined the association between smoking status and total CVD and CVD subtypes, including fatal and nonfatal coronary heart disease, stroke, congestive heart failure, and other CVD deaths. We performed (1) modified Kaplan-Meier analysis to estimate long-term risks, (2) adjusted competing Cox models to estimate joint cumulative risks for CVD or noncardiovascular death, and (3) Irwin's restricted mean to estimate years lived free from and with CVD. Of 106 165 adults, 50.4% were women. Overall long-term risks for CVD events were 46.0% (95% CI, 44.7-47.3) and 34.7% (95% CI, 33.3-36.0) in middle-aged men and women, respectively. In middle-aged men who reported smoking compared with those who did not smoke, competing hazard ratios (HRs) were higher for the first presentation being a fatal CVD event (HR, 1.79 [95% CI, 1.68-1.92]), with a similar pattern among women (HR,1.82 [95% CI, 1.68-1.98]). Smoking was associated with earlier CVD onset by 5.1 and 3.8 years in men and women. Similar patterns were observed in younger and older adults. Conclusions Current smoking was associated with a fatal event as the first manifestation of clinical CVD.

Race- and Sex-Specific Population Attributable Fractions of Incident Heart Failure: A Population-Based Cohort Study From the Lifetime Risk Pooling Project.

BACKGROUND: Race- and sex-specific differences in heart failure (HF) risk may be related to differential burden and effect of risk factors. We estimated the population attributable fraction (PAF), which incorporates both prevalence and excess risk of HF associated with each risk factor (obesity, hypertension, diabetes, current smoking, and hyperlipidemia), in specific race-sex groups. METHODS: A pooled cohort was created using harmonized data from 6 US longitudinal population-based cohorts. Baseline measurements of risk factors were used to determine prevalence. Relative risk of incident HF was assessed using a piecewise constant hazards model adjusted for age, education, other modifiable risk factors, and the competing risk of death from non-HF causes. Within each race-sex group, PAF of HF was estimated for each risk factor individually and for all risk factors simultaneously. RESULTS: Of 38 028 participants, 55% were female and 22% Black. Hypertension had the highest PAF among Black men (28.3% [95% CI, 18.7%-36.7%]) and women (25.8% [95% CI, 16.3%-34.2%]). In contrast, PAF associated with obesity was the highest in White men (21.0% [95% CI, 14.6%-27.0%]) and women (17.9% [95% CI, 12.8%-22.6%]). Diabetes disproportionately contributed to HF in Black women (PAF, 16.4% [95% CI, 12.7%-19.9%]). The cumulative PAF of all 5 risk factors was the highest in Black women (51.9% [95% CI, 39.3%-61.8%]). CONCLUSIONS: The observed differences in contribution of risk factors across race-sex groups can inform tailored prevention strategies to mitigate disparities in HF burden. This novel competing risk analysis suggests that a sizeable proportion of HF risk may not be associated with modifiable risk factors.

Association between marijuana use and electrocardiographic abnormalities by middle age: the Coronary Artery Risk Development in Young Adults (CARDIA) study.

AIMS: To evaluate the prevalence of electrocardiogram (ECG) abnormalities in marijuana users as an indirect measure of subclinical cardiovascular disease (CVD). DESIGN: Longitudinal and cross-sectional secondary data analysis from the CARDIA (Coronary Artery Risk Development in Young Adults) study. SETTING: Four communities in the United States. PARTICIPANTS: A total of 2585 participants from the 5115 black and white men and women recruited at age 18-30 years in 1985 to 1986 in CARDIA. MEASUREMENTS: ECG abnormalities coded as minor and major abnormalities with the Minnesota code of electrocardiographic findings at year 20. Self-reported current (past 30 days) and computed cumulative life-time marijuana use (one 'marijuana-year' corresponds to 365 days of use) through assessments every 2-5 years. We fitted logistic regression models adjusting for sex, race, center, education, age, tobacco smoking, physical activity, alcohol use and body mass index. FINDINGS: Among the 2585 participants with an ECG at year 20, mean age was 46, 57% were women, 45% were black; 83% had past exposure to marijuana and 11% were using marijuana currently. One hundred and seventy-three participants (7%) had major abnormalities and 944 (37%) had minor abnormalities. Comparing current with never use in multivariable-adjusted models, the odds ratio (OR) for major ECG abnormalities was 0.60 [95% confidence interval (CI) = 0.32-1.15] and for minor ECG abnormalities 1.21 (95% CI = 0.87-1.68). Results did not change after stratifying by sex and race. Cumulative marijuana use was not associated with ECG abnormalities. CONCLUSION: In a middle-aged US population, life-time cumulative and occasional current marijuana use were not associated with increases in electrocardiogram abnormalities. This adds to the growing body of evidence that occasional marijuana use and cardiovascular disease events and markers of subclinical atherosclerosis are not associated.

Cardiovascular Health Among Pregnant Women, Aged 20 to 44 Years, in the United States.

Background Pregnancy is a cardiometabolic stressor and thus a critical period to address women's lifetime cardiovascular health (CVH). However, CVH among US pregnant women has not been characterized. Methods and Results We analyzed cross-sectional data from National Health and Nutrition Examination Surveys 1999 to 2014 for 1117 pregnant and 8200 nonpregnant women, aged 20 to 44 years. We assessed 7 CVH metrics using American Heart Association definitions modified for pregnancy; categorized metrics as ideal, intermediate, or poor; assigned these categories 2, 1, or 0 points, respectively; and summed across the 7 metrics for a total score of 0 to 14 points. Total scores 12 to 14 indicated high CVH; 8 to 11, moderate CVH; and 0 to 7, low CVH. We applied survey weights to generate US population-level estimates of CVH levels and compared pregnant and nonpregnant women using demographic-adjusted polytomous logistic and linear regression. Among pregnant women, the prevalences (95% CIs) of ideal levels of CVH metrics were 0.1% (0%-0.3%) for diet, 27.3% (22.2%-32.3%) for physical activity, 38.9% (33.7%-44.0%) for total cholesterol, 51.1% (46.0%-56.2%) for body mass index, 77.7% (73.3%-82.2%) for smoking, 90.4% (87.5%-93.3%) for blood pressure, and 91.6% (88.3%-94.9%) for fasting glucose. The mean total CVH score was 8.3 (95% CI, 8.0-8.7) of 14, with high CVH in 4.6% (95% CI, 0.5%-8.8%), moderate CVH in 60.6% (95% CI, 52.3%-68.9%), and low CVH in 34.8% (95% CI, 26.4%-43.2%). CVH levels were significantly lower among pregnant versus nonpregnant women; for example, 13.0% (95% CI, 11.0%-15.0%) of nonpregnant women had high CVH (adjusted, comparison P=0.01). Conclusions From 1999 to 2014, <1 in 10 US pregnant women, aged 20 to 44 years, had high CVH.

Lifetime Risk of Lower-Extremity Peripheral Artery Disease Defined by Ankle-Brachial Index in the United States.

Background There are no available lifetime risk estimates of lower-extremity peripheral artery disease (PAD). Methods and Results Using data from 6 US community-based cohorts and the vital statistics, we estimated the prevalence and incidence of PAD, defined as an ankle-brachial index < 0.90, at each year of age from birth to 80 years for white, black, and Hispanic men and women. Then, we used Markov Monte Carlo simulations in a simulated cohort of 100 000 individuals to estimate lifetime risk of PAD. On the basis of odds ratios of PAD for traditional atherosclerotic risk factors (eg, diabetes mellitus and smoking), we developed a calculator providing residual lifetime risk of PAD. In an 80-year horizon, lifetime risks of PAD were 30.0% in black men and 27.6% in black women, but ≈19% in white men and women and ≈22% in Hispanic men and women. From another perspective, 9% of blacks were estimated to develop PAD by 60 years of age, while the same proportion was seen at ≈70 years for whites and Hispanics. The residual lifetime risk within the same race/ethnicity varied by 3.5- to 5-fold according to risk factors (eg, residual lifetime risk in 45-year-old black men was 19.9% when current smoking, diabetes mellitus, and history of cardiovascular disease were absent versus 70.4% when all were present). Conclusions In the United States, ≈30% of blacks are estimated to develop PAD during their lifetime, whereas the corresponding estimate is ≈20% for whites and Hispanics. The residual lifetime risk within the same race/ethnicity substantially varies according to traditional risk factors.

10-Year Risk Equations for Incident Heart Failure in the General Population.

BACKGROUND: Primary prevention strategies to mitigate the burden of heart failure (HF) are urgently needed. However, no validated risk prediction tools are currently in use. OBJECTIVES: This study sought to derive 10-year risk equations of developing incident HF. METHODS: Race- and sex-specific 10-year risk equations for HF were derived and validated from individual-level data from 7 community-based cohorts with at least 12 years of follow-up. Participants who were recruited between 1985 and 2000, between 30 to 79 years, and were free of cardiovascular disease at baseline were included to create a pooled cohort (PC) and were randomly split for derivation and internal validation. Model performance was also assessed in 2 additional cohorts. RESULTS: In the derivation sample of the PC (n = 11,771), 58% were women, 22% were black with a mean age of 52 ± 12 years, and HF occurred in 1,339 participants. Predictors of HF included in the race-sex-specific models were age, blood pressure (treated or untreated), fasting glucose (treated or untreated), body mass index, cholesterol, smoking status, and QRS duration. The PC equations to Prevent HF model had good discrimination and strong calibration in internal and external validation cohorts. A web-based tool was developed to facilitate clinical application of this tool. CONCLUSIONS: The authors present a contemporary analysis from 33,010 men and women demonstrating the utility of the sex- and race-specific 10-year PC equations to Prevent HF risk score, which integrates clinical parameters readily available in primary care settings. This tool can be useful in risk-based decision making to determine who may merit intensive screening and/or targeted prevention strategies.

A Simple Community-Based Risk-Prediction Score for Sudden Cardiac Death.

BACKGROUND: Although sudden cardiac death is a leading cause of death in the United States, most victims of sudden cardiac death are not identified as at risk prior to death. We sought to derive and validate a population-based risk score that predicts sudden cardiac death. METHODS: The Atherosclerosis Risk in Communities (ARIC) Study recorded clinical measures from men and women aged 45-64 years at baseline; 11,335 white and 3780 black participants were included in this analysis. Participants were followed over 10 years and sudden cardiac death was physician adjudicated. Cox proportional hazards models were used to derive race-specific equations to estimate the 10-year sudden cardiac death risk. Covariates for the risk score were selected from available demographic and clinical variables. Utility was assessed by calculating discrimination (Harrell's C-index) and calibration (Hosmer-Lemeshow chi-squared test). The white-specific equation was validated among 5626 Framingham Heart Study participants. RESULTS: During 10 years' follow-up among ARIC participants (mean age 54.4 years, 52.4% women), 145 participants experienced sudden cardiac death; the majority occurred in the highest quintile of predicted risk. Model covariates included age, sex, total cholesterol, lipid-lowering and hypertension medication use, blood pressure, smoking status, diabetes, and body mass index. The score yielded very good internal discrimination (white-specific C-index 0.82; 95% confidence interval [CI], 0.78-0.85; black-specific C-index 0.75; 95% CI, 0.68-0.82) and very good external discrimination among Framingham participants (C-index 0.82; 95% CI, 0.79-0.86). Calibration plots indicated excellent calibration in ARIC (white-specific chi-squared 5.3, P = .82; black-specific chi-squared 4.1, P = .77), and a simple recalibration led to excellent fit within Framingham (chi-squared 2.1, P = 0.99). CONCLUSIONS: The proposed risk scores may be used to identify those at risk for sudden cardiac death within 10 years and particularly classify those at highest risk who may merit further screening.

Trends in Racial/Ethnic Disparities in Cardiovascular Health Among US Adults From 1999-2012.

BACKGROUND: In the United States, there are persistent racial and ethnic disparities in cardiovascular disease morbidity and mortality. National efforts have focused on reducing these disparities; however, little is known about the long-term trends in racial/ethnic disparities in cardiovascular health (CVH). METHODS AND RESULTS: We included 11 285 adults aged ≥20 years from the National Health and Nutrition Examination Surveys survey cycles 1999/2000 through 2011/2012. CVH includes 7 health factors and behaviors-diet, physical activity, smoking status, body mass index, blood pressure, blood glucose, and total cholesterol-each scored as ideal (2 points), intermediate (1 point), or poor (0 points). Overall CVH is a summation of these scores (range, 0-14) points. Age-adjusted mean CVH scores were calculated by race/ethnicity (non-Hispanic black, non-Hispanic white, or Mexican American) and sex for each survey cycle. Non-Hispanic black women had significantly lower mean CVH scores as compared with non-Hispanic white women at each survey cycle (difference=0.93; P=0.001 in 2011/2012) and Mexican-American women had significantly lower mean score as compared with non-Hispanic white women at almost all survey cycles (difference=0.71; P=0.02 in 2011/2012). Differences between racial/ethnic groups were smaller for men and were mostly nonsignificant. CONCLUSIONS: From 1999/2000 to 2011/2012, there were enduring disparities in CVH for non-Hispanic black and Mexican-American women as compared with non-Hispanic white women. Disparities that were present in 1999/2000 were present in 2011/2012, though no racial/ethnic differences became more pronounced over time. These findings provide US nationally representative data to evaluate health factors and behaviors of particular concern regarding racial/ethnic disparities in cardiovascular health.

A point-based prediction model for cardiovascular risk in orthotopic liver transplantation: The CAR-OLT score.

Cardiovascular disease (CVD) complications are important causes of morbidity and mortality after orthotopic liver transplantation (OLT). There is currently no preoperative risk-assessment tool that allows physicians to estimate the risk for CVD events following OLT. We sought to develop a point-based prediction model (risk score) for CVD complications after OLT, the Cardiovascular Risk in Orthotopic Liver Transplantation risk score, among a cohort of 1,024 consecutive patients aged 18-75 years who underwent first OLT in a tertiary-care teaching hospital (2002-2011). The main outcome measures were major 1-year CVD complications, defined as death from a CVD cause or hospitalization for a major CVD event (myocardial infarction, revascularization, heart failure, atrial fibrillation, cardiac arrest, pulmonary embolism, and/or stroke). The bootstrap method yielded bias-corrected 95% confidence intervals for the regression coefficients of the final model. Among 1,024 first OLT recipients, major CVD complications occurred in 329 (32.1%). Variables selected for inclusion in the model (using model optimization strategies) included preoperative recipient age, sex, race, employment status, education status, history of hepatocellular carcinoma, diabetes, heart failure, atrial fibrillation, pulmonary or systemic hypertension, and respiratory failure. The discriminative performance of the point-based score (C statistic = 0.78, bias-corrected C statistic = 0.77) was superior to other published risk models for postoperative CVD morbidity and mortality, and it had appropriate calibration (Hosmer-Lemeshow P = 0.33). CONCLUSION: The point-based risk score can identify patients at risk for CVD complications after OLT surgery (available at www.carolt.us); this score may be useful for identification of candidates for further risk stratification or other management strategies to improve CVD outcomes after OLT. (Hepatology 2017;66:1968-1979).

Application of a Lifestyle-Based Tool to Estimate Premature Cardiovascular Disease Events in Young Adults: The Coronary Artery Risk Development in Young Adults (CARDIA) Study.

Importance: Few tools exist for assessing the risk for early atherosclerotic cardiovascular disease (ASCVD) events in young adults. Objective: To assess the performance of the Healthy Heart Score (HHS), a lifestyle-based tool that estimates ASCVD events in older adults, for ASCVD events occurring before 55 years of age. Design, Setting, and Participants: This prospective cohort study included 4893 US adults aged 18 to 30 years from the Coronary Artery Risk Development in Young Adults (CARDIA) study. Participants underwent measurement of lifestyle factors from March 25, 1985, through June 7, 1986, and were followed up for a median of 27.1 years (interquartile range, 26.9-27.2 years). Data for this study were analyzed from February 24 through December 12, 2016. Exposures: The HHS includes age, smoking status, body mass index, alcohol intake, exercise, and a diet score composed of self-reported daily intake of cereal fiber, fruits and/or vegetables, nuts, sugar-sweetened beverages, and red and/or processed meats. The HHS in the CARDIA study was calculated using sex-specific equations produced by its derivation cohorts. Main Outcomes and Measures: The ability of the HHS to assess the 25-year risk for ASCVD (death from coronary heart disease, nonfatal myocardial infarction, and fatal or nonfatal ischemic stroke) in the total sample, in race- and sex-specific subgroups, and in those with and without clinical ASCVD risk factors at baseline. Model discrimination was assessed with the Harrell C statistic; model calibration, with Greenwood-Nam-D'Agostino statistics. Results: The study population of 4893 participants included 2205 men (45.1%) and 2688 women (54.9%) with a mean (SD) age at baseline of 24.8 (3.6) years; 2483 (50.7%) were black; and 427 (8.7%) had at least 1 clinical ASCVD risk factor (hypertension, hyperlipidemia, or diabetes types 1 and 2). Among these participants, 64 premature ASCVD events occurred in women and 99 in men. The HHS showed moderate discrimination for ASCVD risk assessment in this diverse population of mostly healthy young adults (C statistic, 0.71; 95% CI, 0.66-0.76); it performed better in men (C statistic, 0.74; 95% CI, 0.68-0.79) than in women (C statistic, 0.69; 95% CI, 0.62-0.75); in white (C statistic, 0.77; 95% CI, 0.71-0.84) than in black (C statistic, 0.66; 95% CI, 0.60-0.72) participants; and in those without (C statistic, 0.71; 95% CI, 0.66-0.76) vs with (C statistic, 0.64; 95% CI, 0.55-0.73) clinical risk factors at baseline. The HHS was adequately calibrated overall and within each subgroup. Conclusions and Relevance: The HHS, when measured in younger persons without ASCVD risk factors, performs moderately well in assessing risk for ASCVD events by early middle age. Its reliance on self-reported, modifiable lifestyle factors makes it an attractive tool for risk assessment and counseling for early ASCVD prevention.

Association of Cardiovascular Health With Subclinical Disease and Incident Events: The Multi-Ethnic Study of Atherosclerosis.

BACKGROUND: Few adults have ideal cardiovascular health (CVH). We studied associations of an overall CVH score with subclinical cardiovascular disease and events. We assessed whether associations varied by race/ethnicity. METHODS AND RESULTS: Among 5961 participants in the Multi-Ethnic Study of Atherosclerosis, components of CVH were measured at baseline, 2000-2002: systolic blood pressure, total cholesterol, fasting glucose, smoking, physical activity, diet, and body mass index. Levels were classified as ideal (2 points), intermediate (1 point), and poor (0 points) according to American Heart Association definitions. Points were summed to produce a CVH score (0-7 low, 8-11 moderate, 12-14 high). Coronary artery calcium, carotid intima-media thickness, and left ventricular mass were measured at baseline. Cardiovascular disease was defined as myocardial infarction, coronary heart disease death, resuscitated cardiac arrest, stroke, heart failure, or peripheral artery disease. Follow-up was 10.3 years. Regression models were used to examine associations of the CVH score with subclinical disease and events, adjusting for age, sex, and education. Analyses were stratified by race/ethnicity. Adults with high or moderate CVH scores had significantly lower odds of coronary artery calcium and lower carotid intima-media thickness and left ventricular mass than adults with low CVH scores. Adults with high or moderate CVH scores were 67% (95%CI 41% to 82%) and 37% (95%CI 22% to 49%) less likely, respectively, to experience a cardiovascular disease event than adults with low scores. There was no interaction with race/ethnicity. CONCLUSIONS: There is a graded inverse association between CVH scores and measures of subclinical and overt cardiovascular disease that is similar across race/ethnic groups.

Assessing and Refining Myocardial Infarction Risk Estimation Among Patients With Human Immunodeficiency Virus: A Study by the Centers for AIDS Research Network of Integrated Clinical Systems.

Importance: Persons with human immunodeficiency virus (HIV) that is treated with antiretroviral therapy have improved longevity but face an elevated risk of myocardial infarction (MI) due to common MI risk factors and HIV-specific factors. Despite these elevated MI rates, optimal methods to predict MI risks for HIV-infected persons remain unclear. Objective: To determine the extent to which existing and de novo estimation tools predict MI in a multicenter HIV cohort with rigorous MI adjudication. Design, Setting, and Participants: We evaluated the performance of standard of care and 2 new data-derived MI risk estimation models in 5 Centers for AIDS Research Network of Integrated Clinical Systems sites across the United States where a multicenter clinical prospective cohort of 19 829 HIV-infected adults received care in inpatient and outpatient settings since 1995. The new risk estimation models were validated in a separate cohort from the derivation cohort. Exposures: Traditional cardiovascular risk factors, HIV viral load, CD4 lymphocyte count, statin use, antihypertensive use, and antiretroviral medication use were used to calculate predicted event rates. Main Outcomes and Measures: We observed MI rates over the course of follow-up that were scaled to 10 years using the Greenwood-Nam-D'Agostino Kaplan-Meier approach to account for dropout and loss to follow-up before 10 years. Results: Of the 11 288 patients with complete baseline data, 6904 were white and 9250 were men. Myocardial infarction rates were higher among black men (6.9 per 1000 person-years) and black women (7.2 per 1000 person-years) than white men (4.4 per 1000 person-years) and white women (3.3 per 1000 person-years), older participants (7.5 vs 2.2 MI per 1000 person-years for adults 40 years and older vs < 40 years old at study entry, respectively), and participants who were not virally suppressed (6.3 vs 4.7 per 1000 person-years for participants with and without detectable viral load, respectively). The 2013 Pooled Cohort Equations, which predict composite rates of MI and stroke, adequately discriminated MI risk (Harrell C statistic = 0.75; 95% CI, 0.71-0.78). Two data-derived models incorporating HIV-specific covariates exhibited weak calibration in a validation sample and did not discriminate risk any better (Harrell C statistic = 0.72; 95% CI, 0.67-0.78 and 0.73; 95% CI, 0.68-0.79) than the Pooled Cohort Equations. The Pooled Cohort Equations were moderately calibrated in the Centers for AIDS Research Network of Clinical Systems but predicted consistently lower MI rates. Conclusions and Relevance: The Pooled Cohort Equations discriminated MI risk and were moderately calibrated in this multicenter HIV cohort. Adding HIV-specific factors did not improve model performance. As HIV-infected cohorts capture and assess MI and stroke outcomes, researchers should revisit the performance of risk estimation tools.

Prevalence and Prognostic Implications of Coronary Artery Calcification in Low-Risk Women: A Meta-analysis.

Importance: The role of coronary artery calcium (CAC) testing for guiding preventive strategies among women at low cardiovascular disease (CVD) risk based on the American College of Cardiology and American Heart Association CVD prevention guidelines is unclear. Objective: To assess the potential utility of CAC testing for CVD risk estimation and stratification among low-risk women. Design, Setting, and Participants: Women with 10-year atherosclerotic CVD (ASCVD) risk lower than 7.5% from 5 large population-based cohorts: the Dallas Heart Study (United States), the Framingham Heart Study (United States), the Heinz Nixdorf Recall study (Germany), the Multi-Ethnic Study of Atherosclerosis (United States), and the Rotterdam Study (the Netherlands). The 5 cohorts were selected based on the availability of CAC data in a sizable group of low-risk women from the general population together with the long detailed follow-up data. Across the cohorts, events were assessed from the date of CAC scan (performed from 1998 through 2006) until January 1, 2012; January 1, 2014; or March 6, 2015. Fixed-effects meta-analysis was conducted to combine the results of the 5 studies. Exposures: CAC score by computed tomography. Main Outcomes and Measures: Main outcome was incident ASCVD, including nonfatal myocardial infarction, coronary heart disease (CHD) death, and stroke. Association of CAC with ASCVD was examined using Cox proportional hazards models. To assess whether CAC was associated with improved ASCVD risk predictions beyond the traditional risk factors, the C statistic and the continuous net reclassification improvement (cNRI) index were calculated. Results: Among 6739 women with low ASCVD risk from the 5 studies, mean age ranged from 44 to 63 years and CAC was present in 36.1%. Across the cohorts, median follow-up ranged from 7.0 to 11.6 years. A total of 165 ASCVD events occurred (64 nonfatal myocardial infarctions, 29 CHD deaths, and 72 strokes), with the ASCVD incidence rates ranging from 1.5 to 6.0 per 1000 person-years. Compared with the absence of CAC (CAC = 0), presence of CAC (CAC >0) was associated with an increased risk of ASCVD (incidence rates per 1000 person-years, 1.41 for CAC absence vs 4.33 for CAC presence; difference, 2.92 [95% CI, 2.02-3.83]; multivariable-adjusted hazard ratio, 2.04 [95% CI, 1.44-2.90]). The addition of CAC to traditional risk factors improved the C statistic from 0.73 (95% CI, 0.69-0.77) to 0.77 (95% CI, 0.74-0.81) and provided a cNRI of 0.20 (95% CI, 0.09-0.31) for ASCVD prediction. Conclusions and Relevance: Among women at low ASCVD risk, CAC was present in approximately one-third and was associated with an increased risk of ASCVD and modest improvement in prognostic accuracy compared with traditional risk factors. Further research is needed to assess the clinical utility and cost-effectiveness of this additional accuracy.

Long-Term Risk of Atherosclerotic Cardiovascular Disease in US Adults With the Familial Hypercholesterolemia Phenotype.

BACKGROUND: Heterozygous familial hypercholesterolemia (FH) affects up to 1 in 200 individuals in the United States, but atherosclerotic cardiovascular disease (ASCVD) outcomes of FH in the general US population have not been described. We therefore sought to evaluate long-term coronary heart disease (CHD) and total ASCVD risks in US adults with an FH phenotype. METHODS: Using individual pooled data from 6 large US epidemiological cohorts, we stratified participants by low-density lipoprotein cholesterol level at index ages from 20 to 79 years. For the primary analysis, low-density lipoprotein cholesterol levels ≥190 and <130 mg/dL defined the FH phenotype and referent, respectively. Sensitivity analyses evaluated the effects of varying the FH phenotype definition. We used Cox regression models to assess covariate-adjusted associations of the FH phenotype with 30-year hazards for CHD (CHD death or nonfatal myocardial infarction) and total ASCVD (CHD or stroke). RESULTS: We included 68 565 baseline person-examinations; 3850 (5.6%) had the FH phenotype by the primary definition. Follow-up across index ages ranged from 78 985 to 308 378 person-years. After covariate adjustment, the FH phenotype was associated with substantially elevated 30-year CHD risk, with hazard ratios up to 5.0 (95% confidence interval, 1.1-21.7). Across index ages, CHD risk was accelerated in those with the FH phenotype by 10 to 20 years in men and 20 to 30 years in women. Similar patterns of results were found for total ASCVD risk, with hazard ratios up to 4.1 (95% confidence interval, 1.2-13.4). Alternative FH phenotype definitions incorporating family history, more stringent age-based low-density lipoprotein cholesterol thresholds, or alternative lipid fractions decreased the FH phenotype prevalence to as low as 0.2% to 0.4% without materially affecting CHD risk estimates (hazard ratios up to 8.0; 95% confidence interval, 1.0-61.6). CONCLUSIONS: In the general US population, the long-term ASCVD burden related to phenotypic FH, defined by low-density lipoprotein cholesterol ≥190 mg/dL, is likely substantial. Our finding of CHD risk acceleration may aid efforts in risk communication.