Abstract
Background
Studies evaluating the longitudinal relationship between plasma free fatty acids (FFAs) and subclinical atherosclerosis remain limited. We investigated the prospective association between serum FFA levels and the progression of coronary artery calcification (CAC) in a generally healthy, age-, sex-, and racially/ethnically diverse population.
Methods
This study utilized baseline data from the Multi-Ethnic Study of Atherosclerosis cohort, collected from 2000 to 2002, including 2988 women and 2696 men, with outcome data extending to 2012. CAC progression was defined as either the new onset of detectable CAC among participants with a baseline score of zero or a clinically meaningful increase in CAC severity among those with pre-existing CAC, based on established cut-offs. We employed Cox proportional hazards regression to calculate hazard ratios (HR) and 95% confidence intervals (95%CI) for the associations between FFAs and the CAC progression, stratified by sex. We evaluated potential non-linear relationships using restricted cubic splines.
Results
During the 9-year follow-up, we identified 1302 cases of CAC progression in women and 1480 cases in men. Serum FFAs were linearly associated with CAC progression in men, with a multivariable-adjusted HR of 1.41 (95% CI: 1.03–1.93) for each mmol/L increase in FFAs. Moreover, men in the highest relative to the lowest quintile of FFAs had a 20% higher risk (1.20; 1.02–1.43). Among women, we found evidence of a potential nonlinear association between FFAs and CAC progression, suggesting a threshold effect in their relationship (P non-linearity: 0.050). We observed no significant effect modification by age, body mass index, diabetes status, or hypertension in the relationship between FFAs and CAC progression.
Conclusions
This study highlights a significant association between FFAs and CAC progression in both men and women, though the nature of this association differs by gender.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12872-026-05692-1.
Keywords: Free fatty acids, Coronary artery calcification, Sex, Atherosclerosis, Subclinical
Background
Cardiovascular disease (CVD) remains the leading global cause of morbidity and mortality [1]. Atherosclerosis, a chronic inflammatory process, underlies CVD by promoting plaque buildup in arterial walls [2, 3]. Coronary artery calcification (CAC), measurable by computed tomography (CT), reflects subclinical atherosclerosis and predicts future cardiovascular events [4]. CAC progression, the increase in calcified plaque burden over time [5], is strongly linked to advancing atherosclerosis and elevated CVD risk [6–10]. Therefore, identifying determinants of CAC progression is essential for CVD prevention.
Recent studies indicate that free fatty acids (FFAs), may contribute to CVD when chronically elevated.
FFAs are non-esterified fatty acids characterized by a hydrocarbon chain (typically 12–24 carbons) with a free carboxyl group (-COOH), distinguishing them from esterified forms such as those in triglycerides (TGs) or phospholipids. They are generated primarily through lipolysis of stored TGs in adipose tissue by hormone-sensitive lipase, releasing FFAs into circulation for energy utilization via β-oxidation or re-esterification. Disposal occurs via uptake into tissues (e.g., muscle, liver), oxidation, or recycling back into TGs. FFAs encompass saturated (e.g., palmitic acid, C16:0), monounsaturated (e.g., oleic acid, C18:1), and polyunsaturated varieties (e.g., linoleic acid, C18:2; α-linolenic acid, C18:3), mirroring the general classification of fatty acids. Notably, while total circulating FFAs include PUFAs, these represent only minor components compared to saturated/monounsaturated species [11, 12].
High serum FFA levels have been linked to oxidative stress, endothelial dysfunction, and inflammation, key processes in atherosclerosis and CAC [13]. While some studies report associations between elevated FFAs and adverse CVD outcomes, including coronary artery disease (CAD) [14], coronary heart disease (CHD) [15], heart failure in older adults [16], and CVD mortality [15, 17], findings remain inconsistent. For instance, one cross-sectional study found higher FFAs correlated with greater CAC severity in patients with type 2 diabetes (T2DM) [14]. In contrast, another large U.S. cohort found no significant links between FFAs and cardiovascular events or mortality [17].
To date, the relationship between FFAs and subclinical atherosclerosis, particularly CAC, remains underexplored. Since CAC progression reflects early atherosclerotic changes that develop years before clinical events, it offers valuable insights into the initial biological mechanisms of CVD. This study builds on prior research by focusing on CAC progression, bridging the gap between underlying biology and clinical outcomes. Considering well-established sex differences in CVD prevalence, risk factors, pathophysiology, and lipid metabolism [18–22], we examined the associations between serum FFAs and CAC progression separately in men and women within the Multi-Ethnic Study of Atherosclerosis (MESA) cohort.
Methods
The data used for this study were obtained from the Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC), which can be accessed at https://biolincc.nhlbi.nih.gov/.
Study population
Study participants were selected from the MESA, a multicenter, observational cohort that aims to investigate the prevalence, correlates, and progression of subclinical CVD in individuals without known CVD at baseline. Briefly, the main cohort included 6814 women and men aged 45 to 84 years at baseline, recruited from six U.S. cities, representing various ethnicities, including White, Black, Hispanic, and Chinese participants [23]. The rationale and design have been described previously [23]. The institutional review boards at each participating center approved the study protocols, and all participants provided written informed consent at every examination. This study also received ethical approval from the Research Institute for Endocrine Sciences at Shahid Beheshti University of Medical Sciences, Tehran, Iran (approval number: IR.SBMU.ENDOCRINE.REC.1404.021).
For this study, we included 6814 participants aged 45 to 84 years at baseline (2000–2002), excluding 78 individuals lacking baseline FFAs data. We further excluded 738 participants without follow-up CAC measurements through the end of the study (December 2012). Among the remaining 5998 individuals, 314 with missing data on study variables were also excluded, resulting in a final sample of 5684 participants (2696 men) (Supplementary Fig. 1).
Data collection
Measurement of FFAs
Fasting blood samples were collected at baseline examination (2000–2002) after a 12-hour fast through venous puncture and stored at -70 °C until analysis [23]. Serum FFAs (mmol/L) were measured using the Wako enzymatic method. This technique involves the acylation of coenzyme A by fatty acids in the presence of added acyl-CoA synthetase. The resulting acyl-CoA is then oxidized by acyl-CoA oxidase, producing hydrogen peroxide. In the presence of peroxidase, this hydrogen peroxide facilitates the oxidative condensation of 3-methyl-N-ethyl-N (β-hydroxyethyl)-aniline with 4-aminoantipyrine to form a purple-colored adduct, which is measured colorimetrically at 550 nm. The intra-assay coefficients of variation (CV) were as follows: 5.6% at a concentration of 0.6 mmol/L, 5.4% at a concentration of 0.93 mmol/L, and 6.6% at a concentration of 0.21 mmol/L [17].
Outcome ascertainment
After the baseline examination, participants attended up to four follow-up examinations: Exam 2 (2002–2004), Exam 3 (2004–2005), Exam 4 (2005–2007), and Exam 5 (2010–2012) [23]. At both baseline and follow-ups, coronary calcium was measured with cardiac-gated electron-beam CT or a multi-detector CT depending on the study site [23]. CAC scores were determined by Agatston method [15], which calculates scores by multiplying the area of each calcified lesion by a factor related to the maximum plaque attenuation. The scores for all lesions within a given artery (left main, left anterior descending, circumflex, and right coronary) were then summed to obtain the total calcium score. Each participant was scanned twice, and the mean score was used for analysis [23]. All scans were phantom adjusted and were analyzed independently by two trained radiologist at a central reading center. Rescan agreement was high for both electron beam CT and multidetector CT scanning [24].
CAC progression was identified as: (1) incident CAC (Agatston units (AU) > 0) among participants with a baseline score of zero; (2) an absolute annual increase of > 10 AU among those with a baseline CAC score of 0-100; or (3) a relative annual increase of ≥ 10% among those with baseline CAC score > 100 AU [25].
Covariates
At baseline, information on demographics, medical history, family history, tobacco use, and alcohol consumption was collected by interviewer and self-administered questionnaires. Participants’ height and weight were recorded to compute body mass index (BMI) using the formula: weight (kg) divided by height squared (m²). Three readings of resting blood pressure were taken, and the average of the last two readings was used for analysis. Baseline laboratory measurements for fasting plasma glucose (FPG), total cholesterol (TC), TGs, and high-density lipoprotein cholesterol (HDL-C) were conducted following standard protocols [23]. The Participants’ use of blood pressure, cholesterol, and diabetes medications was evaluated via a questionnaire and confirmed by verifying the actual medications taken [23]. Smoking status was categorized into two groups: never/former smokers and current smokers. Alcohol consumption was classified into three groups: never, former, and current drinkers. A semi-quantitative questionnaire, adapted from the Cross-Cultural Activity Participation Study [26] was utilized to assess physical activity in terms of metabolic equivalents of task (METs). Total physical activity level (PAL) was calculated by summing the minutes per week spent on light, moderate, and vigorous physical activities. Education was classified into three categories: (i) high school degree or lower; (ii) some college, technical school, associate degree, or bachelor’s degree; and (iii) graduate degree or professional school. A positive family history of coronary heart disease (FH CHD) was defined as having a heart attack in a parent, sibling, or child. Diabetes was defined as having a FPG level of ≥ 7.0 mmol/L or as the use of insulin or oral hypoglycemic medications. Hypertension was defined as having a systolic blood pressure (SBP) of ≥ 130 mmHg, a diastolic blood pressure (DBP) of ≥ 80 mmHg, or the use of antihypertensive medications.
Plasma concentrations of ω-3 and ω-6 polyunsaturated fatty acids (PUFAs) were measured in stored fasting samples from baseline, following the method previously described by Cao et al. [27]. The analyzed ω-3 PUFAs included Alpha-Linolenic Acid (ALA), Eicosapentaenoic Acid (EPA), Docosapentaenoic Acid (DPA), and Docosahexaenoic Acid (DHA). The measured ω-6 PUFAs were Linoleic Acid (LA), Gamma-Linolenic Acid (GLA), Docosadienoic Acid, Dihomo-Gamma-Linolenic Acid (DGLA), and Arachidonic Acid (AA). These fatty acids were extracted from phospholipids in plasma samples stored in ethylenediaminetetraacetic acid (EDTA) tubes. Phospholipid fatty acid measurements were expressed as a percentage of total fatty acids. The limit of detection for each fatty acid was 0.03%. Interassay CVs were as follows: 13.5% (ALA), 7.6% (EPA), 8.3% (DPA), 8.5% (DHA), 6.8% (LA), and 7.4% (AA). Total PUFA levels were calculated as the sum of the ω-3 and ω-6 fatty acids.
Statistical analysis
Baseline characteristics of study sample were summarized using means and standard deviations (SD) for normally distributed continuous variables and median (interquartile range) for non-normally distributed variables. We assessed normality for all continuous variables using formal tests (Shapiro-Wilk/Kolmogorov-Smirnov) and visual examination of Q-Q plots. Categorical variables were summarized as frequencies and percentages. Between-group comparisons were performed using independent-samples t tests or analysis of variance (ANOVA) for normally distributed continuous variables, Mann-Whitney U tests or Kruskal-Wallis tests for non-normally distributed continuous variables, and chi-square tests for categorical variables.
We also compared baseline characteristics between respondents who comprised the analytic sample and non-respondents who were excluded from the analysis.
Distribution of the FFAs among participants was visualized using violin plots. Participants were categorized into quintiles based on their FFAs levels, and baseline characteristics were compared across FFAs quintiles.
The crude incidence densities of CAC progression per 1000 person-years, along with 95% confidence intervals (95% CI), were calculated by dividing the number of new cases of CAC progression by the total person-time at risk and then multiplying by 1000.
Follow-up time was calculated as the difference between the date of study entry and either the date of CAC progression or censoring, whichever happened first. Participants were censored at the date of their last imaging visit if they remained free of CAC progression, or at the date of loss to follow-up or death for those who did not complete the study. For participants who developed CAC progression, the date of onset was defined as the midpoint between the follow-up visit at which CAC progression was first identified and the most recent visit without progression. For censored participants, follow-up ended at the date of their last available imaging visit or date of death.
For descriptive purposes, sex-stratified Kaplan-Meier (KM) curves were generated to illustrate the probability of CAC progression over time, with comparisons made using the log-rank test.
We used Cox regression models to estimate the HR and 95% CI for the associations between FFAs and CAC progression in men and women. FFAs were examined in both continuous form and as quintiles, with quintile 1 serving as the reference category. We developed sequential models, beginning with an unadjusted model and then minimally adjusting for age, and race. Further adjustments were made for additional potential confounding variables, including education, smoking status, alcohol intake, PAL (continuous), BMI (continuous), diabetes, hypertension, FH CHD, use of lipid-lowering drugs, TC, HDL-C, TG, and plasma total PUFA, with the last four mentioned variables treated as continuous. Plasma total PUFA was included because it reflects long-term dietary intake and metabolic status, both of which independently influence cardiovascular risk [28] and may confound the association between FFAs and CAC progression.
The assumptions of proportional hazards were tested using the Schoenfeld goodness-of-fit test. We assessed the linear trend across quintiles of FFAs by incorporating the medians of the FFAs quintiles as a continuous variable in the Cox models.
In order to evaluate any potential nonlinear relationship between FFAs and outcome, restricted cubic splines (RCS) were utilized in multivariable Cox models, adjusted for all confounders. To achieve the optimal balance between a good fit and avoiding overfitting in the primary cubic splines models for CAC progression, we tested various numbers of knots, ranging from three to seven. The selection of knots was based on the Akaike Information Criterion (AIC), with the model exhibiting the lowest AIC value being chosen. The concentration with the lowest HR on RCS was the FFA concentration associated with the lowest risk of CAC progression.
In subgroup analysis, we explored whether the associations between serum FFAs (continuous) and CAC progression varied by age (< 60 vs. ≥60 years), race, diabetes, hypertension, and BMI (< 30 kg/m2 vs. ≥30 kg/m2). Each subgroup-specific model was adjusted for the same covariates as in the main analysis (model 3) (age, race, education, smoking status, alcohol intake, total physical activity, BMI, diabetes, hypertension, family history of CHD, use of lipid-lowering drugs, and continuous measures of TC, HDL-C, TG, and total PUFA), except for the variable on which the subgroup was defined. To assess potential effect modification, an interaction term between serum FFAs (continuous) and the subgroup variable (e.g., age group, diabetes status) was added to the multivariable Cox model. The P for interaction was derived using the likelihood ratio test, comparing models with and without the interaction term.
We performed several sensitivity analyses to assess the robustness of our findings. First, to address potential bias from competing events, we conducted Fine-Gray competing risk regression [29]. Unlike standard Cox models, which may overestimate risk by censoring individuals who die from non-cardiovascular causes, this approach treats non-CVD death as a competing event. The event of interest remained CAC progression. We applied the same modeling strategy as the primary analysis: Model 1 (unadjusted), Model 2 (adjusted for age and race), and Model 3 (fully adjusted). All analyses were stratified a priori by sex.
Second, to assess potential selection bias arising from the exclusion of 1130 participants due to missing baseline FFA data, missing follow-up CAC measurements, or incomplete covariate data, we performed inverse probability weighting (IPW). We first estimated each participant’s probability of being included in the analytical sample using logistic regression on baseline covariates (age, sex, race/ethnicity, education, smoking, alcohol use, physical activity, BMI, diabetes, hypertension, lipid-lowering medication use, family history of CHD, TC, HDL-C, TGs, total PUFA, and baseline FFAs levels). IPW were calculated as the reciprocal of the predicted probability (1/predicted probability) and incorporated into Cox regression models with the same covariate adjustment as the primary analysis.
Third, to assess whether FFAs differentially associate with plaque initiation versus progression, we stratified participants by baseline CAC status. Among those with baseline CAC = 0, we assessed incident CAC development; among those with baseline CAC > 0, we evaluated progression of established CAC (defined as previously described). FFAs were modeled as continuous variables and in quintiles using Cox regression with the same adjustment as the primary analysis.
All analyses were conducted using R software (version 4.1.1). A difference was considered statistically significant at a two-sided P-value of < 0.05.
Results
Baseline characteristics of participant
The baseline characteristics of the study participants, stratified by sex, are presented in Table 1. Women had lower education levels and were more likely to be never or former smokers compared to men. Additionally, women were less likely to be current drinker.
The prevalence of diabetes and hypertension was higher in men than in women, while the FH CHD was lower in men. Furthermore, women had significantly lower DBP and FPG, along with higher PAL. Lastly, women had higher TC, HDL-C and serum FFAs levels, while TG was significantly higher in men. The CAC score was markedly higher in men than in women, with the substantially wider interquartile range in men, indicating greater variability.
Table 1.
Baseline characteristics of participants stratified by sex
| variable | Women (n = 2988) |
Men (n = 2696) |
P value |
|---|---|---|---|
| Categorical, n (%) | |||
| Race/ethnicity | |||
| White | 1167 (39.1) | 1093 (40.5) | 0.142 |
| Chinese | 344 (11.5) | 330 (12.2) | |
| Black | 840 (28.1) | 685 (25.4) | |
| Hispanic | 637 (21.3) | 588 (21.8) | |
| Education | |||
| Less than high school/ High school | 1161 (38.9) | 800 (29.7) | < 0.001 |
| Some college | 1378 (46.1) | 1277 (47.4) | |
| Graduate degree or professional school | 449 (15.0) | 619 (23.0) | |
| Smoking status | |||
| Never/ Former | 2651 (88.7) | 2318 (86.0) | 0.002 |
| Current | 337 (11.3) | 378 (14.0) | |
| Alcohol consumption | |||
| Never | 868 (29.0) | 271 (10.1) | < 0.001 |
| Former | 603 (20.2) | 714 (26.5) | |
| Current | 1517 (50.8) | 1711 (63.5) | |
| Diabetes (Yes) | 309 (10.3) | 351 (13.0) | 0.002 |
| Hypertension (Yes) | 1706 (57.1) | 1612 (59.8) | 0.041 |
| Family history of CHD (Yes) | 1285 (43.0) | 981 (36.4) | < 0.001 |
| Lipid lowering drugs (Yes) | 482 (16.1) | 433 (16.1) | 0.971 |
| Antidiabetic drugs (Yes) | 260 (8.7) | 276 (10.2) | 0.088 |
| Antihypertensive drugs (Yes) | 1110 (37.1) | 946 (35.1) | 0.109 |
| Continuous, mean (SD) | |||
| Age (years) | 61.8 (10.1) | 61.9 (10.1) | 0.623 |
| *Physical Activity Level (MET.min/wk) | 8317.5 (6005.0) | 6862.5 (5519.4) | < 0.001 |
| BMI (kg/m2) | 28.7 (6.1) | 27.8 (4.4) | < 0.001 |
| SBP (mm Hg) | 126.2 (22.8) | 125.5 (19.1) | 0.186 |
| DBP (mm Hg) | 69.0 (10.1) | 74.9 (9.3) | < 0.001 |
| *FPG (mmol/L) | 4.8 (0.8) | 5.1 (0.9) | < 0.001 |
| TC (mmol/L) | 5.2 (0.9) | 4.9 (0.9) | < 0.001 |
| HDL-C (mmol/L) | 1.5 (0.4) | 1.2 (0.3) | < 0.001 |
| *TG (mmol/L) | 1.2 (0.9) | 1.3 (1.0) | 0.014 |
| *FFAs (mmol/L) | 0.6 (0.3) | 0.5 (0.2) | < 0.001 |
| *CAC score (AU) | 0.0 (29.0) | 15.4 (166.5) | < 0.001 |
| *Total PUFA (% of total FA) | 41.6 (3.8) | 41.6 (3.7) | 0.963 |
Variables with non-normal distributions are presented as median (interquartile range) and were compared between groups using the Mann–Whitney U test
Abbreviations:CHD coronary heart disease, BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, FPG fasting plasma glucose, TC total cholesterol, HDL-C high-density lipoprotein cholesterol,TG triglycerides, FFAs free fatty acids, CAC coronary artery calcium, PUFA polyunsaturated fatty acids,AU Agatston unit
Supplementary Fig. 2 illustrates the distribution of serum FFAs through a violin plot, stratified by sex. Women generally exhibited higher median levels of FFAs compared to men. Additionally, the shape of the violins suggests that while both sexes exhibited variation in FFA levels, women had a more pronounced tail on the higher end, indicating that women tended to have a wider spread or distribution of FFA levels.
Supplementary Table 1 presents the baseline characteristics of men stratified by quintiles of lipid profiles. Men in higher quintiles of FFAs generally tended to be older and had a higher prevalence of diabetes and hypertension. In contrast, the prevalence of smoking decreased across higher quintiles. BMI, SBP, DBP, and FPG, TG, and CAC scores increased progressively in higher quintiles. Furthermore, the use of antidiabetic and antihypertensive medications was more common in higher quintiles of FFAs. A similar pattern was generally observed for women (Supplementary Table 2).
Comparisons between respondents and non-respondents are presented in Supplemental Table 3. Overall, excluded participants were more likely to be Black, had lower educational attainment, higher prevalence of diabetes and hypertension, and greater use of antidiabetic and antihypertensive medications. They also exhibited higher baseline SBP, FPG, TGs, and higher FFAs and CAC scores.
Outcome incidence
Among 2988 women and 2696 men, a total of 1302 and 1480, respectively, experienced CAC progression, during median follow-ups (interquartile range [IQR]) of 9.0 (4.8–9.5) and 8.89 (4.73–9.41) years. Incidence rates of CAC progression by quintiles of serum FFAs, stratified by sex, are displayed in Supplementary Table 4. In both men and women, the risk of CAC progression increased in the top quintiles of serum FFA levels compared with the lowest quintile.
In a KM survival analysis, men demonstrated significantly lower cumulative survival compared to women (P log-rank < 0.0001, Fig. 1), thereby providing justification for sex-specific modeling in subsequent analyses. Supplementary Fig. 3 presents KM plots stratified by quintiles of FFAs for both men and women. Among both sexes, individuals in higher FFA quintiles exhibited significantly lower cumulative survival for CAC progression compared to those in lower quintiles.
Fig. 1.
Kaplan-Meier survival curve of CAC progression stratified by sex
Association between FFA and CAC progression
The sex-stratified association between serum FFAs and the progression of CAC is presented in Table 2. Among women, continuous FFAs were significantly associated with an increased risk of CAC progression in unadjusted models (per one unit increase HR: 1.93, 95% CI: 1.49–2.50). However, this association was not observed in minimally adjusted models (age and race adjusted) or fully adjusted models, although the direction of the association remained consistent. The quintile analysis among women revealed that, in the unadjusted model, there was a pattern of increasing risk with higher FFAs. Specifically, those in the fifth quintile, compared to those in the first quintile, had an estimated 52% higher risk of CAC progression (1.52: 1.28–1.81). However, these associations became non-significant after adjusted for additional covariates.
Table 2.
Association between serum FFAs and progression of CAC, stratified by sex
| Model 1 | Model 2 | Model 3 | |
|---|---|---|---|
| HR (95% CI) | HR (95% CI) | HR (95% CI) | |
| Women | |||
| Continuous | 1.93 (1.49–2.50) | 1.24 (0.95–1.63) | 1.04 (0.79–1.38) |
| Quintile (Range) | |||
| Quintile 1 (0.15–0.45) | Reference | Reference | Reference |
| Quintile 2 (0.46–0.56) | 1.26 (1.06–1.50) | 1.13 (0.95–1.35) | 1.10 (0.92–1.31) |
| Quintile 3 (0.57–0.66) | 1.22 (1.02–1.45) | 1.03 (0.86–1.23) | 1.01 (0.85–1.21) |
| Quintile 4 (0.67–0.79) | 1.30 (1.09–1.54) | 1.07 (0.90–1.28) | 0.98 (0.82–1.18) |
| Quintile 5 (0.80–2.11) | 1.52 (1.28–1.81) | 1.16 (0.97–1.38) | 1.04 (0.87–1.25) |
| P for trend | < 0.001 | 0.228 | 0.865 |
| Men | |||
| Continuous | 2.15 (1.61–2.88) | 1.61 (1.20–2.17) | 1.41 (1.03–1.93) |
| Quintile (Range) | |||
| Quintile 1 (0.13–0.33) | Reference | Reference | Reference |
| Quintile 2 (0.34–0.41) | 1.16 (0.98–1.37) | 1.13 (0.95–1.33) | 1.11 (0.94–1.31) |
| Quintile 3 (0.42–0.49) | 1.24 (1.06–1.46) | 1.19 (1.02–1.40) | 1.13 (0.96–1.33) |
| Quintile 4 (0.50–0.60) | 1.24 (1.06–1.46) | 1.14 (0.97–1.34) | 1.08 (0.92–1.27) |
| Quintile 5 (0.61–1.78) | 1.51 (1.28–1.77) | 1.31 (1.11–1.54) | 1.20 (1.02–1.43) |
| P for trend | < 0.001 | 0.003 | 0.073 |
Model 1: unadjusted
Model2: adjusted for age and race Model
Model3: adjusted for age, race, education, smoking status, alcohol intake, total physical activity level (continuous), BMI (continuous), diabetes, hypertension, FH CHD, use of lipid-lowering drugs, and continuous forms of TC, HDL-C, TG, and total PUFA
Abbreviations: CHD coronary heart disease, CI confidence interval, BMI body mass index, FH-CHD family history of CHD, HR hazard ratio, TC total cholesterol, HDL-C high-density lipoprotein cholesterol, TG triglycerides, PUFA polyunsaturated fatty acids, FFAs free fatty acids, CAC coronary artery calcium, PUFA polyunsaturated fatty acids
Among men (Table 2), a significant positive association was observed between continuous FFAs and CAC progression, with a HR of 1.41 (1.03–1.93) for each one-unit increase in FFAs, in fully adjusted model. The results for the quintiles of FFAs indicated an increasing trend in risk with higher FFAs. Individuals in the highest quintile of FFAs had a significant 51% higher risk of CAC progression compared to those in the lowest quintile in the unadjusted model. The strength of this association diminished with additional adjustments for confounding factors but remained significant in both the minimally adjusted model and the fully adjusted model, showing a 31% and 20% higher risk of CAC progression in the fifth quintile compared to the first quintile, respectively.
In sensitivity analyses, further adjustment for baseline CAC scores did not substantially alter the association between serum FFAs and CAC progression in both men and women (results not shown).
Test of non-linearity
Figure 2 presents sex-stratified RCS plots of serum FFAs. Among women, we observed a suggestive non-linear association with CAC progression (P for non-linearity = 0.050), with a curve indicating a threshold-like pattern around approximately 0.9 mmol/L. Below this value, the HR was only modestly above 1.0, whereas above 0.9 mmol/L the HR increased more steeply and remained clearly elevated. To formally evaluate the spline-suggested threshold, we conducted an ancillary analysis. The female population was stratified into two groups according to serum FFA levels (< 0.9 and ≥ 0.9 mmol/L), and separate multivariable Cox models were fitted within each stratum with FFAs entered as a continuous variable. In the subgroup with FFAs ≥ 0.9 mmol/L, higher serum FFAs were linearly and significantly associated with an increased risk of CAC progression (HR 3.80, 95% CI 1.26–11.48). By contrast, in women with FFAs < 0.9 mmol/L, FFAs were not significantly associated with CAC progression (HR 0.84, 95% CI 0.57–1.24) (data not shown), suggesting no clear evidence of excess risk at lower FFA levels. These results are in line with the non-linear pattern observed in the spline analysis.
Fig. 2.
Restricted cubic spline plots for coronary artery calcification progression according to the serum free fatty acids in men (A) and women (B). The solid lines represent multivariable-adjusted hazard ratios (HR), with the shaded areas indicating the 95% confidence intervals derived from restricted cubic spline regressions. The horizontal dotted lines represent an HR of 1.0, serving as a reference point. multivariable models were adjusted for age, race, education, smoking status, alcohol intake, total physical activity level, BMI, diabetes, hypertension, FH CHD, use of lipid-lowering drugs, and TC, HDL-C, TG, and total PUFA.We tested models with 3 to 7 knots and selected the optimal number based on the lowest Akaike Information Criterion (AIC) value. Accordingly, Four knots were used for women, while three knots were used for men
Subgroup analysis
The subgroup analysis (Fig. 3) revealed varying HRs for CAC progression across demographic and clinical groups. Individuals aged ≥ 60 years (1.54; 1.07–2.23) and those with BMI < 30 (1.56; 1.09–2.22) exhibited higher risks compared to their counterparts, though no significant interactions were observed (all p-values > 0.05). Similarly, non-hypertensive individuals (1.73; 1.01–2.98) showed higher risks compared to hypertensive groups, but without significant interactions (p-value > 0.05). Overall, while some subgroups had higher risks, the associations were consistent across all subgroups, with no significant modifying effects.
Fig. 3.
Hazard ratios for coronary artery calcification progression in subgroups of men (A) and women (B). Models were adjusted for age, race, education, smoking status, alcohol intake, total physical activity level, BMI, diabetes, hypertension, family history of CHD, use of lipid‑lowering drugs, and continuous measures of TC, HDL‑C, TG, and total PUFA, except for the variable used to define each subgroup. BMI: body mass index; CHD: coronary heart disease; TC: total cholesterol; HDL-C: high‑density lipoprotein cholesterol; TG: triglycerides; PUFA: polyunsaturated fatty acids; HR: hazard ratio; CI: confidence interval
Sensitivity analysis
Results from competing risk regression analyses, treating non-cardiovascular death as a competing event, are presented in Supplementary Table 5. The results remained consistent with the primary Cox regression analyses, albeit with slightly attenuated HRs.
Sensitivity analyses using IPW (Supplementary Table 6) confirmed the patterns observed in the primary Cox regression models. Among men, the positive association between serum FFAs and CAC progression remained significant in the fully adjusted Model 3 (HR per unit increase: 1.41; 1.01–1.98). Quintile-based analyses similarly demonstrated a graded relationship, with risk remaining significantly elevated in the top quintile (1.21; 1.02–1.44).
Supplementary Table 7 presents the associations of serum FFAs with CAC incidence among participants free of CAC at baseline (CAC = 0) and with CAC progression among those with prevalent calcification at baseline (CAC > 0). Overall, no significant associations were observed between FFAs and the risk of initiating new coronary calcification. Among participants with baseline CAC > 0, higher FFAs were associated with CAC progression in unadjusted models (Quintile 5 vs. 1 h: 1.26, 1.08–1.46). However, this association was attenuated and became non-significant after adjusting for cardiovascular risk factors (Model 3 h: 1.11: 0.95–1.30).
Discussion
The present study conducted a sex-stratified analysis of a prospective cohort of U.S. adults. Our findings indicated that higher circulating levels of FFAs were positively associated with an increased risk of CAC progression in men, even after adjusting for known risk factors related to CAC. Among women, we observed evidence of a potential non-linear association, where the risk of CAC progression appeared to increase more sharply above FFA levels of approximately 0.9 mmol/L.
Studies investigating the associations between FFAs and subclinical CVD, such as CAC progression, are limited, particularly in the general population. In a retrospective, single-center study involving 426 patients with CAD, participants were classified into severe CAC (SCAC) and non-SCAC groups. Serum FFAs were independently associated with SCAC across the entire cohort and within subgroups with and without T2DM, with the strongest association observed in the those with T2DM [14]. In contrast, numerous studies have investigated the association between FFAs and clinical CVD outcomes, yielding conflicting results. Some research indicates that elevated FFAs levels are linked to increased mortality and adverse cardiovascular events, particularly in patients with CAD [30] and T2DM [31]. A recent large-scale study in China involving over 10,000 CAD patients found a non-linear U-shaped relationship between FFAs levels and ischemic events, with the lowest risk observed at 500 µmol/L. Subgroup analysis showed that this pattern was observed only in individuals with T2DM, not in non-diabetic CAD patients [32]. Conversely, the study by Nomura et al. [17], which analyzed data from the MESA cohort, found that serum FFAs were not associated with CHD, CVD, or CVD-specific mortality, although a modest association with all-cause mortality was observed, independent of demographic and metabolic factors. The discrepancy between the current study’s findings of a significant relationship between serum FFAs and CAC progression in both men and women and the null associations reported in the Nomura et al. study [17], both utilizing data from the MESA, can be attributed to several potential factors: First, the Nomura et al. study [17] focused on clinical CVD outcomes (e.g., CHD or CVD), while the current study examined subclinical CVD, specifically CAC progression. These represent distinct stages of CVD. FFAs may have a more direct role in early vascular calcification or plaque development, which are processes that precede clinical events. While CAC progression is a strong predictor of future CVD events, the transition from subclinical atherosclerosis to clinical outcomes involves additional pathophysiological mechanisms, such as plaque rupture, thrombosis, and systemic inflammation [33], which may dilute or mask the specific contribution of FFAs over time. Second, CAC progression, as a marker of subclinical disease, may be more sensitive to the early metabolic effects of FFAs, particularly in the context of vascular calcification, which is a gradual process that can be detected earlier than clinical events [34]. Other explanation could be that FFAs contribute to a form of vascular calcification that stabilizes plaques rather than destabilizing them. This concept is somewhat analogous to the “statin paradox,” where statins increase calcification but reduce cardiovascular events by stabilizing plaques [35].
The mechanisms by which FFAs may be associated with CAC progression remain unclear; however, research indicates that FFAs, particularly saturated types, contribute to CAC through their effects on vascular endothelial dysfunction (ED). FFAs increase the production of reactive oxygen species (ROS), which contributes to endothelial cell injury and ED. Moreover, FFAs activate the NF-κB pathway, enhancing pro-inflammatory cytokines that exacerbate ED. Additionally, FFAs can induce apoptosis in endothelial progenitor cells, impairing vascular repair mechanisms. FFAs may also influence the accumulation of lipoproteins in vascular tissues, which is associated with inflammation and calcification [13, 36].
In this study, despite the lower average serum FFAs in men compared to women, FFAs showed a direct linear relationship with the risk of CAC progression in men, but not in women. The higher levels of FFAs observed among women in the present study have been consistently reported in previous research [17, 22]. The sex differences in serum FFAs have been attributed to the higher adiposity generally observed in women compared with men across the life course. This increased fat mass contributes to greater overall FFAs availability due to lipolysis, where stored fat is broken down into FFAs for energy. Additionally, estrogen in women may enhance lipolysis and influence the mobilization of FFAs from adipose tissue into circulation [22]. Furthermore, women typically have lower resting energy expenditure compared to men [37], which might lead to the retention of higher FFA levels in circulation.
The lack of a direct association in this study between FFAs and CAC progression among women, despite their higher levels of FFAs, may be attributed to greater FFAs disposal back into subcutaneous adipose tissue in women compared with men [38]. Women have higher FFAs storage rates per kilogram of adipose tissue, particularly in lower body subcutaneous fat, compared to men. This, along with their greater overall adiposity, suggests that female adipose tissue may act as an FFA “sink” providing some protection against metabolic complications [39]. Moreover, studies have shown that women have increased fat oxidation compared with men [40]. This mechanism may help mitigate the risk of metabolic disorders associated with elevated circulating FFAs levels. In addition, hormonal factors, particularly estrogen, may influence adipose tissue distribution, lipolysis, and systemic FFA metabolism, potentially contributing to sex-specific differences in the relationship between FFAs and vascular calcification. Further research incorporating comprehensive body composition assessments and detailed hormonal profiling is needed to clarify these mechanisms and their role in sex-specific associations between FFAs and CAC progression.
Although we did not find a direct linear association between FFAs and CAC progression in women, further analysis using RCS suggested a potential non-linear relationship (p for non-linearity = 0.050). While this finding is borderline in terms of conventional statistical significance, it raises the possibility of a threshold effect at approximately 0.9 mmol/L for FFAs, above which the risk of CAC progression may increase.
This observed threshold effect in women may be explained by the adipose tissue expandability hypothesis [41], which posits that subcutaneous adipose tissue can safely store excess fatty acids up to a certain capacity, protecting against lipotoxicity. In women, the threshold of approximately 0.9 mmol/L could represent the saturation point of their subcutaneous fat depot, below which FFAs are efficiently buffered and vascular damage is minimized. In contrast, men tend to accumulate visceral fat with lower buffering capacity, leading to a more linear relationship between FFAs and CAC progression due to earlier spillover of FFAs into circulation and vascular tissues. These sex-specific differences in adipose tissue distribution and expandability likely underlie the non-linear association seen in women and highlight a mechanistic pathway linking FFA levels to early vascular calcification. However, given the marginal significance of the non-linearity test, this finding should be interpreted with caution and requires further validation in larger studies.
Our analysis revealed no significant effect modification by age, BMI, diabetes status, or hypertension, suggesting the relationship between FFAs and CAC progression appears independent of common metabolic and demographic factors. However, a notable contrast emerges in a cross-sectional study of 426 CAD patients, where elevated FFA concentrations showed stronger associations with CAC severity (OR: 1.93 vs. 1.27). These findings highlight FFAs as a potential contributor to subclinical atherosclerosis across diverse populations, regardless of common metabolic or demographic factors. Future research should explore whether genetic or lifestyle factors may influence this relationship. Sensitivity analyses showed FFAs were not associated with new CAC development in those with baseline CAC = 0, but a modest trend toward progression was seen in those with baseline CAC > 0. These divergent patterns suggest FFAs may play a larger role in propagating existing plaque than initiating new calcification, consistent with literature distinguishing early metabolic drivers [42] from later mechanical and calcific factors [43]. However, this finding should be interpreted cautiously, as the stratified analysis may have been underpowered due to limited sample size. Although we were unable to perform sex-stratified sensitivity analyses, we observed no significant sex interaction in the pooled data. Collectively, these findings support targeted FFA monitoring in individuals with pre-existing CAC and highlight the need for future studies examining FFA subtypes and detailed plaque composition to improve risk prediction.
Strengths
This study is the first to investigate the relationship between FFA concentrations and the progression of CAC in a large cohort of U.S. adults. We utilized data from the MESA, which employs rigorous and standardized data collection and measurement protocols across multiple locations throughout the United States. A significant strength of the MESA cohort is its diversity in age and race/ethnicity, along with a balanced representation of male and female participants. Furthermore, serum FFAs were measured using a valid and reproducible method within the MESA framework. An additional strength of this study is its comprehensive approach to addressing potential bias from competing events, such as death. The consistency of findings between the primary Cox regression models and the competing risk analyses strengthens confidence in the robustness and validity of our results.
Limitations
Our study has several limitations. First, our study focused on total serum FFAs and their association with CAC progression but did not examine individual FA subtypes, such as omega-3 or omega-6 FAs, which may have distinct effects on cardiovascular risk. Future investigations could clarify whether specific FA species or their ratios modulate CAC development. Second, serum FFAs were assessed only at a single point during the baseline examination. It remains unknown whether repeated measures and an evaluation of changes in FFAs over time would have yielded different associations with CAC progression. Third, residual confounding by unmeasured factors cannot be excluded, including biomarkers of mineral metabolism such as serum phosphate. Emerging evidence suggests that even high-normal phosphate levels may contribute to the development and progression of vascular calcification, independent of chronic kidney disease [44]. Fourth, serum albumin was not measured at baseline in the MESA cohort, precluding direct adjustment for albumin, which may be influenced by inflammation and age-related hypoalbuminemia, a known cardiovascular risk factor [45]. Future studies incorporating both total and unbound FFA measures alongside serum albumin would be valuable to more precisely disentangle these relationships. Fifth, we used a binary outcome to define CAC progression. Although this aligns with established clinical thresholds and facilitates clear risk stratification, it may oversimplify the continuous nature of CAC change and miss nuanced information captured by continuous measures. Modeling such continuous progression is inherently challenging due to non-normal distributions and complex temporal patterns. Our analysis was therefore restricted to the first observed progression event to avoid methodological issues like regression to the mean and sparse data from repeated measures, potentially limiting insights into long-term CAC dynamics. Despite this considerations, the binary definition remains dominant in CAC research for its clinical interpretability and direct linkage to guideline-endorsed risk categories. Future studies should aim to incorporate more sophisticated modeling of continuous CAC progression alongside repeated FFAs measurements to better elucidate the underlying disease mechanisms. Sixth, selection bias is possible due to the exclusion of 1130 participants (16% of the original cohort) who lacked follow-up CAC measurements, FFA levels, or complete covariate data. These excluded participants exhibited a more adverse baseline cardiometabolic profile, suggesting they were at greater risk for CAC progression. Consequently, their exclusion would most likely bias our estimates toward the null, potentially underestimating the true association between FFAs and CAC progression. To address this limitation, we conducted IPW cox regression as a sensitivity analysis. The IPW-weighted estimates were virtually identical to the primary results, suggesting that any bias introduced by participant exclusion is minimal and unlikely to have substantively influenced our findings. Finally, our findings may not be generalizable to adults < 45 years or populations from Europe or other regions with different demographic, genetic, or lifestyle characteristics.
Conclusion
We demonstrate for the first time that serum FFA concentrations are independently associated with the progression of CAC, exhibiting a linear effect in men and a potential non-linear effect in women. Future studies are warranted to elucidate the mechanisms linking FFAs with CAC, and to identify modifiable lifestyle factors and pharmacological intervention that influence serum FFA levels.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We express our gratitude to the MESA Study staff and participants for their invaluable contributions.
Abbreviations
- AA
Times New Roman, serif Arachidonic Acid
- AIC
Akaike Information Criterion
- ALA
Alpha-Linolenic Acid
- ANOVA
Analysis of Variance
- AU
Agatston Units
- BMI
Body Mass Index
- CAC
Coronary Artery Calcification
- CAD
Coronary Artery Disease
- CHD
Coronary Heart Disease
- CVD
Cardiovascular Disease
- CV
Coefficient of Variation
- CT
Computed Tomography
- DBP
Diastolic Blood Pressure
- DGLA
Dihomo-Gamma-Linolenic Acid
- DHA
Docosahexaenoic Acid
- DPA
Docosapentaenoic Acid
- ED
Endothelial Dysfunction
- EDTA
Ethylenediaminetetraacetic Acid
- EPA
Eicosapentaenoic Acid
- FFAs
Free Fatty Acids
- FH CHD
Family History of Coronary Heart Disease
- FPG
Fasting Plasma Glucose
- GLA
Gamma-Linolenic Acid
- HDL-C
High-Density Lipoprotein Cholesterol
- HR
Hazard Ratio
- IPW
Inverse Probability Weighting
- IQR
Interquartile Range
- KM
Kaplan-Meier
- LA
Linoleic Acid
- MESA
Multi-Ethnic Study of Atherosclerosis
- METs
Metabolic Equivalents of Task
- NEFAs
Non-Esterified Fatty Acids
- PAL
Physical Activity Level
- PUFAs
Polyunsaturated Fatty Acids
- RCS
Restricted Cubic Splines
- SBP
Systolic Blood Pressure
- SCAC
Severe Coronary Artery Calcification
- SD
Standard Deviation
- TC
Total Cholesterol
- T2DM
Type 2 Diabetes Mellitus
- TGs
Triglycerides
- ω-3 PUFAs
Omega-3 Polyunsaturated Fatty Acids
- ω-6 PUFAs
Omega-6 Polyunsaturated Fatty Acids
Authors' contributions
FH and AR contributed to conceptualizing the study and its design; AR analyzed the data and wrote the initial manuscript. FE, FH, and AR collaborated on the interpretation of results and the discussion. All authors reviewed the manuscript and provided final approval.
Funding
None declared.
Data availability
The MESA cohort contributes to the National Heart, Lung, and Blood Institute’s (NHLBI) Biologic Specimen and Data Repository (BioLINCC). The data utilized in this study were sourced from the BioLINCC, which can be accessed at https://www.biolincc.nhlbi.nih.gov/.
Declarations
Ethics approval and consent to participate
The institutional review boards at each participating center approved the study protocols, and all participants provided written informed consent at every examination. This study also received ethical approval from the Research Institute for Endocrine Sciences at Shahid Beheshti University of Medical Sciences, Tehran, Iran (IR.SBMU.ENDOCRINE.REC.1404.021).
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Vaduganathan M, Mensah GA, Turco JV, Fuster V, Roth GA. The global burden of cardiovascular diseases and risk: a compass for future health. J Am Coll Cardiol. 2022;80(25):2361–71. [DOI] [PubMed] [Google Scholar]
- 2.Malekmohammad K, Bezsonov EE, Rafieian-Kopaei M. Role of lipid accumulation and inflammation in atherosclerosis: focus on molecular and cellular mechanisms. Front Cardiovasc Med. 2021;8:707529. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Gusev E, Sarapultsev A. Atherosclerosis and inflammation: insights from the theory of general pathological processes. Int J Mol Sci. 2023; 26;24(9):7910. 10.3390/ijms24097910. [DOI] [PMC free article] [PubMed]
- 4.Greenland P, Blaha MJ, Budoff MJ, Erbel R, Watson KE. Coronary calcium score and cardiovascular risk. J Am Coll Cardiol. 2018;72(4):434–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Paixao AR, Chakravorty R, Khera A, Leonard D, DeFina LF, Barlow CE, et al. Disagreement between different definitions of coronary artery calcium progression. JACC Cardiovasc imaging. 2015;8(6):743–4. [DOI] [PubMed] [Google Scholar]
- 6.Budoff MJ, Young R, Lopez VA, Kronmal RA, Nasir K, Blumenthal RS, et al. Progression of coronary calcium and incident coronary heart disease events: MESA (Multi-Ethnic Study of Atherosclerosis). J Am Coll Cardiol. 2013;61(12):1231–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ceponiene I, Nakanishi R, Osawa K, Kanisawa M, Nezarat N, Rahmani S, et al. Coronary artery calcium progression is associated with coronary plaque volume progression: results from a quantitative semiautomated coronary artery plaque analysis. JACC: Cardiovasc Imaging. 2018;11(12):1785–94. [DOI] [PubMed] [Google Scholar]
- 8.Kälsch H, Lehmann N, Moebus S, Hoffmann B, Stang A, Jöckel KH, et al. Aortic calcification onset and progression: association with the development of coronary atherosclerosis. J Am Heart Association. 2017;6(4):e005093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Meng Q, Zhao L, Zhao N, An Y, Lu B, Gao Y. Progression of coronary artery calcium density and major adverse cardiovascular events. Atherosclerosis. 2024;398:118593. [DOI] [PubMed]
- 10.Eghtedari B, Kinninger A, Roy SK, Budoff MJ. Coronary artery calcium progression and all-cause mortality. Coron Artery Dis. 2023;34(4):244–9. [DOI] [PubMed] [Google Scholar]
- 11.Henderson GC. Plasma free fatty acid concentration as a modifiable risk factor for metabolic disease. Nutrients. 2021;13(8). [DOI] [PMC free article] [PubMed]
- 12.Fahy E, Subramaniam S, Brown HA, Glass CK, Merrill AH Jr., Murphy RC, et al. A comprehensive classification system for lipids. J Lipid Res. 2005;46(5):839–61. [DOI] [PubMed] [Google Scholar]
- 13.Ghosh A, Gao L, Thakur A, Siu PM, Lai CWK. Role of free fatty acids in endothelial dysfunction. J Biomed Sci. 2017;24(10):50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Xin Y, Zhang J, Fan Y, Wang C. Serum free fatty acids are associated with severe coronary artery calcification, especially in diabetes: a retrospective study. BMC Cardiovasc Disord. 2021;21(1):343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Huang NK, Bůžková P, Matthan NR, Djoussé L, Hirsch CH, Kizer JR, et al. Associations of serum nonesterified fatty acids with coronary heart disease mortality and nonfatal myocardial infarction: the CHS (cardiovascular health study) cohort. J Am Heart Association. 2021;10(6):e019135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Djousse L, Biggs ML, Matthan NR, Ix JH, Fitzpatrick AL, King I, et al. Serum individual nonesterified fatty acids and risk of heart failure in older adults. Cardiology. 2021;146(3):351–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nomura SO, Karger AB, Weir NL, Duprez DA, Tsai MY. Free fatty acids, cardiovascular disease, and mortality in the multi-ethnic study of atherosclerosis. J Clin Lipidol. 2020;14(4):531–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Pinho-Gomes AC, Peters SA, Thomson B, Woodward M. Sex differences in prevalence, treatment and control of cardiovascular risk factors in England. Heart. 2021;107(6):462–7. [DOI] [PubMed] [Google Scholar]
- 19.Peters SAE, Muntner P, Woodward M. Sex differences in the prevalence of, and trends in, cardiovascular risk factors, treatment, and control in the United States, 2001 to 2016. Circulation. 2019;139(8):1025–35. [DOI] [PubMed] [Google Scholar]
- 20.Beale AL, Meyer P, Marwick TH, Lam CSP, Kaye DM. Sex differences in cardiovascular pathophysiology: why women are overrepresented in heart failure with preserved ejection fraction. Circulation. 2018;138(2):198–205. [DOI] [PubMed] [Google Scholar]
- 21.Holven KB, van Roeters J. Sex differences in lipids: a life course approach. Atheroscler. 2023;384:117270. [DOI] [PubMed] [Google Scholar]
- 22.Zhao L, Hao F, Huang J, Liu X, Ma X, Wang C, et al. Sex-and age-related metabolic characteristics of serum free fatty acids in healthy Chinese adults. J Proteome Res. 2020;19(4):1383–91. [DOI] [PubMed] [Google Scholar]
- 23.Bild DE, Bluemke DA, Burke GL, Detrano R, Diez Roux AV, Folsom AR, et al. Multi-ethnic study of atherosclerosis: objectives and design. Am J Epidemiol. 2002;156(9):871–81. [DOI] [PubMed] [Google Scholar]
- 24.Detrano RC, Anderson M, Nelson J, Wong ND, Carr JJ, McNitt-Gray M, et al. Coronary calcium measurements: effect of CT scanner type and calcium measure on rescan reproducibility—MESA study. Radiology. 2005;236(2):477–84. [DOI] [PubMed] [Google Scholar]
- 25.Berry JD, Liu K, Folsom AR, Lewis CE, Carr JJ, Polak JF, et al. Prevalence and progression of subclinical atherosclerosis in younger adults with low short-term but high lifetime estimated risk for cardiovascular disease: the coronary artery risk development in young adults study and multi-ethnic study of atherosclerosis. Circulation. 2009;119(3):382–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bertoni AG, Whitt-Glover MC, Chung H, Le KY, Barr RG, Mahesh M, et al. The association between physical activity and subclinical atherosclerosis: the multi-ethnic study of atherosclerosis. Am J Epidemiol. 2009;169(4):444–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Cao J, Schwichtenberg KA, Hanson NQ, Tsai MY. Incorporation and clearance of omega-3 fatty acids in erythrocyte membranes and plasma phospholipids. Clin Chem. 2006;52(12):2265–72. [DOI] [PubMed] [Google Scholar]
- 28.Abdelhamid AS, Martin N, Bridges C, Brainard JS, Wang X, Brown TJ, et al. Polyunsaturated fatty acids for the primary and secondary prevention of cardiovascular disease. Cochrane Database Syst Rev. 2018;11(11):Cd012345. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Li L, Hu B, Kattan MW. Modeling potential time to event data with competing risks. Lifetime Data Anal. 2014;20(2):316–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Pilz S, Scharnagl H, Tiran B, Seelhorst U, Wellnitz B, Boehm BO, et al. Free fatty acids are independently associated with all-cause and cardiovascular mortality in subjects with coronary artery disease. J Clin Endocrinol Metabolism. 2006;91(7):2542–7. [DOI] [PubMed] [Google Scholar]
- 31.Hu T, Zhang W, Han F, Zhao R, Liu L, An Z. Plasma fingerprint of free fatty acids and their correlations with the traditional cardiac biomarkers in patients with type 2 diabetes complicated by coronary heart disease. Front Cardiovasc Med. 2022;9:903412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Pan Y, Wu T-T, Mao X-F, Hou X-G, Yang Y, Deng C-J, et al. Decreased free fatty acid levels associated with adverse clinical outcomes in coronary artery disease patients with type 2 diabetes: findings from the practice study. Eur J Prev Cardiol. 2023;30(8):730–9. [DOI] [PubMed] [Google Scholar]
- 33.Henein MY, Vancheri S, Longo G, Vancheri F. The role of inflammation in cardiovascular disease. Int J Mol Sci. 2022;23(21). [DOI] [PMC free article] [PubMed]
- 34.Mendieta G, Pocock S, Mass V, Moreno A, Owen R, García-Lunar I, et al. Determinants of progression and regression of subclinical atherosclerosis over 6 years. J Am Coll Cardiol. 2023;82(22):2069–83. [DOI] [PubMed] [Google Scholar]
- 35.German CA, Liao JK. Understanding the molecular mechanisms of statin pleiotropic effects. Arch Toxicol. 2023;97(6):1529–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Yu F, Zong B, Ji L, Sun P, Jia D, Wang R. Free fatty acids and free fatty acid receptors: role in regulating arterial function. Int J Mol Sci. 2024;25(14):7853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Jagim AR, Jones MT. Sex Differences in resting metabolic rate among athletes and association with body composition parameters: a follow-up investigation. J Funct Morphology Kinesiol. 2023;3(3):109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Gavin KM, Bessesen DH. Sex differences in adipose tissue function. Endocrinol Metab Clin North Am. 2020;49(2):215–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Mundi MS, Koutsari C, Jensen MD. Effects of increased free fatty acid availability on adipose tissue fatty acid storage in men. J Clin Endocrinol Metab. 2014;99(12):E2635–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Beckner ME, Thompson L, Radcliffe PN, Cherian R, Wilson M, Barringer N, et al. Sex differences in body composition and serum metabolome responses to sustained, physical training suggest enhanced fat oxidation in women compared with men. Physiol Genom. 2023;55(5):235–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Virtue S, Vidal-Puig A. Adipose tissue expandability, lipotoxicity and the metabolic syndrome–an allostatic perspective. Biochim Biophys Acta. 2010;1801(3):338–49. [DOI] [PubMed] [Google Scholar]
- 42.Nicoll R, Zhao Y, Ibrahimi P, Olivecrona G, Henein M. Diabetes and hypertension consistently predict the presence and extent of coronary artery calcification in symptomatic patients: a systematic review and meta-analysis. Int J Mol Sci. 2016;17(9). [DOI] [PMC free article] [PubMed]
- 43.Kim KA, Jung HO, Lee SY, Ahn Y, Jung MH, Chung WB, et al. Differences in risk factors associated with the initiation and progression of mitral annular calcification in asymptomatic individuals. Sci Rep. 2025;15(1):16528. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Campos-Obando N, Bosman A. Genetic evidence for a causal role of serum phosphate in coronary artery calcification: the rotterdam study. J Am Heart Association. 2022;11(15):e023024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Tong J, Wang T. Serum Albumin is linearly and negatively associated with the risk of all-cause and cardiovascular death in coronary heart disease patients. Rev Cardiovasc Med. 2025;26(8):38034. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The MESA cohort contributes to the National Heart, Lung, and Blood Institute’s (NHLBI) Biologic Specimen and Data Repository (BioLINCC). The data utilized in this study were sourced from the BioLINCC, which can be accessed at https://www.biolincc.nhlbi.nih.gov/.