Dietary and lifestyle inflammation scores and incidence of cardiovascular disease and heart failure: evidence from a longitudinal cohort

Abstract

Background

We investigated the associations of dietary and lifestyle inflammation scores (DIS and LIS) with the incidence of all cardiovascular disease (CVD) events, its different subtypes, including hard CVD, stroke, myocardial infarction (MI), as well as heart failure (HF).

Methods

We analyzed data from 5866 participants aged ≥ 45 years from the Multi-Ethnic Study of Atherosclerosis (MESA). DIS and LIS were constructed using food frequency and lifestyle questionnaire data. Multivariable Cox regression models were used to estimate the hazard ratios (HRs) and 95% confidence intervals (95% CIs) of DIS and LIS for CVD outcomes.

Results

Over 14 years, there were 822, 600, 290, 252, and 270 events of all CVD, hard CVD, HF, stroke, and MI, respectively. In continuous scale, each one-point increase in DIS was associated with higher risk of all CVD (HR: 1.03; 95% CI: 1.01–1.07), hard CVD (1.05; 1.01–1.08), and MI (1.06; 1.01–1.12). In quintile analysis, the highest DIS quintile was associated with an elevated risk of all CVD (1.28; 1.02–1.61; P trend = 0.067) and hard CVD events (1.47; 1.12–1.92; P trend = 0.013). In subgroup analysis, the impact of DIS on all CVD events was more pronounced among overweight participants compared to normal weight or obese counterparts (P-interaction < 0.001). Furthermore, each one-point increase in LIS was associated with a 36% higher risk of HF (1.36; 1.12–1.62). Participants in the highest LIS quintile had a 75% higher risk of HF (1.75: 1.18–2.59), compared to those in the lowest quintile.

Conclusion

Diet and lifestyle, through their contributions to inflammation, may be associated with higher CVD risk, particularly hard CVD, MI, and HF.

Trial registration

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12937-026-01303-z.

Background

Cardiovascular diseases (CVDs) are the leading cause of death worldwide, contributing to loss of health and imposing significant costs on healthcare systems globally [1, 2]. Systemic and local inflammation are key factors in the development and progression of CVD, from endothelial dysfunction to clinical syndromes [3]. The meta-analysis of 33 observational studies found that high levels of inflammatory biomarkers, including fibrinogen, interleukin-6, high-sensitivity C-reactive protein (hsCRP), and galectin-3, were associated with an increased risk of CVD [4]. Epidemiologic findings indicate that dietary and lifestyle exposures may be associated with CVD risk, likely in part through affecting chronic inflammation [5]. For example, consuming foods rich in antioxidants, fiber, and long-chain n-3 polyunsaturated fatty acids has been associated with reduced levels of pro-inflammatory biomarkers [6, 7]. Beyond single nutrients, adherence to healthy dietary patterns, including the Mediterranean diet, have been associated with lower concentrations of some inflammatory biomarkers [8, 9]. Taken together, the combined impact of multiple dietary components on inflammation and the risk of CVD may be more substantial than the influence of individual nutrients alone [10].

To capture the overall dietary influence on systemic inflammation, indices such as the dietary inflammatory index (DII) [11] and the empirical dietary inflammatory pattern (EDIP) [12] were developed. A meta-analyses of observational studies found that each one-point increase of the DII score was associated with an 8% increased risk of CVD (relative risk = 1.08, 95% confidence interval (95% CI): 1.04–1.12) [13]. However, the DII and EDIP have several limitations. The derivation of EDIP scores requires data on inflammatory markers in the target population, which may not be available in all studies. Additionally, the original DII has several methodological limitations, as noted by its developers [14]. It primarily focuses on individual nutrients without accounting for their complex interactions within the body or the influence of unmeasured bioactive compounds present in whole foods and beverages. Moreover, the inflammatory weights assigned to dietary components were derived from a literature review, which may be limited by the quality, consistency, and generalizability of the underlying studies. Furthermore, certain key anti-inflammatory nutrients, such as flavonoids, known to play an important role in modulating systemic inflammation, were omitted from the original scoring system [14]. Recently, Byrd et al. [10] developed Dietary Inflammation Score (DIS) to address the limitations of previous dietary inflammation scores. Using dietary data from the Reasons for Geographic and Racial Differences in Stroke Study (REGARDS), they categorized intake into 19 predefined food groups encompassing thousands of bioactive compounds, thereby shifting the focus from individual nutrients to whole foods that better reflect actual dietary patterns. Moreover, the weights for each food group were empirically derived based on their associations with a panel of circulating systemic inflammation biomarkers (comprising high-sensitivity C-reactive protein (hsCRP), interleukin-6 (IL-6), interleukin-8 (IL-8), and interleukin-10 (IL-10) in a diverse population [10].

Previous evidence also indicates that inflammatory biomarkers, including hsCRP, IL-6, tumor necrosis factor-alpha (TNF-α), and fibrinogen, are influenced by lifestyle factors, such as physical inactivity, smoking, and excess body weight [5, 15, 16]. In parallel, Byrd et al. [10], introduced the Lifestyle Inflammation Score (LIS), designed to capture the inflammatory effects of non-dietary lifestyle factors. The LIS was constructed from four components, body mass index (BMI), physical activity, smoking status, and alcohol intake, which were selected based on their association with a panel of inflammatory biomarkers, including CRP, IL-6, IL-8, and IL-10. Each component was then assigned a specific weight according to its impact on this biomarker profile [10].

Both the DIS and LIS were originally developed and validated in the REGARDS cohort, a large and racially diverse U.S. population-based study. In the original validation, both scores showed strong and graded associations with inflammatory biomarkers, confirming their criterion validity. Furthermore, in comparative analyses across three independent populations, the DIS proved to have a significantly stronger positive association with inflammatory biomarkers than the DII, a finding that remained robust under various sensitivity analyses [10].

The associations of DIS and LIS with a variety of health outcomes have been investigated [17–21]. However, their association with CVD, its major subtypes, and congestive heart failure (HF) have not yet been examined. Therefore, in the present study, we aimed to evaluate the associations of DIS and LIS with all CVD events, hard CVD events, stroke, myocardial infarction (MI), and HF in middle-late adulthood (ages 45–85) using data from the prospective Multi-Ethnic Study of Atherosclerosis (MESA). The racial and ethnic diversity of MESA, along with its harmonized and rigorously collected dietary and lifestyle data consistent with those used in original score development, makes the application of DIS and LIS in this cohort both methodologically sound and scientifically appropriate.

Methods

The de-identified data used in this study were sourced from The Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC), accessible at https://www.biolincc.nhlbi.nih.gov.

Study design and study population

We used data from MESA, a large-scale cohort study with longitudinal data collection for this study. The study, which began in 2000, has enrolled 6814 men and women aged 45–84 years from six U.S. communities: Baltimore, Maryland; Chicago, Illinois; Forsyth County, North Carolina; Los Angeles County, California; northern Manhattan, New York; and St. Paul, Minnesota. The participants were free of CVD at baseline and did not receive active cancer treatment. Thorough descriptions of the study design and procedures have been previously published [22, 23]. From the 6814 participants at baseline examination (July 2000 and August 2002), we excluded 577 individuals with missing dietary information and 288 individuals with implausible energy intakes (< 600 or > 6000 kcal/d). Of the remaining 5949 participants, we excluded 22 individuals without follow-up data and 61 individuals with missing data on baseline study variables, resulting in a final study population of 5866 (2840 men), who were followed until 2015 for CVD events (Fig. 1). The MESA study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional review boards at each examination site. Written informed consent was obtained from all participants at each examination. Approval for undertaking the current project was also obtained from the Ethics Committee of Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran (approval code: IR.SBMU.ENDOCRINE.REC.1403.056).

Fig. 1.

The flow chart of participants included in the study

This study was conducted in accordance with the STROBE-nutreporting guidelines [24].

Data collection

Covariates

The self-administered and interviewer-based questionnaires were used to gather information on demographics, medical and family history, tobacco use, and alcohol consumption at baseline. The use of medication for diabetes, hypertension, and high cholesterol was assessed through a questionnaire and verified by actual prescriptions. Height and weight were measured and body mass index (BMI) was calculated by dividing weight (in kg) by height (in meters) squared. Three seated resting blood pressure measurements were taken, and the average of the last two was used for analysis. Blood samples were collected to measure fasting plasma glucose (FPG), total cholesterol (TC), and high-density lipoprotein cholesterol (HDL-C) levels using standard protocols. Physical activity was assessed using a semi-quantitative questionnaire adapted from the Cross-Cultural Activity Participation Study [25]. The questionnaire calculated the total minutes per week of intentional activities like walking, sports, dance, and conditioning (e.g., aerobics, bicycling, running, jogging, rowing, swimming, judo, karate). The total minutes were then multiplied by the individual MET (metabolic equivalent of task) value for each activity. Smoking was classified as never, former, or current smoker. Individuals who had smoked at least 100 cigarettes in their lifetime were classified as current or former smokers depending on whether they had smoked cigarettes in the past month. Those who had not smoked were categorized as never-smokers. The alcohol consumption categories included never-drinkers (no lifetime alcohol consumption), current drinkers (currently consuming alcohol), and former drinkers (past alcohol consumers who have stopped). Current drinkers and former drinkers were questioned about the number of drinks per week when drinking. Heavy drinking was characterized as alcohol intake exceeding one drink per day for women or two drinks per day for men, while moderate drinking was defined as an alcohol intake of one drink per day for women or one to two drinks per day for men. Education levels were grouped as (i) high school or lower; (ii) some college, technical school, associate degree, or bachelor’s degree; and (iii) graduate or professional school. A positive family history of coronary heart disease (CHD) was characterized by the occurrence of a heart attack in a parent, sibling, or child. Diabetes was defined as having an FPG level of ≥ 7.0 mmol/l or using insulin or oral hypoglycemic medications. Hypertension was defined as having a systolic blood pressure (SBP) of ≥ 130 mmHg, or a diastolic blood pressure (DBP) of ≥ 80 mmHg, or the use of antihypertensive medications. Hypercholesterolemia was defined as having a TC level of ≥ 6.2 mmol/L or the use of lipid-lowering medication.

Dietary data

Dietary assessments were conducted at baseline between 2000 and 2002 using a modified Block-style 120-item food-frequency questionnaire (FFQ) [26]. The FFQ quantified the consumption frequency and serving size of each food or beverage, with serving sizes categorized as small, medium, or large, and corresponding weights (g) based on NHANES data [27]. Total energy intake was calculated by summing energy from all foods, and the frequencies of alcohol and supplemental micronutrient intakes were also measured.

Constructing the DIS and LIS

In our analysis, we applied the same weighted scoring structure validated in the original cohort [10] without modification. In the original study, weights were derived as regression coefficients from multivariable linear models predicting a composite inflammation biomarker score. Positive weights indicate that higher dietary intake or lifestyle exposure contributes to a more pro-inflammatory profile, whereas negative weights reflect anti-inflammatory effects.

Constructing the DIS

The DIS [10] has 19 components, including various food groups (leafy greens and cruciferous vegetables, legumes, refined grains, and starchy vegetables, apples and berries, deep yellow or orange vegetables and fruit, tomatoes, other fruits, and real fruit juices, other vegetables, added sugars, red and organ meats, processed meats, fish, poultry, high-fat dairy, low-fat dairy and tea, nuts, other fats) and supplemental micronutrients. In the original study [10], the weights for each component were derived from multivariable-adjusted associations with a panel of circulating inflammatory biomarkers, as previously mentioned.

In our study, we applied the same food components as defined in Byrd et al. [10]. To calculate the DIS, we quantified the daily intake (in servings) of each food group from both single-ingredient foods and mixed dishes. Mixed dishes reported in the FFQ were disaggregated into their constituent food groups using the MyPyramid Equivalents Database version 2.0 [28]. For example, pizza was decomposed into refined grains, tomatoes, cheese, and processed meat, with serving sizes estimated for each component. The servings from mixed dishes were then combined with those from single-ingredient foods to derive the total daily intake of each food group for each participant. The total daily intake of each food group was then standardized to a mean of 0 and a standard deviation (SD) of 1.0, while supplemental micronutrients were categorized into sex-specific tertiles. Each standardized or categorized value was then multiplied by its respective weight, and the weighted components were summed to obtain the total DIS [10, 21].

Constructing the LIS

We constructed the LIS from four key components: BMI, physical activity, smoking status, and alcohol intake. Each component was first categorized according to the established protocol [10]: BMI as normal (18.5–24.99 kg/m²), overweight (25–29.99 kg/m²), or obese (≥ 30 kg/m²); physical activity into tertiles of weekly MET-minutes; smoking status as current versus former/never; and alcohol intake as non-drinker, moderate, or heavy drinker. We then assigned corresponding weights to each category. The weighted values were then summed across all components to calculate each participant’s LIS score [10, 21]. For instance, for smoking status, a value of 0 was assigned to former or never smokers, while current smokers received a value of 0.5. Thus, the total score for the smoking component was 0 + 0.5 = 0.5. The same method of calculation was consistently applied across all components.

For both DIS and LIS, a higher score indicates more pro-inflammatory exposure profile.

Outcomes

The MESA study conducts follow-up telephone interviews every 9–12 months to identify clinical events, which are subsequently adjudicated by an independent MESA committee [29].

The outcomes in this study included all CVD, MI, stroke, hard CVD, and HF. MI included all cases of definite and probable MI, based primarily on combinations of symptoms, electrocardiogram (ECG) changes, and cardiac biomarker levels. Definite or probable MI required either abnormal cardiac biomarkers (2 times upper limits of normal) regardless of pain or ECG findings; evolving Q waves regardless of pain or biomarker findings; or a combination of chest pain, and ST-T evolution or new left bundle branch block (LBBB), and biomarker levels 1–2 times upper limits of normal. Stroke was defined as a focal neurologic deficit lasting 24 h or until death, or if < 24 h, there was a clinically relevant lesion on brain imaging and no nonvascular cause. The focal neurological deficits secondary to brain trauma, tumors, infections, or other nonvascular causes were excluded. In this study, incident stroke included all cases of definite and possible stroke and transient ischemic attack (TIA) [29].

All CVD events included the first occurrence of MI, resuscitated cardiac arrest, angina, stroke, and CVD death. Hard CVD events were defined as MI, resuscitated cardiac arrest, stroke, death from CHD, and stroke death. HF was characterized as any event indicating probable or definite HF. Both definite and probable HF required symptoms like shortness of breath or edema. Definite HF additionally rerquired meeting specific criteria such as pulmonary edema on chest radiography, left ventricular dilation or decreased systolic function, or evidence of diastolic dysfunction. Probable HF required a physician diagnosis of HF in the clinical record along with documentation of medical treatment for HF [22, 29].

Descriptive analysis

Participants were categorized into LIS and DIS quintiles at baseline. Their selected characteristics were summarized and compared across score quintiles using the chi-square test for categorical variables and one-way analysis of variance (ANOVA) for continuous variables.

The crude incidence rate of CVD events was calculated by dividing the number of new cases of CVD events by the total person-time at risk. Follow-up time was defined as the time from baseline exam to incident CVD events.

Cox regression analysis

We used Cox proportional hazard regression models to investigate the association between DIS and LIS scores and the incidence of CVD events during follow-up. Both scores were analyzed as continuous variables (per one-unit increment) and as categorical variables (quintiles). Because DIS and LIS have no fixed scale or predetermined minimum and maximum values, a one-unit increase should be interpreted as a relative increment in the exposure score rather than as an absolute change in inflammation level. Continuous analyses were thus conducted to estimate dose-response relationships and to allow comparability with previous studies [13], whereas the categorical analyses were performed to improve interpretability by contrasting relative risks across the exposure distribution, particularly between participants with high versus low scores. In the categorical analyses, the first quintile of each score was used as the reference group.

Assosiation between DIS and outcomes

To evaluate the relationship between DIS and outcomes, Model 1 was adjusted for age, sex, and race (White, Chinese, Black and Hispanic). Model 2 was adjusted for additional covariates including BMI (continuous), educational level (less than high school/ high school, some college, and graduate degree or professional school), family history of CHD, physical activity (tertiles of MET), smoking (never, former and current), alcohol consumption (never, former anf current), and total energy intake (continuous). In Model 3, we included all variables in Model 2, with additional adjustment for HDL-C, TC, lipid-lowering medications, diabetes, and hypertension at baseline. Supplementary Fig. 1 illustrates the conceptual framework of this analysis.

Assosiation between LIS and outcomes

Regarding the associations between the LIS and outcomes, Model 1 was adjusted for age, sex, and race (White, Chinese, Black and Hispanic). Model 2 was further adjusted for education level, family history of CHD, total energy intake, and former smoking status (yes/no), as former smoker status is not included in the LIS [30]. In the final model, we sought to additionally adjust for dietary confounding using the 19 food groups that constitute the DIS. However, including all groups would have substantially increased model complexity. Therefore, we instead adjusted for a single composite variable: an equally weighted DIS [20, 30]. This score was calculated by assigning a value of + 1 to pro-inflammatory components and − 1 to anti-inflammatory components. Thus, Model 3 included all covariates from Model 2, with additional adjustment for this equally weighted DIS, HDL-C, TC, lipid-lowering medications, diabetes, and hypertension. The conceptual framework for this analysis is presented in Supplementary Fig. 2.

We assessed the proportionality assumptions of the Cox proportional hazards model using Schoenfeld’s global test of residuals and confirmed that the assumptions were met. To test the linear trend with CVD outcomes, the median of DIS and LIS quintiles were included in the multivariable Cox models (Model 3) as a continuous variable.

Subgroup analysis

In subgroup analysis, we investigated the potential effect modification by age, sex, race, BMI category, diabetes, and hypertension on the associations between the DIS and incident CVD by including interaction terms (DIS × subgroup variable) in multivariable Cox models and calculating P-values for interaction using the likelihood ratio test to compare the fully adjusted models with and without the interaction term. In these models, DIS was categorized into quintiles, and confounders from the primary analysis (Model 3) were included, excluding the subgroup variable under consideration.

For the LIS, we did not investigate effect modification by BMI because it is a component of the score. Therefore, subgroup analyses were stratified by age, sex, race, diabetes, and hypertension only. The LIS was also modeled as quintiles in multivariable Cox models, adjusting for the same confounders as in primary model 3, but excluding the variable used for stratification in each subgroup analysis.

To account for multiple testing across subgroups, the Bonferroni correction was applied, setting the significance threshold at a P-value of 0.008 (0.05 divided by 6).

Sensitivity analysis

In an sensitivity analysis, we repeated the primary analysis for all CVD after excluding participants with baseline diabetes, hypertension, or hypercholesterolemia. Furthermore, we investigated the associations of each individual LIS component with study outcomes. Finally, to further examine the relationship between DIS and LIS and systemic inflammation, we conducted separate linear regression analyses for hsCRP and IL-6. Baseline measurements were available for 5839 participants for hsCRP and 5718 participants for IL-6, out of a total of 5866 study participants. Linear regression models were used to estimate the change in inflammatory marker levels associated with each unit increase in DIS or LIS. To account for potential confounding, we fitted three progressively adjusted models incorporating confounders consistent with those used in the Cox proportional hazards analyses.

All analyses were performed using R version 4.2.1. All statistical tests were two-sided, and significance was determined at a threshold of P < 0.05.

Results

Incidence rates and baseline characteristics

Over a median follow-up of 14.1 years (interquartile range [IQR]: 13.5–14.7), 822 of the 5866 participants developed all CVD events. The incidence rate (95% CI) of all CVD development was 11.7 (11.0-12.6) per 1000 person-years. The number of events for other study outcomes was as follows: 600 hard CVD, 290 HF, 252 stroke, and 270 MI. Incidence rates for all study outcomes by sex, as well as by DIS and LIS, are shown in Supplementary Tables 1 and 2, respectively.

The baseline characteristics of individuals by DIS and LIS quintiles are presented in Table 1. Individuals in the highest quintiles of DIS were more likely to be Black, have lower levels of formal education, be current smokers, former drinkers, and were less likely to use lipid-lowering medication compared to those in the lowest quintiles of DIS. Individuals in the highest quintiles of DIS were also younger and had lower MET value, HDL-C, total carbohydrate, calcium, and protein intake, but had higher BMI, DBP, total energy, and fat intake compared to those in the lowest quintiles of DIS.

Table 1.

Baseline characteristics of participants (n=5866) by quintiles of DIS and LIS

	DISa				LISa
	Q1 (n = 1173)	Q3 (n = 1174)	Q5 (n = 1173)	P Value	Q1 (n = 1505)	Q3 (n = 1332)	Q5 (n = 1329)	P Value
Score range	(-12.9, -1.82)	(-0.46, 0.64)	(1.82, 13.5)	-	(-1.07,0.00)	(0.50, 0.89)	(1.57, 2.67)	-
Age (year) (mean, SD)	(63.6, 10.1)	(62.5, 10.4)	(60.2, 10.2)	< 0.001	(63.1, 10.7)	(63.1, 10.4)	(61.1, 9.7)	< 0.001
Men (%)	49.3	46.1	47.8	0.070	48.3	48.8	37.2	< 0.001
Race/ethnicity (%)				< 0.001				< 0.001
White	53.8	39.1	35.5		43.7	35.6	33.1
Chinese	8.8	14.9	4.8		29.2	12.5	1.3
Black	19.4	22.7	38.2		14.0	24.7	35.6
Hispanics	18.0	23.3	21.6		13.1	27.2	30.0
Education (%)				< 0.001				< 0.001
Less than high school/ High school	22.9	38.2	41.4		30.0	45.3	43.1
Some college	47.4	45.2	47.6		46.8	41.0	44.5
Graduate degree or professional school	29.7	16.6	11.0		23.3	13.7	12.4
Smoking (%)				< 0.001				< 0.001
Never	50.8	52.6	43.4		60.9	53.8	47.8
Former	42.6	37.1	34.4		37.3	35.5	34.5
Current	6.6	10.2	22.2		1.8	10.7	17.8
Alcohol consumption (%)				< 0.001				< 0.001
Never	16.3	21.1	16.5		24.5	22.6	19.6
Former	21.4	21.8	27.2		20.2	25.2	29.6
Current	62.3	57.1	56.3		55.3	52.3	50.8
bPhysical activity (MET.min/wk) (mean, SD)	(785.3, 1634.9)	(415.2, 1071.1)	(288.8, 864.8)	< 0.001	(615.4, 1486.4)	(32.3, 263.5)	(83.1, 570.1)	< 0.001
BMI (kg/m2) (mean, SD)	(27.5, 4.9)	(27.9, 5.1)	(29.6, 5.8)	< 0.001	(22.8, 1.9)	(26.9, 2.3)	(34.4, 4.3)	< 0.001
SBP (mm Hg) (mean, SD)	(125.9, 20.6)	(126.2, 21.3)	(125.8, 21.1)	0.619	(121.8, 21.9)	(127.4, 21.1)	(130.3, 20.4)	< 0.001
DBP (mm Hg) (mean, SD)	(71.4, 10.2)	(71.5, 9.8)	(72.5, 10.3)	< 0.001	(70.0, 10.1)	(72.4, 10.4)	(72.4, 10.1)	< 0.001
HDL-C (mmol/L) (mean, SD)	(1.7, 0.3)	(1.3, 0.3)	(1.2, 0.3)	< 0.001	(1.4, 0.4)	(1.2, 0.3)	(1.2, 0.3)	< 0.001
TC (mmol/L) (mean, SD)	(4.9, 0.8)	(5.0, 0.8)	(5.0, 0.9)	0.227	(5.0, 0.8)	(5.0, 0.9)	(5.0, 0.9)	0.267
FPG (mmol/L) (mean, SD)	(5.3, 1.6)	(5.4, 1.6)	(5.4, 1.7)	0.271	(5.1, 1.6)	(5.3, 1.4)	(5.8, 2.0)	< 0.001
Positive family history of CHD (%)	43.9	37.5	41.3	0.010	33.8	37.3	45.4	< 0.001
Diabetes (%)	10.5	12.7	12.2	0.263	6.9	11.6	19.9	< 0.001
Hypertension (%)	58.3	59.0	57.5	0.505	47.3	62.3	68.5	< 0.001
Hypercholesterolemia (%)	25.8	25.1	22.0	0.119	21.1	26.0	27.6	0.001
Lipid-lowering medication (%)	18.0	16.1	12.3	< 0.001	12.2	17.4	18.4	< 0.001
Antihypertensive medications (%)	36.3	37.0	34.8	0.338	25.5	37.1	47.3	< 0.001
Insulin or oral hypoglycemic medications (%)	7.7	9.0	7.8	0.235	4.7	8.0	13.5	< 0.001
Energy (Kcal/day) (mean, SD)	(1588.1, 738.3)	(1439.3, 656,1)	(1927.1, 907.1)	< 0.001	(1415.3, 654.1)	(1551.5, 765.9)	(1707.1, 844.9)	< 0.001
Total protein (%Kcal/day) (mean, SD)	(16.8, 3.4)	(16.1, 3.1)	(14.3, 2.9)	< 0.001	(16.2, 3.4)	(15.8, 3.2)	(15.5, 3.2)	< 0.001
Total carbohydrate (%Kcal/day) (mean, SD)	(55.2, 9.0)	(52.7, 8.6)	(49.4, 8.3)	< 0.001	(54.3, 8.6)	(53.4, 8.5)	(51.3, 8.8)	< 0.001
Total fat (%Kcal/day) (mean, SD)	(28.3, 6.6)	(30.6, 6.4)	(35.6, 6.5)	< 0.001	(29.7, 6.8)	(31.1, 6.8)	(32.9, 6.9)	< 0.001
Total calcium (mg/day) (mean, SD)	(1520.2, 4366.4)	(1306.8, 6562.0)	(976.8, 1925.3)	< 0.001	(1349.2, 6522.0)	(1094.4, 2206.6)	(1087.2, 1405.3)	0.251

DIS Dietary inflammation score, LIS Lifestyle inflammation score, MET Metabolic equivalent of task, BMI Body mass index, SBP Systolic blood pressure, DBP Diastolic blood pressure, HDL-C High-density lipoprotein cholesterol, TC Total cholesterol, FPG Fasting plasma glucose, CHD Coronary heart disease, SD Standard deviation

^aParticipants were categorized into LIS and DIS quartiles at baseline

^bConditioning physical activity such as walking, sports, dance, and conditioning (e.g. aerobics, bicycling, running, jogging, rowing, swimming, judo, karate)

An analysis of the baseline characteristics of participants based on their LIS quintiles revealed that those in the highest quintiles were younger, less likely to be men, and had lower levels of formal education. Additionally, they had a higher prevalence of positive family history of CHD and were more likely to be current smokers, former drinkers, and have diabetes, hypercholesterolemia and hypertension. Individuals in the top LIS quintiles were more likely to use lipid-lowering, antihypertensive, and hypoglycemic medications. Furthermore, they demonstrated elevated levels of BMI, SBP, DBP, FPG, total energy, and fat intake, while displaying lower levels of MET value, HDL-C, total protein, and carbohydrate intake.

Supplementary Table 3 displays the dietary inflammatory components among participants categorized by quintiles of DIS. Generally, participants in the highest DIS quintiles had lower levels of all dietary components except for red and organ meats, processed meats, added sugars, high-fat dairy, other fats, refined grains and starchy vegetables.

The associations of the DIS with CVD events

A positive assosiation was observed between the DIS and incidence of all CVD, with each one-point increase in DIS corresponding to a 3% higher risk (HR: 1.03; 95% CI: 1.01–1.07) in the fully adjusted model. When analyzed categorically, individuals in the highest DIS quintile (Q5) exhibited a significantly higher risk of all CVD compared to those in the lowest quintile (Q1) in the fully adjusted model (1.28; 1.02–1.61; P-trend = 0.067) (Table 2). The DIS was also positively associated with the risk of hard CVD and MI when analyzed as a continuous variable. Each one-unit increase in DIS corresponded to a 5% higher risk of hard CVD (1.05; 1.01–1.08) and a 6% higher risk of MI (1.06; 1.01–1.12). When examined categorically, participants in the highest DIS quintile (Q5) had higher risks of hard CVD (1.47; 1.12–1.92; P-trend = 0.013) and MI (1.43; 0.98–2.09; P-trend = 0.079), compared to those in lowest quintile (Q1) (Tables 3 and 4, respectively). DIS was not associated with risk of CHF or stroke (Tables 5 and 6).

Table 2.

Associations of DIS and LIS with the risk of all CVD

		DIS				LIS
		Model 1 HR (95% CI)	Model 2a HR (95% CI)	Model 3b HR (95% CI)		Model 1 HR (95% CI)	Model 2c HR (95% CI)	Model 3d HR (95% CI)
Score as continuous		1.05 (1.02–1.08)	1.03 (1.01–1.06)	1.03 (1.01–1.07)		1.29 (1.16–1.43)	1.26 (1.13–1.40)	1.08 (0.96–1.20)
Quintiles	No. cases/event				No. cases/event
Q1	1173/162	Reference	Reference	Reference	1505/179	Reference	Reference	Reference
Q2	1173/162	1.09 (0.87–1.35)	1.05 (0.85–1.31)	1.02 (0.82–1.27)	835/111	0.99 (0.78–1.26)	0.97 (0.76–1.24)	0.87 (0.68–1.11)
Q3	1174/167	1.14 (0.92–1.42)	1.10 (0.89–1.38)	1.08 (0.86–1.34)	1332/201	1.27 (1.03–1.56)	1.24 (1.01–1.53)	1.03 (0.84–1.28)
Q4	1173/151	1.10 (0.88–1.38)	1.02 (0.81–1.29)	0.99 (0.79–1.25)	865/136	1.56 (1.23–1.97)	1.51 (1.20–1.91)	1.20 (0.95–1.53)
Q5	1173/180	1.44 (1.16–1.79)	1.28 (1.02–1.61)	1.28 (1.02–1.61)	1329/195	1.44 (1.16–1.79)	1.38 (1.11–1.71)	1.01 (0.80–1.26)
P for trend				0.067				0.354

Model 1: adjusted for age (continuous), sex, and race (White, Chinese, Black, Hispanic)

^aModel 2: adjusted for model 1 and education (less than high school/ high school, some college, graduate degree or professional school), smoking (never, former, current), alcohol consumption (never, former, current), physical activity level (as tertiles of MET), BMI (continuous), family history of CHD (yes/no) and total energy intake (continuous)

^bModel 3: adjusted for model 2 and HDL-C (continuous), TC (continuous), lipid-lowering medication (yes/no), hypertension (yes/no), and diabetes (yes/no)

^cModel 2: adjusted for model 1 and education, former smoker (yes/no), family history of CHD, and total energy intake

^dModel 3: adjusted for model 2^‡ and equally weighted DIS (continuous), HDL-C, TC, lipid-lowering medication, hypertension, and diabetes

CVD Cardiovascular diseases, DIS Dietary inflammation score, LIS Lifestyle inflammation score, MET Metabolic equivalent of task, BMI Body mass index, HDL-C High-density lipoprotein cholesterol, TC Total cholesterol, CHD Coronary heart disease, HR Hazard ratio, CI Confidence interval

Table 3.

Associations of DIS and LIS with the risk of hard CVD

		DIS				LIS
		Model 1 HR (95% CI)	Model 2a HR (95% CI)	Model 3b HR (95% CI)		Model 1 HR (95% CI)	Model 2c HR (95% CI)	Model 3d HR (95% CI)
Score as continuous		1.06 (1.02–1.10)	1.05 (1.01–1.09)	1.05 (1.01–1.08)		1.26 (1.11–1.42)	1.23 (1.09–1.40)	1.06 (0.93–1.21)
Quintiles	No. cases/event				No. cases/event
Q1	1173/106	Reference	Reference	Reference	1505/128	Reference	Reference	Reference
Q2	1173/123	1.26 (0.97–1.63)	1.22 (0.94–1.59)	1.18 (0.90–1.53)	835/83	1.01 (0.76–1.34)	0.99 (0.75–1.31)	0.90 (0.68–1.20)
Q3	1174/125	1.32 (1.02–1.71)	1.29 (0.99–1.68)	1.26 (0.96–1.63)	1332/145	1.21 (0.95–1.55)	1.19 (0.93–1.52)	1.01 (0.78–1.29)
Q4	1173/114	1.28 (0.98–1.68)	1.20 (0.92–1.58)	1.17 (0.89–1.53)	865/102	1.55 (1.18–2.04)	1.51 (1.15–1.99)	1.21 (0.91–1.60)
Q5	1173/132	1.62 (1.25–2.11)	1.48 (1.13–1.94)	1.47 (1.12–1.92)	1329/142	1.35 (1.05–1.74)	1.30 (1.01–1.68)	0.96 (0.74–1.26)
P for trend				0.013				0.670

Model 1: adjusted for age (continuous), sex, and race (White, Chinese, Black, Hispanic)

^aModel 2: adjusted for model 1 and education (less than high school/ high school, some college, graduate degree or professional school), smoking (never, former, current), alcohol consumption (never, former, current), physical activity level (as tertiles of MET), BMI (continuous), family history of CHD (yes/no) and total energy intake (continuous)

^bModel 3: adjusted for model 2 and HDL-C (continuous), TC (continuous), lipid-lowering medication (yes/no), hypertension (yes/no), and diabetes (yes/no)

^cModel 2: adjusted for model 1 and education, former smoker (yes/no), family history of CHD, and total energy intake

^dModel 3: adjusted for model 2^‡ and equally weighted DIS (continuous), HDL-C, TC, lipid-lowering medication, hypertension, and diabetes

CVD Cardiovascular diseases, DIS Dietary inflammation score, LIS lifestyle inflammation score, MET Metabolic equivalent of task, BMI Body mass index, HDL-C High-density lipoprotein cholesterol, TC Total cholesterol, CHD Coronary heart disease, HR Hazard ratio, CI Confidence interval

Table 4.

Associations between DIS and LIS and the risk of MI

		DIS				LIS
		Model 1 HR (95% CI)	Model 2a HR (95% CI)	Model 3b HR (95% CI)		Model 1 HR (95% CI)	Model 2c HR (95% CI)	Model 3d HR (95% CI)
Score as continuous		1.07 (1.02–1.13)	1.05 (1.01–1.11)	1.06 (1.01–1.12)		1.22 (1.02–1.46)	1.16 (0.97–1.40)	1.02 (0.84–1.23)
Quintiles	No. cases/event				No. cases/event
Q1	1173/54	Reference	Reference	Reference	1505/55	Reference	Reference	Reference
Q2	1173/48	0.98 (0.66–1.45)	0.96 (0.65–1.43)	0.93 (0.63–1.38)	835/35	0.93 (0.61–1.44)	0.91 (0.59–1.40)	0.86 (0.55–1.32)
Q3	1174/54	1.14 (0.78–1.67)	1.14 (0.78–1.68)	1.12 (0.76–1.64)	1332/73	1.42 (0.99–2.02)	1.36 (0.95–1.95)	1.18 (0.82–1.70)
Q4	1173/48	1.07 (0.72–1.59)	1.02 (0.68–1.52)	1.00 (0.67–1.49)	865/50	1.64 (1.10–2.44)	1.55 (1.04–2.31)	1.29 (0.86–1.94)
Q5	1173/66	1.59 (1.10–2.29)	1.43 (0.97–2.10)	1.43 (0.98–2.09)*	1329/57	1.25 (0.85–1.84)	1.15 (0.78–1.70)	0.88 (0.58–1.32)
P for trend				0.079				0.958

Model 1: adjusted for age (continuous), sex, and race (White, Chinese, Black, Hispanic)

^aModel 2: adjusted for model 1 and education (less than high school/ high school, some college, graduate degree or professional school), smoking (never, former, current), alcohol consumption (never, former, current), physical activity level (tertiles of MET), BMI (continuous), family history of CHD (yes/no) and total energy intake (continuous)

^bModel 3: adjusted for model 2 and HDL-C (continuous), TC (continuous), lipid-lowering medication (yes/no), hypertension (yes/no), and diabetes (yes/no)

^cModel 2: adjusted for model 1 and education, former smoker (yes/no), family history of CHD, and total energy intake

^dModel 3: adjusted for model 2^‡ and equally weighted DIS (continuous), HDL-C, TC, lipid-lowering medication, hypertension, and diabetes

MI Myocardial infarction, DIS Dietary inflammation score, LIS Lifestyle inflammation score, MET Metabolic equivalent of task, BMI Body mass index, HDL-C High-density lipoprotein cholesterol, TC Total cholesterol, CHD Coronary heart disease, HR Hazard ratio, CI Confidence interval

^*Pvalue = 0.065

Table 5.

Associations between DIS and LIS and the risk of CHF

		DIS				LIS
		Model 1 HR (95% CI)	Model 2a HR (95% CI)	Model 3b HR (95% CI)		Model 1 HR (95% CI)	Model 2c HR (95% CI)	Model 3d HR (95% CI)
Score as continuous		1.01 (0.97–1.06)	0.98 (0.93–1.02)	0.98 (0.93–1.03)		1.59 (1.32–1.90)	1.54 (1.28–1.85)	1.36 (1.12–1.65)
Quintiles	No. cases/event				No. cases/event
Q1	1173/60	Reference	Reference	Reference	1505/47	Reference	Reference	Reference
Q2	1173/48	0.87 (0.59–1.28)	0.86 (0.58–1.26)	0.83 (0.57–1.22)	835/45	1.47 (0.97–2.23)	1.49 (0.99–2.27)	1.36 (0.89–2.07)
Q3	1174/67	1.28 (0.90–1.83)	1.25 (0.88–1.79)	1.24 (0.87–1.77)	1332/60	1.37 (0.93–2.02)	1.33 (0.90–1.97)	1.18 (0.80–1.76)
Q4	1173/58	1.21 (0.84–1.74)	1.12 (0.77–1.63)	1.15 (0.79–1.66)	865/51	2.17 (1.44–3.28)	2.11 (1.39–3.19)	1.73 (1.13–2.64)
Q5	1173/57	1.22 (0.84–1.77)	0.93 (0.63–1.37)	0.95 (0.65–1.40)	1329/87	2.40 (1.65–3.49)	2.27 (1.56–3.32)	1.75 (1.18–2.59)
P for trend				0.616				0.003

Model 1: adjusted for age (continuous), sex, and race (White, Chinese, Black, Hispanic)

^aModel 2: adjusted for model 1 and education (less than high school/ high school, some college, graduate degree or professional school), smoking (never, former, current), alcohol consumption (never, former, current), physical activity level (as tertiles of MET), BMI (continuous), family history of CHD (yes/no) and total energy intake (continuous)

^bModel 3: adjusted for model 2 and HDL-C (continuous), TC (continuous), lipid-lowering medication (yes/no), hypertension (yes/no), and diabetes (yes/no)

^cModel 2: adjusted for model 1 and education, former smoker (yes/no), family history of CHD, and total energy intake

^dModel 3: adjusted for model 2^‡ and equally weighted DIS (continuous), HDL-C, TC, lipid-lowering medication, hypertension, and diabetes

CHF Congestive heart failure, DIS Dietary inflammation score, LIS Lifestyle inflammation score, MET Metabolic equivalent of task, BMI Body mass index, HDL-C High-density lipoprotein cholesterol, TC Total cholesterol, CHD Coronary heart disease, HR Hazard ratio, CI Confidence interval

Table 6.

Associations between DIS and LIS and the risk of stroke

		DIS				LIS
		Model 1 HR (95% CI)	Model 2a HR (95% CI)	Model 3b HR (95% CI)		Model 1 HR (95% CI)	Model 2c HR (95% CI)	Model 3d HR (95% CI)
Score as continuous		1.03 (0.98–1.09)	1.03 (0.97–1.09)	1.02 (0.97–1.08)		1.19 (0.99–1.44)	1.20 (0.99–1.45)	1.03 (0.84–1.26)
Quintiles	No. cases/event				No. cases/event
Q1	1173/44	Reference	Reference	Reference	1505/56	Reference	Reference	Reference
Q2	1173/58	1.39 (0.94–2.07)	1.34 (0.90–1.99)	1.29 (0.87–1.92)	835/40	1.12 (0.74–1.70)	1.13 (0.74–1.70)	1.02 (0.67–1.56)
Q3	1174/59	1.47 (0.99–2.19)	1.40 (0.94–2.09)	1.35 (0.91–2.02)	1332/55	1.00 (0.68–1.46)	1.00 (0.68–1.46)	0.84 (0.57–1.24)
Q4	1173/43	1.15 (0.75–1.76)	1.07 (0.70–1.65)	1.03 (0.67–1.59)	865/37	1.25 (0.81–1.93)	1.26 (0.82–1.95)	1.00 (0.64–1.56)
Q5	1173/48	1.39 (0.91–2.10)	1.31 (0.85–2.01)	1.28 (0.83–1.97)	1329/64	1.28 (0.88–1.88)	1.29 (0.88–1.89)	0.96 (0.64–1.43)
P for trend				0.575				0.845

Model 1: adjusted for age (continuous), sex, and race (White, Chinese, Black, Hispanic)

^aModel 2: adjusted for model 1 and education (less than high school/ high school, some college, graduate degree or professional school), smoking (never, former, current), alcohol consumption (never, former, current), physical activity level (as tertiles of MET), BMI (continuous), family history of CHD (yes/no) and total energy intake (continuous)

^bModel 3: adjusted for model 2 and HDL-C (continuous), TC (continuous), lipid-lowering medication (yes/no), hypertension (yes/no), and diabetes (yes/no)

^cModel 2: adjusted for model 1 and education, former smoker (yes/no), family history of CHD, and total energy intake

^dModel 3: adjusted for model 2^‡ and equally weighted DIS (continuous), HDL-C, TC, lipid-lowering medication, hypertension, and diabetes

CVD Cardiovascular diseases, DIS Dietary inflammation score, LIS Lifestyle inflammation score, MET Metabolic equivalent of task, BMI Body mass index, HDL-C High-density lipoprotein cholesterol, TC Total cholesterol, CHD Coronary heart disease, HR Hazard ratio, CI Confidence interval

The associations of the LIS with CVD events

The LIS was found to be positively associated with the risk of HF, but not with other outcomes. In the continuous form, each one-point increase in LIS was associated with a 36% higher risk (1.36; 1.14–1.62) for HF. In categorical form, individuals in the highest quintile of the LIS had a 75% (1.75; 1.18–2.59; P-trend = 0.003) higher risk of HF compared to those in the lowest quintile (Table 4).

Subgroup analyses

In subgroup analyses, there were no significant variations in the association between the LIS and all CVD across categories of age, sex, race, BMI, diabetes, and hypertension. However, the association between the DIS and all CVD was stronger among individuals who were overweight (1.34; 0.99–1.88), compared to those with normal weight (1.29; 0.83–2.01) or with obesity (1.30; 0.88–1.90) (P interaction < 0.001) (Supplementary Table 4).

Sensitivity analysis

In the sensitivity analysis, excluding participants with baseline diabetes, hypertension, or hypercholesterolemia (n = 1829; 127 all CVD events), higher DIS remained significantly associated with increased all CVD risk (HR per 1-unit increase = 1.15; 95% CI: 1.06–1.25).

For the LIS, higher scores were associated with a 31% higher risk of all CVD (1.31; 0.99–1.73; P = 0.056). Overall, the results of the sensitivity analysis were consistent with the primary analysis.

The associations for individual LIS components (Supplementary Table 5) were more pronounced than those observed for the overall LIS. Specifically, moderate and heavy drinkers, compared to non-drinkers, showed a lower risk of all CVD with HRs of 0.65 (0.45–0.94) and 0.73 (0.55–0.96), respectively. In contrast, current smokers had a 73% higher risk (1.73; 1.38–2.17) for all CVD compared with never smokers.

The results examining the relationship of DIS and LIS with inflammatory markers demonstrated a positive and graded association, with higher DIS and LIS scores corresponding to elevated levels of hsCRP and IL-6 (Supplementary Table 7).

Discussion

Using data from a well-known American multi-ethnic cohort with participants free of CVD at the baseline, we found that the highest quintile of DIS compared to the lowest, was associated with about 30% higher risk of all CVD events. Additionally, a more pro-inflammatory diet was associated with an increased risk of hard CVD events including MI, stroke, and CV death. Interestingly, we found that BMI status modified the association between DIS and CVD events, with overweight individuals being more susceptible to the negative effects of a pro-inflammatory diet. Our analysis did not find a significant association between LIS and CVD outcomes excluding CHF. Individuals with the most pro-inflammatory lifestyle had a 75% higher risk of CHF compared to those with the least pro-inflammatory lifestyle (quintile 5 vs. quintile 1 of LIS).

Although many original studies as well as meta-analyses have examined the impact of DII on CVD events, to our knowledge, no previous studies have examined the impact of DIS on these outcomes. However, two previous studies reported the impact of DIS on mortality outcomes, including CVD mortality [17, 20]. In the Iowa Women’s Health Study (IWHS), researchers discovered that the highest quintile of DIS was associated with a 12% higher risk of CVD mortality compared to the lowest quintile. Additionally, each point increase in DIS was associated with a 3% increased risk of CVD mortality [20]. In the REGARDS Study, researchers discovered that compared to the first quintile, the fifth quintile of DIS was associated with an 83% higher risk of all CVD mortality, but this association was nullified in the multivariate analysis [17]. We found about 50% higher risk for hard CVD events among individuals in the highest quintile compared with those in the lowest quintile of dis, independent of a large set of covariates, such as BMI and known CVD risk factors (P trend = 0.013). Additionally, the REGARDS study found that the association between DIS and all-CVD mortality was stronger among women (P interaction = 0.07). However, in our study, we did not find an effect modification for gender in the association between DIS and all CVD events (P interaction = 0.274). Notably, Byrd et al. previously demonstrated that the association between DIS and high-sensitivity C-Reactive Protein (hs-CRP) tended to be stronger among normal and overweight participants compared to their obese counterparts, a finding that was consistent across other dietary inflammatory indices, such as DII and EDIP [10]. Similarly, in this prospective study, we found that overweight participants were at the highest risk of CVD compared with normal and obese individuals, even though obese individuals had a higher DIS score (data not shown). This finding suggests that the association between obesity/overweight and CVD risk may not be fully explained by inflammation alone. It is also possible that variations in fat distribution and other unmeasured factors play a role in the increased risk of CVD in the overweight group.

A meta-analysis of 11 prospective cohorts found that the highest compared to the lowest category of Dietary Inflammatory Index (DII) had a 36% (1.19–1.57) higher risk for CVD events, with significant heterogeneity among the included studies. The risk was 8% higher per one-score increase in DII, with a higher risk among women than men [13]. Similarly, another meta-analysis of 15 prospective cohorts showed a 41% higher CVD risk for the highest compared to the lowest DII category, without obvious heterogeneity [31]. Additionally, a dose-response meta-analysis demonstrated that the most pro-inflammatory diet (highest DII score) compared to the most anti-inflammatory diet (lowest DII score) increased the risk of CVD mortality by 30%, with a 4% increase per one-score increment in DII [32]. However, A recent umbrella review of meta-analyses found convincing evidence for the association between DII and MI, but weak evidence for the association between DII and CVD risk, CVD mortality, and no evidence for stroke and CHD risk [33]. In the current study, we also found that one score increase in DIS was associated with a 6% higher risk for MI. Although a few studies found an association between dietary inflammatory potential and prevalent heart failure (HF), these studies were limited by their cross-sectional design or lack of adjustments for potential HF risk factors [34–36]. As the first prospective study, we did not confirm the role of diets with higher pro-inflammatory potential in the development of HF. Numerous studies support the role of individual LIS components, including BMI, smoking status, physical activity, and alcohol consumption, in the development of CVD and HF [37, 38]. Therefore, it is speculated that the cumulative impact of LIS components might be stronger than the individual influence of each component. In line with this, the only two previous studies, REGARDS [17] and IWHS Study [20], found that the highest level of LIS was associated with a 30% and 80% higher risk of CVD mortality, respectively. In the current study, we did not find an impact of the pro-inflammatory state of LIS on the development of all CVD events. However, a significant association between LIS and HF was demonstrated, with a one-score increase in LIS correlating to a 36% increased risk. Although there was no clear link between LIS (as a combined effect) and CVD in the MESA participants, a significant strong deleterious impact of current smoking was found, which is supported by the body of evidence [39]. Unexpectedly, a heavy drinking habit was also associated with a lower risk of CVD. While our study was not designed to determine the causal mechanisms underlying this specific finding, several potential explanations should be considered. First, the proportion of heavy drinkers in our study population was relatively small (5.4%), which may have resulted in unstable estimates. Second, residual confounding by unmeasured socioeconomic or other lifestyle factors cannot be excluded. Third, the reliance on self-reported alcohol intake may have also introduced misclassification [40, 41]. Taken together, these considerations underscore the need for cautious interpretation of this result and highlight the importance of replication in other cohorts.

For the first time, we demonstrated that a highly pro-inflammatory lifestyle, measured by LIS, significantly increased the risk of HF. Upon further analysis, we found that among components of LIS, obesity and smoking were significant predictors for HF (Supplementary Table 6). The role of inflammation in the development of HF was addressed in a recent state-of-the-art review [42]. Additionally, research has demonstrated that obesity can contribute to inflammation through changes in pro-inflammatory cytokines, including IL-6, which are released by adipose tissue [43]. In agreement with our findings, a previous study in MESA found a robust association between obesity and HF [1.72 (1.21–2.45)]. However, no significant association was found between obesity and atherosclerotic CVD [44]. Our findings, which noted an almost doubling of the risk of HF and a 73% increased CVD risk in current smokers, were in line with previous studies reporting the unfavorable impact of current smoking on the development of CVD and HF. The detrimental impact of smoking is believed to be due to several factors, such as inducing an inflammatory response, endothelial dysfunction, activation of proatherogenic molecules, and pro-thrombotic conditions [45].

Our findings suggest that reducing pro-inflammatory dietary patterns may may contribute to a lower risk of CVD. Promoting anti-inflammatory diets, particularly among overweight individuals who demonstrated stronger associations, could reduce the overall burden of CVD. In addition, the strong link between the LIS and CHF highlights the importance of addressing lifestyle factors such as smoking and alcohol consumption in prevention strategies.

Strengths and limitations

The main strength of this study lies in its use of a well-characterized, multi-ethnic community-based cohort with a prospective design and long follow-up period. An additional strength was the availability of baseline measurements of key inflammatory biomarkers, including hsCRP and IL-6, which allowed direct assessment of the biological relevance of the DIS and LIS. However, some limitations should be noted. First, the study relied on self-reported data for dietary and some lifestyle variables, which can be prone to errors and biases. Second, the study did not account for all potential confounding variables, such as sleep quality, or other health conditions, which could impact the results. Third, changes in diet and lifestyle over time were not accounted for beyond the baseline, which may have resulted in misclassification. Finally, the study’s findings may not generalize to other non-US populations with different demographics, lifestyles, or health profiles.

Conclusions

During more than a decade of follow-up, we examined the association of the DIS and LIS as tools to assess the collective role of dietary and lifestyle exposures, weighted for their impact on systemic inflammation, with the incidence of CVD, its different subtypes, and HF. A significant positive association between DIS and CVD/hard CVD suggests that inflammation may be the primary mechanism through which diet affects CVD risk, independent of the covariates considered in our models. Additionally, a more pro-inflammatory lifestyle was directly associated with increased HF risk. Our findings, taken together with previous literature, suggest that reducing inflammation, through dietary and lifestyle interventions, could potentially reduce the risk for these inflammatory diseases. Our study supports using the DIS and LIS in further research on inflammation and CVD events.

Abbreviations

DIS

Dietary Inflammation Scores

LIS

Lifestyle Inflammation Scores

CVD

Cardiovascular Disease

MI

Myocardial Infarction

HF

Heart Failure

MESA

Multi-Ethnic Study of Atherosclerosis

HRs

Hazard Ratios

hsCRP

High-sensitivity C-Reactive Protein

EDIP

Empirical Dietary Inflammatory Pattern

BioLINCC

Biologic Specimen and Data Repository Information Coordinating Center

BMI

Body Mass Index

FPG

Fasting Plasma Glucose

TC

Total Cholesterol

HDL-C

High-Density Lipoprotein Cholesterol

MET

Metabolic Equivalent of Task

CHD

Coronary Heart Disease

SBP

Systolic Blood Pressure

DBP

Diastolic Blood Pressure

FFQ

Food-Frequency Questionnaire

NHANES

National Health and Nutrition Examination Survey

LBBB

Left Bundle Branch Block

TIA

Transient Ischemic Attack

ANOVA

Analysis of Variance

IQR

Interquartile Range

IWHS

Iowa Women’s Health Study

REGARDS

Reasons for Geographic And Racial Differences in Stroke

DII

Dietary Inflammatory Index

Authors’ contributions

FH and AR contributed to conceptualizing the study and design; AR analyzed and interpreted the data; AR wrote the initial manuscript. PH and AA contributed to the interpretation of results and discussion. All authors reviewed the manuscript and provided final approval of the manuscript.

Funding

None declared.

Data availability

The data for this study were obtained from the Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC), available at https://www.biolincc.nhlbi.nih.gov.

Declarations

Ethics approval and consent to participate

This study is a secondary analysis of data from the MESA. The MESA study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional review boards at each examination site. Written informed consent was obtained from all participants at each examination. Approval for undertaking the current project was also obtained from the Ethics Committee of Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran (approval code: IR.SBMU.ENDOCRINE.REC.1403.056).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.