Each fatty acid was fit as a cubic spline with 4 knots. The presentation excludes the top 1% of values of each fatty acid to illustrate the shape of the association with most of the data. Vertical dashed lines indicate the quintile cut points. Solid lines show hazard ratios, and the shading around the line indicates the pointwise 95% CIs. Density plots for the fatty acid values are shown just above the x-axes. Models were adjusted for age, race (Black vs White), sex, field center, education (less than high school, high school, some college, college), physical activity (number of blocks walked last week), body mass index, waist circumference, alcohol consumption, smoking status (never, former, or current), self-reported health (excellent or very good, good, and fair or poor), prevalent diabetes, systolic blood pressure, use of hypertensive medications, and depression score.
eTable 1. Definitions of Unhealthy Aging Events
eTable 2. Tests of Nonlinearity Comparing Cubic Spline With Linear Model
eTable 3. Hazard Ratio (95% CI) of Unsuccessful Aging Associated With Higher Quintiles of Plasma Phospholipid VLSFAs With Adjustment for EPA+DHA+DPA
eTable 4. Hazard Ratio (95% CI) of Unsuccessful Aging Associated With Higher Quintiles of Plasma Phospholipid VLSFAs in Models Including Potential Mediators (Triglycerides and LDL)
eTable 5. Hazard Ratio (95% CI) of Unsuccessful Aging Associated With Higher Quintiles of Plasma Phospholipid VLSFAs With Adjustment for Lipid-Lowering Medications
eTable 6. Hazard Ratio (95% CI) of Unsuccessful Aging Associated With Higher Quintiles of Plasma Phospholipid VLSFAs With Adjustment for 14:0+16:0+18:0
eTable 7. Hazard Ratio (95% CI) of Unsuccessful Aging Associated With One SD Higher Plasma Phospholipid VLSFAs Excluding First 2 Years of Events After Baseline Measure
eTable 8. Hazard Ratio (95% CI) of Unsuccessful Aging Associated With One SD Higher Plasma Phospholipid VLSFAs Among Those With and Without Prevalent Diabetes
eFigure. Hazard Ratio (Pointwise 95% CI) of Unsuccessful Aging Associated With Plasma Phospholipid Levels of VLSFAs Relative to the Median Value
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Bockus LB, Biggs ML, Lai HTM, et al. Assessment of Plasma Phospholipid Very-Long-Chain Saturated Fatty Acid Levels and Healthy Aging. JAMA Netw Open. 2021;4(8):e2120616. doi:10.1001/jamanetworkopen.2021.20616
Are higher levels of the circulating very-long-chain saturated fatty acids arachidic acid, behenic acid, and lignoceric acid associated with healthy aging?
In this cohort study of 2680 older individuals, participants in the highest quintile of levels of plasma phospholipid lignoceric acid showed a significant 16% reduction in incident unhealthy aging events compared with those in the lowest quintile, with a similar finding for behenic acid (15% reduction in unhealthy events).
These findings suggest that increasing circulating levels of very-long-chain saturated fatty acids may be a novel target to promote healthy aging.
Identifying novel factors that protect against age-related diseases and promote healthy aging is critical to public health. Higher levels of circulating very-long-chain saturated fatty acids (VLSFAs) are integrated biomarkers of diet and metabolism shown to have beneficial associations in cardiovascular disease and total mortality, but whether they are associated with overall healthy aging is unknown.
To examine the association of circulating levels of 3 VLSFAs with unhealthy aging events, including incident chronic disease (cardiovascular disease, cancer, lung disease or severe kidney disease), physical dysfunction, and cognitive decline.
Design, Setting, and Participants
This cohort study used 1992 to 2014 data from the Cardiovascular Health Study (CHS). The CHS is a multicenter, population-based study of cardiovascular disease among older adults. Among the 4559 CHS participants with available fatty acid data, 1879 participants who had an age-related event before their first measurement were excluded. Data analysis was performed in 2020.
Main Outcomes and Measures
Plasma phospholipid VLSFA levels were measured by thin-layer chromatography followed by gas chromatography. The main outcome was the hazard ratio (HR) of an incident unhealthy aging event associated with serial measures of plasma arachidic acid, behenic acid, and lignoceric acid.
Among the 2680 study participants (976 men [36.4%]), the mean (SD) age was 74.7 (4.8) years old at entry. During a median (interquartile range) of 6.4 (2.9-12.9) years of follow-up, 2484 participants experienced an unhealthy event. Compared with the lowest quintile, levels of behenic acid in the highest quintile of the fatty acid distribution were associated with 15% lower risk of an unhealthy event (HR, 0.85; 95% CI, 0.74-0.97; P for trend = .01) after adjustment for demographic characteristics, lifestyle factors, and clinical conditions. In analogous comparisons, levels of lignoceric acid were similarly associated with 16% lower risk of an unhealthy event (HR, 0.84; 95% CI, 0.73-0.95; P for trend = .001).
Conclusions and Relevance
These findings suggest that higher levels of circulating behenic acid and lignoceric acid are associated with lower risk of unhealthy aging events. These results highlight the need to explore determinants of circulating VLSFAs for potential novel efforts to promote healthy aging.
With increasing life expectancy, the population of older adults is growing rapidly worldwide.1 However, increased longevity does not necessarily translate into an increase in healthy life span, as older adults experience high rates of cardiovascular and other chronic diseases.2 Identifying novel factors that protect against age-related disease and promote healthy aging is critical to public health.
Higher levels of circulating very-long-chain saturated fatty acids (VLSFAs), saturated fatty acids with 20 carbons or more, are integrated biomarkers of diet and metabolism that are associated with lower risk of heart failure,3,4 atrial fibrillation,5 and diabetes,6-8 which are chronic diseases that contribute to unhealthy aging, and lower risk of sudden cardiac arrest9 and total mortality.10 These widely beneficial associations led us to hypothesize that higher circulating levels of VLSFAs may be broadly associated with a greater likelihood of healthy aging.
We used data from the Cardiovascular Health Study (CHS), a cohort study of factors associated with risk of cardiovascular disease among older adults,11 to examine the associations of plasma phospholipid levels of 3 circulating VLSFAs—arachidic acid (saturated fatty acid with 20 carbons), behenic acid (saturated fatty acid with 22 carbons), and lignoceric acid (saturated fatty acid with 24 carbons)—measured at the current study baseline and up to 2 more times during follow-up, with risk of an incident unhealthy aging event.
CHS is a population-based cohort study of cardiovascular disease among older adults.11 Participants were recruited from 4 US communities (Forsyth County, North Carolina; Sacramento County, California; Washington County, Maryland; and Allegheny County, Pennsylvania) from a random sample generated from the Health Care Financing Administration files. The cohort consists of 5201 noninstitutionalized men and women, aged 65 years or older, recruited in 1989 through 1990, plus an additional 687 predominantly Black participants recruited in 1992 through 1993. The study was conducted using data from 1992 to 2014, and participants were followed up for a median (interquartile range [IQR]) follow up of 6.4 (2.9-12.9) years. Race was classified by self-identification. Race was assessed in the study to investigate potential differences in associations of risk factors with disease outcomes. Each center’s institutional review board approved the study, and all participants provided written informed consent. The present cohort study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for reporting methods, findings, and study limitations for cohort studies.
Plasma phospholipid fatty acids were measured up to 3 times: on blood drawn in 1992 through 1993 (the study baseline; 3941 participants), on blood drawn in 1998 through 1999 (2609 participants), and on blood drawn in 2005 through 2006 (933 participants). Among the 4559 participants with available fatty acid data from at least 1 time point, we excluded 1879 participants who had an age-related event before their first fatty acid measurement. The remaining 2680 participants were included in the analyses. Among these, 2434 had fatty acid data in 1992 to 1993, 1696 had data in 1998 to 1999, and 700 had data in 2005 to 2006.
Blood was drawn after at least 8 hours of fasting, and plasma specimens were stored at −70 °C. Plasma lipids were extracted by the method of Folch.12 Phospholipids were separated from other lipids by thin-layer chromatography. Fatty acid methyl esters were prepared by direct trans-esterification of the phospholipid fraction13 and were separated by gas chromatography using a fused-silica 100-m capillary column, as previously described.14 Fatty acids were expressed as a weighted percentage of total fatty acids. Interassay coefficients of variation for the VLSFA measurements were less than or equal to 3.5%.
Participants were followed up by means of annual study-clinic examinations with interim telephone contacts for 10 years and telephone contacts every 6 months thereafter. Cardiovascular events were adjudicated by a centralized Event Committee. Unhealthy aging was defined on the basis of the established definition in CHS, as previously described,15 and as shown in eTable 1 in the Supplement. In brief, we defined an unhealthy event as any of the following: incident cardiovascular disease (myocardial infarction, heart failure, stroke, transient ischemic attack, or claudication), incident severe kidney disease (estimated glomerular filtration rate <10 mL/min/1.73 m2 or dialysis), incident chronic obstructive pulmonary disease, incident cancer, a decrease in cognition (first Mini Mental State Examination score ≤80), or an increase in difficulties of activities of daily living (first record of difficulty with ≥1 daily living activity).
Information on medical history, medications, lifestyle, and clinical risk factors was collected at annual clinic visits, as previously reported.11 Diabetes was defined by glucose greater than or equal to 126 mg/dL (to convert to millimoles per liter, multiply by 0.0555) when participants reported fasting for 8 hours or more before venipuncture, glucose greater than or equal to 200 mg/dL when fasting was less than 8 hours, or use of insulin or oral hypoglycemic medication. Body mass index (BMI) was calculated as body weight in kilograms divided by height in meters squared. Waist circumference was measured at the umbilicus. Lipids, glucose, insulin, and inflammatory biomarkers were assessed on fasting blood samples using enzymatic methods.11 The self-reported depression score of 1 to 10 was based on the 10-item Centers for Epidemiological Studies’ Depression Scale.16
Baseline (unadjusted) demographic characteristics, cardiometabolic risk factors, and lifestyle habits for the study population were summarized according to quintiles of lignoceric acid. We used a Cox proportional-hazards model to evaluate the association between time-varying VLSFA levels, adjusting for time-varying covariates (updated at each fatty acid measurement) and the likelihood of unhealthy aging. Time to event was calculated as the time elapsed between VLSFA measurement and the earliest of ascertainment of first unhealthy event, death, or date of last follow-up in 2014. The functional forms of the associations between VLSFAs and unhealthy aging were evaluated using natural cubic splines, and tests for nonlinearity were performed by comparing the spline model with a linear model using the likelihood ratio test. Quintiles of VLSFA levels were defined from the baseline distributions. Comparisons between quintiles of VLSFA levels were evaluated using time-dependent indicator variables, with tests for trend based on a time-dependent linear variable for quintile. The associations were adjusted for age, race (Black vs White), sex, field center, education (less than high school, high school graduate, some college, and college graduate), physical activity (number of city blocks walked in the previous week), BMI, waist circumference, alcohol consumption, smoking status (never, former, or current), self-reported health (excellent or very good, good, and fair or poor), prevalent diabetes, systolic blood pressure, use of hypertensive medications, and depression score. Modification of the association of VLSFA with the risk of an unhealthy event was evaluated for age (linear), sex, BMI (linear), and prevalent diabetes with statistical significance assessed via the Wald test for the interaction term, in models where VLSFAs were modeled linearly.
Several sensitivity analyses were performed. To assess the role of potential mediators, we fit a model that also included variables for triglycerides and low-density lipoprotein (LDL) and one further adjusted for palmitic acid (saturated fatty acid with 16 carbons). To assess whether other factors might confound the associations of VLSFA, we fit separate models adjusting for plasma phospholipid eicosapentaenoic acid, docosahexaenoic acid, and docosapentaenoic acid; lipid-lowering medications; and the sum of plasma phospholipid long-chain saturated fatty acids (myristic acid plus palmitic acid plus stearic acid). Finally, to examine whether reverse causation might have created these associations, we performed analyses excluding the first 2 years of events.
Missing covariates (<4% for all covariates) were imputed using data on age, sex, race, education, smoking, alcohol use, BMI, physical activity, self-reported health status, and prevalent coronary heart disease at baseline, as previously described.15 Results with imputed data are presented. The results were unchanged from analyses that excluded participants with missing values.
The aforementioned statistical tests were 2-sided, and significance was set at P < .05. Analyses were performed using Stata statistical software version 14.0 (StataCorp) and R statistical software version 4.05 (R Project for Statistical Computing). Data analysis was performed in 2020.
A total of 2680 participants (976 men [36.4%]) were included in this analysis. Participants’ mean (SD) age was 74.7 (4.8) years at entry. Median (IQR) levels of the plasma phospholipids were 1.38% (1.21%-1.58%) of total fatty acids for lignoceric acid, 1.66% (1.46%-1.87%) for behenic acid, and 0.50% (0.44%-0.55%) for arachidic acid. Baseline characteristics of the study participants across quintiles of lignoceric acid are shown in Table 1. Participants in the highest quintile of lignoceric acid were more likely than participants in the lowest quintile to be male (217 men [40.8%] vs 146 men [26.7%]), Black (98 participants [18.4%] vs 47 participants [8.6%]), college graduates (158 participants [29.7%] vs 118 participants [21.6%]), and current smokers (52 participants [9.8%] vs 36 participants [6.6%]), with higher levels of LDL (mean [SD], 139.18 [30.79] vs 112.62 [31.76] mg/dL; to convert to millimoles per liter, multiply by 0.0259) and physical activity (mean [SD] number of blocks walked in previous week, 47.75 [69.76] vs 37.80 [62.56]), but they were less likely to have diabetes (91 participants [17.1%] vs 143 participants [26.2%]) and had lower mean levels of triglycerides (mean [SD], 107.91 [46.45] vs 192.32 [121.45]) mg/dL; to convert to millimoles per liter, multiply by 0.0113).
During up to 23 years of follow-up, 2484 participants in the previously healthy population experienced an event of unhealthy aging. The distribution of the first events is shown in Table 2, with the most frequent first event being CVD (880 participants) and the least frequent being severe kidney disease (12 participants). After adjustment for demographic variables, adiposity measures, physical activity, smoking, alcohol, diabetes, self-reported health, hypertension, and depression score, higher levels of plasma phospholipids lignoceric acid and behenic acid were each associated with lower risk of unhealthy aging (comparing the highest quintile with the lowest, 16% lower risk for lignoceric acid [HR, 0.84; 95% CI, 0.73-0.95; P for trend = .001] and 15% lower risk for behenic acid [HR, 0.85; 95% CI, 0.74-0.97; P for trend = .01]) (Table 3). A 1 SD higher value of lignoceric acid was associated with 15% lower risk of unhealthy aging (HR, 0.85; 95% CI, 0.77-0.94). Analyses using cubic splines showed that the associations with lower risk were linear over most of the data, as suggested from the quintile analyses shown in Table 3, with possibly higher risk at very high levels of behenic acid and lignoceric acid (Figure; eFigure in the Supplement). However, tests of nonlinearity comparing cubic splines with linear models did not show evidence of nonlinearity (eTable 2 in the Supplement). Further adjustments for levels of eicosapentaenoic acid, docosahexaenoic acid, and docosapentaenoic acid, which may be associated with healthier aging17; for the potential mediators triglycerides and LDL; for lipid-lowering medications; and for the sum of plasma phospholipid long-chain saturated fatty acids (myristic acid plus palmitic acid plus stearic acid) did not appreciably change the results (eTable 3, eTable 4, eTable 5, and eTable 6 in the Supplement). Adjustment for the combination of triglycerides, LDL, and plasma phospholipid palmitic acid attenuated the associations slightly (Table 4). Finally, excluding the unhealthy aging events that occurred in the first 2 years to assess the possibility of reverse causation did not change the study results (eTable 7 in the Supplement).
The associations of each plasma phospholipid VLSFA with unhealthy aging did not vary according to age, sex, and BMI (smallest observed P for interaction, .22). However, we found significant interactions between prevalent diabetes and levels of arachidic acid (P for interaction = .002) and behenic acid (P for interaction = .01). We observed associations of arachidic acid and behenic acid with lower risk of unhealthy aging in the larger group of participants without diabetes (HR for 1 SD higher level, 0.83 [95% CI, 0.74-0.94] for arachidic acid and 0.79 [95% CI, 0.70-0.90] for behenic acid) but not in the smaller group of participants with diabetes (HR for 1 SD higher level, 1.20 [95% CI, 0.97-1.48] for arachidic acid and 1.10 [95% CI, 0.88-1.37] for behenic acid) (eTable 8 in the Supplement).
To our knowledge, this cohort study is the first to show that higher levels of plasma phospholipid VLSFAs may be associated with improved healthy life span, as shown by lower risk of unhealthy aging events. The most significant association was with lignoceric acid. Compared with the lowest quintile, the quintile with the highest levels of lignoceric acid was associated with a 16% lower risk of an unhealthy aging event, after adjustment for demographic variables, lifestyle factors, and diabetes.
The primary factor associated with incident unhealthy aging events in this cohort was cardiovascular disease, followed by declining physical function, cancer, and declining cognition. Higher VLSFA levels have beneficial associations with mortality risk,10 multiple metabolic and inflammatory parameters,18-21 and a wide range of incident cardiovascular and metabolic diseases, such as heart failure,4 atrial fibrillation,5 and diabetes.7,8 Further studies will be needed to investigate potential associations between VLSFAs and cognition and other age-related conditions. Our purpose here was to study healthy aging, which is considered as living without chronic disease and with intact physical and mental functions. Identifying people resistant to the effects of aging compared with their peers and identifying potentially modifiable determinants of this resistance are both of great public health importance.
The univariable association of higher levels of VLSFAs with male sex, Black race, smoking, and an increase in LDL is surprising, because none of these factors has been previously associated with healthy aging. Considering that high LDL levels are associated with coronary artery disease,22 it is interesting that higher levels of VLSFAs have consistent and beneficial associations with risk of heart disease, including coronary artery disease,23,24 despite their LDL association. There are also negative associations between VLSFA levels and levels of triglycerides and palmitic acid, markers of de novo lipogenesis,20 and we previously reported that these markers mediate the associations of VLSFAs with lower risk of diabetes.7,8 However, adjusting for levels of triglycerides, palmitic acid, and LDL or prevalent diabetes did not negate the associations between VLSFAs and healthy aging. Interestingly, we found significant effect modification by diabetes and no benefits of VLSFAs on healthy aging in patients with prevalent diabetes. Although this interaction should be interpreted cautiously until replicated, it further illustrates the complex physiological associations of these compounds.
VLSFAs may be associated with the aging process through multiple mechanisms. They are major components of ceramides and sphingomyelins, lipids that contain 1 fatty acid acylated to a sphingoid backbone.25 Ceramides, specifically, are known for their role in apoptosis26 and inflammation27 and appear to contribute to the pathogenesis of a broad range of disorders, including insulin resistance and atherosclerosis.28-30 Interestingly, in experimental cell and animal studies, ceramides that contain a VLSFA may actually counteract this tendency by having biological activities opposite from ceramides that contain the shorter saturated fatty acid palmitic acid.31,32 We speculate that VLSFAs may have a positive effect on healthy aging by lowering endogenous levels of shorter-chain ceramides.
The sources of circulating VLSFAs are both dietary and metabolic. Dietary sources include peanuts, macadamia nuts, and canola oil.33 Small short-term feeding trials have shown an increase in circulating levels of VLSFAs with consumption of both macadamia nuts34 and peanut butter,35 illustrating that dietary intake may directly be associated with circulating levels of VLSFAs. We reported previously that higher levels of behenic acid and lignoceric acid were associated with greater peanut intake measured 3 years earlier.8 VLSFA may also be synthesized endogenously from stearic acid by elongases, such as the ubiquitously distributed ELOVL1.36 In agreement with the predominant occurrence of VLSFAs in ceramides and other sphingolipids, ELOVL1 appears to be coregulated with ceramide synthase 2, which produces ceramide by addition of a VLSFA to sphingosine.37 Furthermore, we have previously shown an association of circulating VLSFAs with genetic variation in the sphingolipid synthesis pathway.38 The extent to which dietary intake and endogenous metabolism contribute to circulating levels of VLSFA is currently unknown.
The study has several limitations and strengths. This is an observational study and causality cannot be established, as residual confounding by unknown factors is possible. However, the results were robust to adjustment for multiple characteristics. Reverse causality is also of potential concern in association studies39 and would operate, for example, if frailty led to low levels of VLSFAs. However, we eliminated participants with a wide range of medical conditions and symptoms in this study of incident unhealthy aging and examined the association of VLSFAs in a particularly healthy group at entry. We also observed associations in the healthier subgroup without prevalent diabetes. Furthermore, we did not see attenuation of the associations over time, and exclusion of the first 2 years of follow-up did not remove or attenuate the observed associations, suggesting that reverse causation does not account in large part for the observed associations. The study was conducted among older adults, and the results may not be generalizable to other populations; however, healthy aging is of high relevance to this older population. The study was limited to White and Black participants. Strengths include the prospective design, population-based enrollment, use of an objective marker of diet and metabolism, repeated measurements, and the rich information available on demographic variables, risk factors, and lifestyle habits.
In conclusion, to our knowledge, we report for the first time an association of higher levels of plasma phospholipid lignoceric acid and behenic acid with lower risk of incident unhealthy aging events. Together with our previous reports of inverse associations of VLSFA with incident heart failure,4 incident atrial fibrillation,5 incident sudden cardiac arrest,9 and total mortality,10 the study findings should prompt further research on determinants of VLSFA levels and may lead to novel approaches to promote healthy aging.
Accepted for Publication: June 6, 2021.
Published: August 12, 2021. doi:10.1001/jamanetworkopen.2021.20616
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Bockus LB et al. JAMA Network Open.
Corresponding Author: Rozenn N. Lemaitre, PhD, MPH, Department of Medicine, University of Washington, 1730 Minor Ave, Ste 1360, Seattle, WA 98101 (email@example.com).
Author Contributions: Dr Lemaitre had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: de Olivera Otto, King, Siscovick, Mozaffarian, Lemaitre.
Acquisition, analysis, or interpretation of data: Bockus, Biggs, Lai, Fretts, McKnight, Sotoodehnia, Song, Siscovick, Mozaffarian, Lemaitre.
Drafting of the manuscript: Bockus.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Lai, de Olivera Otto, Fretts, McKnight, Siscovick.
Obtained funding: Siscovick, Mozaffarian, Lemaitre.
Administrative, technical, or material support: Biggs, Lai, Fretts, King, Song.
Supervision: McKnight, Lemaitre.
Conflict of Interest Disclosures: Dr Biggs reported receiving grants from the National Institutes of Health (NIH) outside the submitted work. Dr Fretts reported receiving grants from the NIH outside the submitted work. Dr McKnight reported receiving grants from NIH outside the submitted work. Dr Song reported receiving grants from NIH outside the submitted work. Dr Mozaffarian reported receiving chapter royalties from UpToDate; receiving grants from the NIH, Gates Foundation, and Rockefeller Foundation; receiving personal fees from Acasti Pharma, America’s Test Kitchen, Barilla, Cleveland Clinic Foundation, Danone, GOED, and Motif FoodWorks; and serving on the scientific advisory boards of Beren Therapeutics, Brightseed, Calibrate, DayTwo, Elysium Health, Filtricine, Foodome, HumanCo, January, Inc, Perfect Day, Season, and Tiny Organics outside the submitted work. Dr Lemaitre reported receiving grants from NIH outside the submitted work. No other disclosures were reported.
Funding/Support: This research was supported by grants R01-HL085710 and R01-HL094555 from the National Heart, Lung, and Blood Institute (NHLBI), with cofunding from the Office of Dietary Supplements. The CHS cohort was supported by contracts HHSN268201200036C, HHSN268200800007C, HHSN268201800001C, N01HC55222, N01HC85079, N01HC85080, N01HC85081, N01HC85082, N01HC85083, N01HC85086, and 75N92021D00006, and grants U01HL080295 and U01HL130114 from the NHLBI, with additional contribution from the National Institute of Neurological Disorders and Stroke. Additional support was provided by grant R01AG023629 from the National Institute on Aging.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Additional Information: A full list of principal CHS investigators and institutions can be found at https://www.CHS-NHLBI.org.