Associations of Serum Nonesterified Fatty Acids With Coronary Heart Disease Mortality and Nonfatal Myocardial Infarction: The CHS (Cardiovascular Health Study) Cohort

Abstract

Background

Significant associations have been reported between serum total nonesterified fatty acid (NEFA) concentrations and coronary heart disease (CHD) mortality and incident nonfatal myocardial infarction (MI) in some prospective cohort studies. Little is known about whether individual or subclasses (saturated, polyunsaturated [n‐6 and n‐3], and trans fatty acids) of serum NEFAs relate to CHD mortality and nonfatal MI.

Methods and Results

CHS (Cardiovascular Health Study) participants (N=1681) who had no history of MI, angina, or revascularization or were free of MI at baseline (1996–1997) were included. NEFAs were quantified using gas chromatography. Cox regression analysis was used to evaluate associations of 5 subclasses and individual NEFAs with CHD composite (CHD mortality and nonfatal MI), CHD mortality, and incident nonfatal MI. During a median follow‐up of 11.7 years, 266 cases of CHD death and 271 cases of nonfatal MI occurred. In the fully adjusted model, no significant associations were identified between individual NEFA and CHD composite. Exploratory analyses indicated that lauric acid (12:0) was negatively associated (hazard ratio [HR], 0.76; 95% CI, 0.59–0.98; P=0.0328) and dihomo‐γ‐linolenic acid (20:3n‐6) was positively associated with CHD mortality (HR, 1.34; 95% CI, 1.02–1.76; P=0.0351). Elaidic acid (18:1n‐7t) was positively associated with incident nonfatal MI (HR, 1.46; 95% CI, 1.01–2.12; P=0.0445). No significant associations were observed for NEFA subclass and any outcomes.

Conclusions

In CHS participants, 2 NEFAs, dihomo‐γ‐linolenic and elaidic acids, were positively associated with CHD mortality and nonfatal MI, respectively, suggesting potential susceptibility biomarkers for risks of CHD mortality and nonfatal MI.

Nonstandard Abbreviations and Acronyms

CHS

Cardiovascular Health Study

WHI

Women's Health Initiative

Clinical Perspective

What Is New?

• This is the first study to profile subclasses and individual nonesterified fatty acids and assess their relations with coronary heart disease mortality and incident nonfatal myocardial infarction in a community‐based prospective study of older adults.

• Fasting serum nonesterified dihomo‐γ‐linolenic (20:3n‐6) and elaidic acids (18:1n‐9t) were associated with a 34% higher risk of coronary heart disease mortality and 46% higher risk of nonfatal myocardial infarction, respectively.

What Are the Clinical Implications?

• The findings suggest that serum nonesterified dihomo‐γ‐linolenic (20:3n‐6) and elaidic acids (18:1n‐9t) may be potential susceptibility biomarkers for coronary heart disease mortality and nonfatal myocardial infarction risk in older adults.

Cardiovascular disease, including coronary heart disease (CHD), is the leading cause of morbidity and mortality in the United States. By 2030, the medical costs associated with CHD are estimated to double from current levels. ¹ Inflammation and disorders of lipid metabolism are linked to coronary artery plaque initiation and progression. ² , ³ , ⁴ Elevated concentrations of nonesterified fatty acids (NEFA) have been associated with increased local and systemic inflammation and induction of oxidative stress, ⁵ insulin resistance, ⁶ , ⁷ endothelial dysfunction, ⁸ and foam cell formation. ⁵ , ⁶ , ⁸

Data from cross‐sectional and prospective studies identified significant positive associations between plasma total NEFA concentrations and CHD risk factors, including hypertension, ⁸ , ⁹ obesity, ¹⁰ , ¹¹ diabetes mellitus, insulin resistance, ¹¹ , ¹² , ¹³ and systemic inflammation. ⁴ Total plasma NEFA has also been linked to adverse cardiovascular events. ¹⁴ , ¹⁵ , ¹⁶ In the CHS (Cardiovascular Health Study) cohort, plasma total NEFA concentrations were positively associated with heart failure ¹⁶ but not ischemic stroke ¹⁷ and cardiac arrest. ¹⁸ In other cohorts plasma total NEFA concentrations have been associated with severity of coronary artery disease. ¹⁹ However, lacking is an assessment of the relation between NEFA subclasses (saturated FA, omega (n)‐6 polyunsaturated [PUFA], n‐3 PUFA, and trans fatty acid [FA]) or individual NEFAs and CHD risk in a large population of older adults. Preliminary support for the importance of resolving this uncertainty comes from data for other plasma circulating lipid subfractions. ²⁰ Specifically, plasma long‐chain phospholipid (PL) saturated and trans FAs have been positively associated with incident CHD, ²⁰ , ²¹ , ²² whereas both plasma PL total n‐6 and n‐3 PUFA, as well as 3 individual plasma long‐chain PL FAs (linoleic acid, eicosapentaenoic acid, and docosahexaenoic acid), have been negatively associated with CHD risk. ²¹ , ²²

The objective of this study was to assess the relation between serum NEFA subclasses and individual NEFA(s), and a composite outcome of CHD (CHD composite, including CHD mortality and nonfatal myocardial infarction [MI]), CHD mortality and nonfatal MI in CHS participants. We hypothesized that fasting serum nonesterified n‐3 and n‐6 PUFAs, either as a subclass or individually, would be inversely associated, and nonesterified saturated FA and trans FA, either as a subclass or individually, would be positively associated with CHD composite, CHD mortality, and incident nonfatal MI.

Methods

Data Disclosure Statement

The data that support the findings of this study are available from the CHS Coordinating Center upon reasonable request.

Study Population and Design

The CHS cohort is a population‐based, longitudinal study of CHD and stroke in adults aged 65 years and older. ²³ Briefly, in 1989 to 1990, a total of 5201 Medicare‐eligible residents were recruited from 4 field centers (Allegheny County, PA; Forsyth County, NC; Sacramento County, CA; Washington County, MD). In 1992 to 1993, using similar recruitment methods, 687 mostly Black participants were recruited from the same field centers with the exception of Washington County, MD. Participants attended clinic exams at baseline and annually through 1999. Of the 4413 participants who attended the 1996 to 1997 visit, NEFA concentrations were determined in serum from 2139 participants who had a 2‐hour oral glucose tolerance test blood specimen available. Among these participants, 458 were excluded because of prevalent CHD, resulting in a final sample size of 1681 for the current analysis. The institutional review committee of each field center approved the study, and all participants provided informed written consent. Separate approval to use de‐identified samples and data for the analyses proposed in this study was obtained under exemption category 4, from the Tufts University/Tufts Medical Center Institutional Review Board.

Nonesterified Fatty Acid Determinations

All samples used for NEFA analysis were collected in 1996 to 1997 and stored at −80℃ and never thawed before the NEFA determinations in 2017. No antioxidants and triglyceride lipolysis inhibitors were used during sample collection. Prior work has demonstrated that plasma phospholipids, cholesteryl ester, and triglyceride fatty acid profiles were stable for at least 10 years when stored at −80°C. ²⁴ Lipids were extracted from serum using a modified Folch method ²⁵ , ²⁶ , ²⁷ after addition of an internal standard (heptadecanoic acid). The serum NEFA fraction was isolated using solid‐phase chromatography (aminopropyl columns), saponified, methylated, and the resulting fatty acid methyl esters were quantified using an Autosystem XL gas chromatograph (Perkin Elmer, Boston, MA) equipped with a 100 m×0.25 mm capillary column (HP INNOWQAX, Agilent Technologies, DE) as previously described. ²⁸ Thirty‐five individual FAs were identified by comparison with authenticated standards (NuCheck Prep, MN), and are reported as absolute concentrations (µmol/L). Additionally, 5 NEFA subclasses were calculated; total saturated FA, total monounsaturated FA (cis), total n‐3 PUFA, total n‐6 PUFA (cis), and total trans FA.

A pooled human serum sample as a quality control was included at the beginning and the middle of each run (81 samples/run) for quality control purposes. The lower limits of quantification of the NEFA assay is 0.01% total fatty acid (% weight). The intra‐ and interassay coefficients of variation were 0.5% to 4.3% for FAs present at concentrations >25 µmol/L, 1.8% to 7.1% for FAs present at concentrations between 5 and 25 µmol/L, and 2.8% to 11.1% for FAs present at concentrations <5 µmol/L. ²⁴ The NEFA analysis was conducted at the Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA, USA.

Follow‐Up and Ascertainment of CHD Composite, CHD Mortality, and MI

Surveillance for cardiovascular events occurred during annual clinic visits and intervening 6‐month telephone contacts through 1999 and thereafter by twice yearly telephone contacts through June 2015. At each 6‐month interval participants were contacted to request updates on new cardiovascular events and hospitalizations. Medicare data were used to verify cardiovascular events. All interview data, medical records, death certificates, and next of kin reports of the cases for CHD mortality and incident nonfatal MI were reviewed and adjudicated by an events committee. ²⁹ The median follow‐up time for CHD composite was 11.7 years.

CHD mortality and incident nonfatal MI were adjudicated by the CHS events committee. CHD composite was a composite end point that included CHD mortality and nonfatal MI. Briefly, CHD mortality was ascertained using available interview data, medical records, and death certificates by a centralized events committee blinded to other participant data, ²⁹ and CHD was classified as a history of MI, angina, or revascularization, including percutaneous coronary intervention or coronary artery bypass graft. ²⁹ Nonfatal MI included the traditional elements of chest pain, cardiac enzymes, and ECG records. ³⁰

Other Covariates

Age, sex, race, and educational level were reported by participants at enrollment. All participant characteristics were collected at the 1996 to 1997 visit, which serves as baseline for the current analysis. Smoking status (never, former, and current), alcohol intake (none, 1–6 drinks/week, 7–14 drinks/week, >14 drinks/week), and health status were ascertained by questionnaire. Weight, height, waist circumference, fasting glucose, serum albumin, and C‐reactive protein were measured using standard methods. Physical activity (quantified as metabolic equivalents per week) was assessed using Minnesota Leisure‐Time Activities questionnaire. ³¹ Renal function was assessed based on cystatin C for estimated glomerular filtration rate (mL/min per 1.73 m²). ³² Information regarding diabetes mellitus (defined as fasting glucose ≥7 mmol/L [126 mg/dL], oral glucose tolerance test ≥11.1 mmol/L [200 mg/dL], or use of oral hypoglycemic medications or insulin) and hypertension (defined as systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or treatment with blood pressure lowering medications plus reported physician diagnosis of hypertension) were collected using standardized protocols.

Statistical Analysis

To characterize the study population at the analysis baseline, we calculated means and SDs for continuous measures and percentages for binary and categorical variables. Five NEFA subclasses and individual NEFA were expressed as absolute serum concentration (µmol/L). The adjustments were made by inclusions of all 5 NEFA subclasses for NEFA subclasses analysis or all 35 NEFAs for individual NEFA analysis in the Cox regression models, without the need for correction of multiple testing. The subclasses and the individual NEFAs were standardized and scaled as association with CHD mortality and nonfatal MI per 1 SD. Cox proportional hazards regression was used to estimate the risk of CHD composite, CHD mortality, and incident nonfatal MI. Time to first event or censoring was calculated as the time from the date of the 1996 to 1997 study visit to the earliest date of nonfatal MI or CHD, date of death, date of loss to follow‐up, or date of administrative censoring (June 2015). The NEFA subclasses and individual NEFAs were used for the primary and secondary analyses, respectively.

Multivariable analyses were adjusted in nested models as follows: model 1—age, sex, race, field center, education, and other NEFA subclasses for the primary analysis or all other 35 NEFAs for the secondary analysis; model 2—all the covariates in model 1 plus weight, height, smoking, physical activity, serum albumin, alcohol intake, and renal function; and model 3—all the covariates in the model 2 plus hypertension and diabetes mellitus.

All analyses were conducted in R software (version 3.6.3; Vienna, Austria) using survival and ggplot2 packages, and statistical significance was defined as 2‐tailed α≤0.05.

Results

Characteristics of Study Participants

The mean age of the 1681 participants at year 9 was 77.6±4.4 years, body mass index was 26.7±4.4, and waist circumference was 96.2±12.9 (Table 1). Of the participants, over one‐third were male, one‐seventh were Black, and approximately three‐fifths were diagnosed with hypertension. The participants were relatively young and healthy, compared with those who were not included (Table S1).

Table 1.

Baseline Characteristics of 1681 Participants in the Cardiovascular Health Study Cohort at Baseline in 1996 to 1997

Characteristics	Participants (N=1681)
Age, y	77.6±4.44
Male, %	35.8
Black, %	14.5
Cardiovascular Health Study clinic, %
California	29.0
Maryland	20.5
North Carolina	23.4
Pennsylvania	27.1
Educational attainment, %
≥ High school	50.8
Smoking status, %
Never smoked	51.4
Former smoker	40.3
Current smoker	8.3
Alcoholic drinks/wk, %
0	55.0
1–6	30.7
7–14	8.4
>14	5.9
Hypertension, %	59.7
Diabetes mellitus, %	2.4
Prevalent atrial fibrillation, %	3.3
Prevalent congestive heart failure, %	2.9
Prevalent stroke, %	3.6
Prevalent transient ischemic attack, %	2.7
Hypertension medication, %	46.5
Estrogen, %	19.5
Fasting glucose, mg/dL	97.7±14.5
Albumin, g/dL	3.8±0.29
Body mass index, kg/m2	26.7±4.4
Cystatin C for estimated glomerular filtration rate	73.2±18.5
C‐reactive protein, mg/dL, log2	1.20±1.58
Waist circumference, cm	96.2±12.9

Values are presented as mean±SD for continuous variables and percent for categorical variables.

The most abundant serum individual NEFAs (µmol/L) in the CHS participants were oleic acid (153±63.4 µmol/L), palmitic acid (125±44.8 µmol/L), linoleic acid (80.0±32.7 µmol/L), and stearic acid (60.5±16.8), contributing to 83.6% of total serum NEFAs at the baseline measurement (Table S2). Long‐chain saturated FAs were more strongly correlated with long‐chain monounsaturated FAs and trans FAs and correlated less strongly with n‐6 and n‐3 PUFAs (cutoff: 0.5; Figure S1). Additionally, long‐chain monounsaturated FAs were strongly correlated with long‐chain n‐6 PUFAs and trans FAs and had moderate correlations with long‐chain n‐3 PUFAs.

Association of Serum Nonesterified FAs and CHD Composite

During the 11.7‐year follow‐up period, 434 cases of CHD composite, 266 cases of CHD death, and 271 cases of incident nonfatal MI occurred.

In the NEFA analysis, no significant associations were observed between NEFA subclasses (Table S3) or individual NEFAs (Table S4) and CHD composite in the fully adjusted model.

Association of Serum Nonesterified FAs and CHD Mortality

In NEFA subclass analysis, no significant associations were identified between NEFA subclasses and CHD mortality (Table S3).

Table S5 provides the results for 35 individual NEFAs. Table 2 is an excerpt from the table to highlight the key findings. In the multivariable fully adjusted Cox regression model nonesterified lauric acid (12:0) was associated with a 24% (95% CI, 2%–41%) lower risk of CHD mortality, whereas dihomo‐γ‐linolenic acid (20:3n‐6) was associated with a 34% higher risk of CHD mortality (95% CI, 2%–76%) after adjusting for all covariates (Table 2).

Table 2.

Significant Findings From Multivariable Adjusted Hazard Ratios Relating 35 Individual Serum NEFAs With Coronary Heart Disease Mortality and Incident Nonfatal Myocardial Infarction in the Cardiovascular Health Study Cohort With Baseline in 1996 to 1997

NEFAs, µmol/L Per SD	Model 1*		Model 2 †		Model 3 ‡
	HR (95% CI)	P Value	HR (95% CI)	P Value	HR (95% CI)	P Value
CHD mortality
Lauric acid, 12:0	0.75 (0.59–0.97)	0.026	0.74 (0.57–0.96)	0.024	0.76 (0.59–0.98)	0.033
Dihomo‐γ‐linolenic acid, 20:3n‐6	1.40 (1.06–1.85)	0.018	1.34 (1.02–1.76)	0.037	1.34 (1.02–1.76)	0.035
Nonfatal MI
Elaidic acid, 18:1n‐9t	1.40 (0.98–2.02)	0.069	1.53 (1.06–2.22)	0.025	1.46 (1.01–2.12)	0.045

Values are hazard ratio (95% CI) per SD (n=1681). All 35 individual NEFAs were included in a single model in the Cox proportional hazard regression to estimate the risk of CHD mortality and incident nonfatal MI with 1 SD increment of each NEFA. CHD indicates coronary heart disease; HR, hazard ratio; MI, myocardial infarction; and NEFA, nonesterified fatty acid.

^*Model 1 adjusted for age, sex, race, field center, education, and all 35 NEFAs.

^†Model 2 adjusts for model 1 covariates plus smoking status, serum albumin, alcohol consumption, cystatin C for estimated glomerular filtration rate, weight, height, and physical activity.

^‡Model 3 adjusts for Model 2 covariates plus hypertension and diabetes mellitus.

Association of Serum Nonesterified FAs and Risk of Incident Nonfatal MI

No significant associations were identified between NEFA subclasses and incident nonfatal MI (Table S3).

Table S6 provides the results for 35 individual NEFAs. Table 2 is an excerpt from the table to highlight the key findings. In the multivariable fully adjusted Cox regression model, elaidic acid (trans‐18:1n‐9, 18:1n‐9t) was associated with a 46% higher incident of nonfatal MI (95% CI, 1%–112%) per each SD increment (Table 2).

Discussion

To our knowledge, this is the first study to profile subclasses and individual NEFAs and assess their relations with incident nonfatal MI and CHD mortality, in a community‐based prospective study among older adults. Overall, no significant associations between subclasses of NEFA and incident nonfatal MI and CHD mortality were observed. In individual NEFA analysis, the findings were associated with 2 exploratory outcomes, and there were no significant associations between individual NEFAs and CHD composite. In a multivariable adjusted model, positive associations of nonesterified dihomo‐γ‐linolenic and elaidic acids with CHD mortality and incident nonfatal MI were identified, respectively. An inverse relation of medium‐chain nonesterified lauric acid with CHD mortality was also observed. Although the concentrations of these individual NEFAs were relatively low in serum, the physiological effects may be of clinical importance. These findings extend the knowledge of previous work reporting positive associations of plasma total NEFAs with CHD risk. ¹⁴ , ¹⁵ , ¹⁹

In the CHS participants, no NEFA subclasses or individual NEFAs were associated with CHD composite, and the associations for exploratory outcomes were inconsistent. A plausible explanation for this observation might be the differences of selection criteria between these 2 outcomes. Nonfatal MI was adjudicated by chest pain, cardiac enzymes, and ECG records ³⁰ and involves both type 1 (atherothrombosis) and type 2 MI (supply‐demand mismatch from heterogeneity events) in this cohort. ²⁹ , ³⁰ These were different pathophysiologically from fatal MI, which could lead to severe cardiac pump dysfunction, myocardial rupture, or arrhythmic death, ³³ and was considered as CHD death in this cohort. An alternate explanation relates to the FA reporting unit, absolute concentration versus percentage. Some prior work has suggested this variable may influences the final conclusions. ³⁴ , ³⁵

Observed was a positive association between nonesterified dihomo‐γ‐linolenic acid and CHD mortality. This finding is consistent with prior reports from prospective studies that identified a positive association between plasma dihomo‐γ‐linolenic acid in both the PL and cholesteryl ester fractions and CHD risk. ²² , ³⁶ Also consistent with this finding, data from a nested case‐control study of the WHI (Women's Health Initiative) Observational Study. ³⁷ WHI participants with CHD, compared with those without CHD, had lower estimated δ‐5 desaturase activity, higher PL γ‐linolenic acid, and dihomo‐γ‐linolenic acid concentrations. ³⁷ A plausible mechanism may be through hyperinsulinemia, ³⁸ , ³⁹ a critical risk factor involved in the development of cardiovascular disease (CVD), including CHD. Dihomo‐γ‐linolenic acid is a precursor of arachidonic acid and can be synthesized endogenously from dietary linoleic acid, an essential FA, by a series of enzymic reactions including desaturation (δ‐6 and δ‐5 desaturases) and elongation (FA elongase 5). Hyperinsulinemia has been linked to increased concentration of PL dihomo‐γ‐linolenic acid and lower activity of δ‐5 desaturase (fatty acid desaturase 1, FADS1). ³⁹ , ⁴⁰ Although the causality and molecular mechanism remain unclear, genetic variation in FADS1 (TT genotype in rs174537) may play an important role in dihomo‐γ‐linolenic acid concentration ⁴¹ and is associated with insulin resistance. ⁴⁰ Together, the observation in this and other studies suggests that nonesterified dihomo‐γ‐linolenic acid may be a risk biomarker for CHD risk.

An inverse association was observed between nonesterified lauric acid (12:0) and CHD mortality. Prior work has identified a positive association between higher intake of lauric acid and incident MI and CHD. ⁴² , ⁴³ Lauric acid (12:0) is a medium‐chain FA derived primarily from dairy fat and palm kernel oil ⁴⁴ or de novo lipogenesis. It is an intermediate for long‐chain FA synthesis and has been reported to be rapidly oxidized by the liver. ⁴⁵ Because circulating nonesterified lauric acid in the fasting state is most likely synthesized endogenously, ⁴⁶ caution should be taken when interpreting these data.

In the CHS cohort a positive association was observed between serum nonesterified elaidic acid and incident nonfatal MI, but not with other nonesterified total or other trans FAs isomers. Previously, on the basis of a CHS nested case‐control study, higher serum PL 18:2t and lower 18:1t were positively associated with fatal ischemic heart disease and sudden cardiac death. ⁴⁷ These data are consistent with some but not all prior work. The relation between serum PL and nonesterified total trans FAs and CVD, CHD, and CHD mortality ²⁰ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² appears to be dependent on whether the trans FA isomers were more likely to be derived from ruminant or partially hydrogenated fat. ⁵³ , ⁵⁴ Regardless of the serum lipid fraction (PL, triacylglycerol or cholesteryl ester), 2 systematic reviews have reported a positive association between incident CHD and CVD risk factors and trans FA isomers of predominance in partially hydrogenated fat, particularly elaidic acid (18:1n‐9t) and linelaidic acid (18:2t). ⁵³ , ⁵⁵ An inverse association was also reported between incident CHD and CVD risk factors and trans FA isomers of predominance in ruminant fat, particularly trans‐palmitoleic acid (16:1n‐7t) and conjugated linoleic acid (18:2CLA). ⁵³ , ⁵⁵ Similar associations were observed in adipose tissues and MI risk. ⁴⁸ , ⁵⁶ , ⁵⁷ , ⁵⁸ Discrepancies between the current and prior reports may be, in part, owing to the relative proportion of the ruminant or partially hydrogenated fat in the habitual diets of the cohorts, preferential incorporation into different lipid fractions, analytical ability to distinguish among the trans FA isomers, and/or covariates included in the analytical models. Food and Drug Administration legislation for mandatory inclusion of trans FA on Nutrition Facts labels and removal of partially hydrogenated fat from the Generally Recognized as Safe list has resulted in a drastic decline in the trans FA content of the food supply, ⁵⁹ and hence may have an impact on any observed nonesterified trans FA‐CVD association. This also emphasized the importance of dietary source of trans FAs.

Strengths of this study are that the serum samples, incident nonfatal MI and CHD mortality data were collected in well‐established research centers that also compiled extensive data on cardiometabolic risk factors, lifestyle, and demographics with little loss to follow‐up. To account for the intercorrelation among individual NEFAs, adjusting for all 35 individual NEFAs in a single model allowed for the calculation of independent associations between individual NEFAs and incident nonfatal MI and CHD mortality. From an analytical perspective, coefficients of variation for individual NEFAs were low, particularly when some NEFA were present at low concentrations. A limitation was that the CHS participants were older adults, which limits the generalizability of the findings to other populations. Measures of the individual serum NEFAs were available at only 1 time point and not longitudinally, and thus serial measurements of individual NEFAs need to be considered in the future studies. Lastly, we were not aware of long‐term studies that have assessed the long‐term stability serum NEFA. There was a potential risk of PUFA oxidation before analysis, although the samples were stored at −80°C and never thawed.

Conclusions

In CHS participants, no significant association with CHD composite was identified. Both nonesterified elaidic and dihomo‐γ‐linolenic acids were positively associated with risk of nonfatal MI and CHD mortality, respectively, suggesting potential susceptibility biomarkers for these 2 cardiovascular events.

Sources of Funding

This study is supported by the National Institute on Aging, National Institutes of Health (R01AG053325, PIs Djoussé and Mukamal) and the U.S. Department of Agriculture (agreement no. 58‐1950‐4‐401, PI Lichtenstein). The CHS was supported by contracts HHSN268201200036C, HHSN268200800007C, HHSN268201800001C, N01HC55222, N01HC85079, N01HC85080, N01HC85081, N01HC85082, N01HC85083, N01HC85086, and grants U01HL080295 and U01HL130114 from the National Heart, Lung, and Blood Institute, with additional contribution from the National Institute of Neurological Disorders and Stroke. Additional support was provided by R01AG023629 from the National Institute on Aging. A full list of principal CHS investigators and institutions can be found at CHS‐NHLBI.org. The content is solely the responsibility of the authors and does not necessarily represent the official views of the U.S. Department of Agriculture and the National Institutes of Health.

Disclosures

Kizer reports stock ownership in Bristol‐Myers Squibb, Johnson & Johnson, Medtronic, Merck, and Pfizer. The remaining authors have no disclosures to report.

Supporting information

Tables S1–S6

Figure S1

(J Am Heart Assoc. 2021;10:e019135. DOI: 10.1161/JAHA.120.019135.)