Association of Higher Dietary Omega‐3 Fatty Acid Intake With Cardiovascular‐Kidney‐Metabolic Syndrome Stage Severity and Mortality in US Adults: A Cross‐Sectional and Prospective Cohort Analysis of NHANES 1999 to 2018

Abstract

Background

Cardiovascular‐kidney‐metabolic (CKM) syndrome interlinks obesity, diabetes, kidney disease, and cardiovascular dysfunction. Omega‐3 polyunsaturated fatty acids (n‐3 PUFAs) possess anti‐inflammatory and metabolic properties potentially relevant to CKM health, yet their links to disease severity and mortality remain undefined.

Methods

This cross‐sectional and longitudinal analysis included 27 934 US adults with CKM syndrome from NHANES (National Health and Nutrition Examination Survey) 1999 to 2018. Dietary n‐3 PUFA intake (mg/kg per day) was ascertained from 2 24‐hour recalls. Mortality was ascertained via linkage to the National Death Index through 2019. Cross‐sectional CKM stage was analyzed using weighted ordered logistic regression. Mortality risk was evaluated using Cox models, restricted cubic splines, and weighted quantile sum regression.

Results

Each 10 mg/kg per day higher n‐3 PUFA intake was associated with 13% lower odds of advanced CKM stage (odds ratio [OR], 0.87 [95% CI, 0.83–0.91]). Over a median 9‐year follow‐up (5150 deaths), a nonlinear, L‐shaped relationship with all‐cause mortality was observed, with a threshold at 22.35 mg/kg per day. Below this point, each 1 mg/kg per day increment was associated with a 1% lower mortality risk (hazard ratio [HR], 0.99 [95% CI, 0.98–1.00]). Participants in the highest versus lowest intake quartile had 48% lower odds of severe CKM stage (OR, 0.52 [95% CI, 0.46–0.60]) and 25% lower mortality (HR, 0.75 [95% CI, 0.61–0.92]). Docosapentaenoic acid was the primary n‐3 PUFA species associated with mortality risk. Variation in glucose disposal and biological aging collectively explained 6.7% of the mortality association.

Conclusions

Higher n‐3 PUFA intake is associated with less severe CKM stage and linked to lower mortality in an L‐shaped, threshold‐dependent manner.

Nonstandard Abbreviations and Acronyms

ALA

alpha‐linolenic acid

CKM

cardiovascular‐kidney‐metabolic

DHA

docosahexaenoic acid

DPA

docosapentaenoic acid

EPA

eicosapentaenoic acid

KDM

Klemera‐Doubal method

n3‐PUFA

omega‐3 polyunsaturated fatty acids

NHANES

National Health and Nutrition Examination Survey

SDA

stearidonic acid

SGLT2i

sodium–glucose cotransporter‐2 inhibitors

WQS

weighted quantile sum

Clinical Perspective.

What Is New?

• Higher dietary omega‐3 polyunsaturated fatty acid intake demonstrates dose‐dependent association with reduced prevalence of advanced cardiovascular‐kidney‐metabolic syndrome progression and mortality, with docosapentaenoic acid emerging as the predominant protective subtype.

• This inverse association was more pronounced in younger, physically active individuals and was partially accounted for by improved metabolic health and decelerated biological aging.

What Are the Clinical Implications?

• Clinicians may consider personalized dietary strategies emphasizing omega‐3 intake, particularly from docosapentaenoic acid sources; these observational data suggest an inverse association that warrants further investigation.

Cardiovascular‐kidney‐metabolic (CKM) syndrome, ¹ which encompasses the overlapping epidemiology of obesity, diabetes, chronic kidney disease (CKD), and cardiovascular disease (CVD), poses a substantial public health crisis. ² , ³ Individuals suffering from comorbidities experience significantly higher health care resource use and costs compared with those with isolated health conditions; notably, these costs escalate exponentially as the burden of comorbidities increases. ⁴ , ⁵ Recent epidemiological studies indicate that ~90% of US adults qualify for stage 1 or higher CKM status, with 14.6% progressing to stages 3 to 4, where the risk of mortality rises dramatically. ⁶ , ⁷ Additionally, stage 2 affects 56.5% of this population, marked by metabolic and CKD alterations that signify a critical preclinical phase. ⁸ By stratifying this heterogeneous cohort, we can implement targeted preventive strategies designed to mitigate the markedly elevated cardiovascular morbidity and mortality rates. The pathophysiological progression of CKM is influenced by interconnected mechanisms, including chronic inflammation, endothelial dysfunction, and insulin resistance, all of which present potential targets for modification through nutritional interventions. ⁹

Omega‐3 polyunsaturated fatty acids (n‐3 PUFAs)—eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), and plant‐derived alpha‐linolenic acid (ALA)—demonstrate cardioprotective potential through triglyceride reduction, anti‐inflammatory effects, and plaque stabilization. ¹⁰ , ¹¹ Landmark trials, including REDUCE‐IT (A Study of AMR101 to Evaluate Its Ability to Reduce Cardiovascular Events in High‐Risk Patients With Hypertriglyceridemia and on Statin), reported a 25% relative risk reduction in cardiovascular events with high‐dose EPA therapy (icosapent ethyl at 4 g daily) in established CVD. ¹² Mechanistic studies further suggest n‐3 PUFAs mitigate renal fibrosis via TGF‐β (transforming growth factor beta) suppression and enhance insulin sensitivity through GPR120 (G‐protein coupled receptor 120) activation. ¹³ Nevertheless, evidence remains contentious: although GISSI‐P (Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto Miocardio‐Prevenzione) demonstrated significant mortality benefits post myocardial infarction, ¹⁴ subsequent OMEGA (Effect of Omega 3‐Fatty Acids on the Reduction of Sudden Cardiac Death After Myocardial Infarction), STRENGTH (Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High Cardiovascular Risk Patients With Hypertriglyceridemia), and OMEMI (Omega‐3 Fatty Acids in Elderly With Myocardial Infarction) trials failed to replicate these effects in contemporary cohorts receiving optimal medical therapy. ¹⁵ , ¹⁶ , ¹⁷ Recent meta‐analyses have partially reconciled benefits for fatal myocardial infarction (35% risk reduction), yet inconsistencies persist regarding dose–response relationships, population heterogeneity, and outcomes across the CKM spectrum. ¹⁸

Critical translational barriers persist due to 3 interconnected knowledge gaps in CKM research. First, the predominant focus on pharmaceutical n‐3 PUFAs in secondary CVD prevention has marginalized investigation of dietary exposures in multimorbid populations with CKM —where nutritional interventions may yield maximal benefit. Second, evidence remains siloed within general populations or isolated disease subgroups (eg, CKD, osteoarthritis), failing to address how n‐3 PUFA efficacy varies across the CKM severity continuum. This oversight ignores disease‐stage modulation of nutritional effects, a fundamental epidemiological blind spot. Third, observational studies exhibit methodological constraints, including: inadequate adjustment for energy and concurrent fat intake, inconsistent CKM phenotyping, and insufficient analysis of individual n‐3 PUFA subtypes and their mechanisms. Most critically, prospective evidence linking weight‐adjusted n‐3 PUFA intake to (1) CKM stage severity, (2) mortality risk, and (3) potential explanatory pathways like insulin resistance and accelerated aging in high‐burden cohorts remains absent. To bridge these gaps, we leveraged the NHANES (National Health and Nutrition Examination Survey) data to conduct a comprehensive analysis with 2 primary, sequential aims: first, to examine the cross‐sectional association between dietary n‐3 PUFA intake and CKM stage severity; and second, to prospectively investigate the association between n‐3 PUFA intake and mortality risk among individuals with established CKM syndrome.

METHODS

All data are publicly available and can be accessed at the NHANES website (https://wwwn.cdc.gov/nchs/nhanes/Default.aspx). The complete analysis code for this study has been deposited in a GitHub repository and is available through the persistent Zenodo DOI: 10.5281/zenodo.17398447.

Study Design, Setting, and Participants

Data for this study were derived from the 1999 to 2018 NHANES, an ongoing biennial study commissioned by the National Center for Health Statistics within the Centers for Disease Control and Prevention. NHANES employs a stratified multistage probability design aimed at assessing the health and nutritional status of nationally representative samples of noninstitutionalized US civilians. Participants were subjected to household interviews, which were followed by comprehensive examinations conducted at the mobile examination center.

The study received ethical approval from the National Center for Health Statistics Ethics Review Board, and written informed consent was obtained from all participants. This retrospective cohort study strictly adheres to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines, ensuring the reliability and transparency of its methodology. ¹⁹

Starting with an initial cohort of 55 081 adults aged 20 years and older, we systematically excluded individuals based on specific criteria: those with baseline pregnancy (n=1541), missing CKM staging parameters (n=13 313), unreported n‐3 PUFA intake (n=4005), missing or zero dietary weights on both recall days (n=3399), incomplete mortality follow‐up (n=43), or unavailable covariate data (n=4846). As a result, the final analytical sample consisted of 27 934 eligible participants, resulting in a data set suitable for robust analysis (see Figure S1 for the selection flow). For the analysis of CKM stage progression, all 27 934 participants with defined CKM staging were included in a cross‐sectional analysis. For the prospective mortality analysis, the analytic sample was restricted to the 26 400 participants with CKM syndrome (stages 1–4) at baseline.

Definition of CKM Syndrome

CKM syndrome staging followed established criteria (Data S1). ²⁰ , ²¹ Participants were stratified according to the established CKM staging system (stages 0 to 4): stage 0: metabolically healthy; stage 1: prediabetes or overweight/obesity; stage 2: single additional metabolic risk factor or CKD; stage 3: Subclinical CVD with metabolic risk factors/CKD; and stage 4: established clinical CVD. Stage 3 encompassed participants with (1) ≥20% 10‐year CVD risk (calculated using American Heart Association [AHA]‐endorsed PREVENT equations), ²² or(2b) very high‐risk CKD (classified per Kidney Disease: Improving Global Outcomes guidelines ²³ ). Stage 4 included participants with clinical CVD manifestations (coronary heart disease, heart failure, or stroke). Comprehensive staging criteria appear in Data S1. For our analytical models examining stage severity, we treated this as an ordinal outcome (0–4). Following reported prevalence patterns in US adults, we defined CKM syndrome: stage ≥1; advanced CKM stages: stages 3 to 4 (associated with substantially elevated CVD risk). ²⁴ This operationalization of the CKM staging framework using NHANES data is consistent with the approach used in other contemporary epidemiological studies of CKM syndrome in this population. ⁶ , ²⁵ , ²⁶ , ²⁷

Dietary Assessment of n3‐PUFA

The NHANES dietary assessment employed a fully computerized recall system with standardized food‐specific questions and response options to ensure rigorous data collection. Nutritional values were derived using the Continuing Survey of Food Intakes by Individuals database and Automated Multiple‐Pass Method. ²⁸ Detailed methodology is documented in the NHANES Dietary Interviewer Procedure Manual. ²⁹

For NHANES cycles other than 1999 to 2002 (which used only a single 24‐hour dietary recall), we estimated n3‐PUFA intake levels using the mean values derived from 2 24‐hour recalls. The first interview was conducted in person at the mobile examination center, with the second administered telephonically 3 to 10 days later. We defined n‐3 PUFA as the sum of 18:3, 18:4, 20:5, 22:5, and 22:6 fatty acids, and n‐6 PUFA comprised 18:2 and 20:4. ³⁰ Body weight‐adjusted mean daily intake (mg/kg/d) of n3‐PUFAs was calculated from the 2 recalls after exclusion of participants with incomplete dietary data. Intakes were stratified into quartiles: Q1 (<25th percentile, reference), Q2 (25th–50th percentile), Q3 (50th–75th percentile), and Q4 (≥75th percentile). The n‐3 PUFA intake values used in our analyses were derived solely from dietary food sources and did not include intake from dietary supplements (eg, fish oil capsules). The primary dietary sources of ALA included plant oils, nuts, and seeds, and the primary sources of EPA, DHA, and docosapentaenoic acid (DPA) were fish and seafood, as corroborated by NHANES data on fish consumption over the past 30 days.

Composite Indicators

The equation for calculating estimated glucose disposal rate (eGDR, mg/kg per min) is as follows: eGDR=21.158−(0.09×waist circumference [cm])−(3.407×hypertension)−(0.551×glycated hemoglobin). ³¹

Biological age was quantified using the Klemera–Doubal method (KDM‐BA). ³² Accelerated aging was assessed by computing Klemera–Doubal‐biological age acceleration, which compares the discrepancy between Klemera–Doubal‐biological age and chronological age. This approach has been validated in prior studies demonstrating high concordance between calculations incorporating and excluding CRP (C‐reactive protein). ³³ , ³⁴

Assessment of Covariates

Final adjusted covariates included sociodemographics (sex, age, education, race, marital status, poverty‐income ratio), lifestyle factors (smoking status, alcohol consumption, physical activity), body mass index, clinical parameters (self‐reported cancer, lipid‐lowering medications), nutritional factors (dietary supplements, vitamin B6, total protein, fiber, energy, carbohydrates). Covariate selection and definitions are detailed in Data S1 and Figure S2.

Definition of Mortality Outcomes

Mortality information, encompassing the dates and reasons for death, was sourced from the National Death Index records up to December 31, 2019. The categorization of causes relied on the International Classification of Diseases, Tenth Revision (ICD‐10). Follow‐up began on the date of survey participation and extended until either death occurred or the observation period concluded. The main outcome measure was all‐cause mortality, and secondary outcomes encompassed deaths attributed to both cardiovascular and noncardiovascular factors.

Statistical Analysis

To ensure nationally representative estimates for the noninstitutionalized US civilian population with CKM syndrome, all analyses incorporated NHANES sampling weights. Analyses using dietary recall data applied the 2‐day dietary weight (wtdr2d) per NHANES guidelines. The complex sampling design was addressed through a survey object created with the svydesign function (R survey package), ³⁵ specifying strata, clusters, and wtdr2d weights. This design‐informed approach was implemented in all primary analyses: survey‐weighted Cox regression (svycoxph), ordinal logistic regression (svyolr), Kaplan–Meier survival estimation (svysurvfit), rms‐based restricted cubic splines, ³⁶ and mediation analyses (mediation). ³⁷ , ³⁸ These particular methods lack survey‐weight support (Fine–Gray subdistribution hazard models, weighted quantile sum [WQS] and quantile g‐computation [Qgcomp]), and unweighted analyses were necessarily performed. These particular models were performed without survey weights, as the statistical methodology for integrating complex survey designs into these specific algorithms is not yet fully established or implemented in standard software.

Continuous variables are expressed as weighted means (±standard error), and categorical variables as unweighted counts (weighted percentages). Group comparisons employed the Wilcoxon rank‐sum test for complex survey designs and the Rao–Scott second‐order corrected chi‐square test. Dietary n‐3 PUFA intake was modeled as (1) a continuous variable (per 10‐mg/kg per day increment), and (2) a categorical variable (quartiles Q1–Q4, with Q1 as reference). Trend tests across quartiles were computed using consecutive integer scoring (1–4).

Survey‐weighted ordinal logistic regression models examined prospective associations between dietary n‐3 PUFA intake and CKM stage progression (0–4). Sequential adjustment models were constructed: Model 1: unadjusted; Model 2: sociodemographic factors (sex, age, education, race, marital status, poverty–income ratio [PIR]); Model 3: Model 2+behavioral covariates (smoking, alcohol consumption, physical activity); Model 4: Model 3+clinical parameters (self‐reported cancer, lipid‐lowering medications)+nutritional covariates (dietary supplements, vitamin B6, total protein, total fiber, total energy, total carbohydrate). Effect estimates are presented as odds ratios (ORs) with 95% CIs. Trend tests across tertiles were conducted by assigning ordinal scores.

Kaplan–Meier curves were generated to illustrate survival probabilities, and log‐rank tests were employed to compare mortality across quartiles of n‐3 PUFA intake. Cox regression results for n‐3 PUFA intake are presented. To explore potential nonlinear relationships between n‐3 PUFA intake and mortality, we used 3‐knot restricted cubic splines. Threshold analysis for n‐3 PUFA intake was conducted using the ‘segmented’ package in R (version 2.1–4). The algorithm was initialized near the median n‐3 PUFA intake level (17.74 mg/kg per day) and iteratively optimized to the final estimate of 22.35 mg/kg per day. Bootstrap resampling with 1000 replicates was used to estimate the 95% CI for the threshold. The analyses were rigorously adjusted for a range of covariates, including sex, age, education level, race, marital status, PIR, smoking status, alcohol consumption, physical activity, body mass index, self‐reported cancer history, lipid‐lowering medication use, dietary supplement use, vitamin B6 intake, and macronutrient intake (total energy, protein, carbohydrate, and fiber). Subgroup and interaction analyses were conducted to explore potential effect modifications across key demographic and clinical factors. These analyses should be interpreted as exploratory and hypothesis‐generating. Subgroup analyses were conducted to evaluate the associations stratified by key demographic and health‐related factors: age (<65 or ≥65 years), sex, marital status, PIR (<1.3, 1.3–3.0, ≥3.0), physical activity (<600, 600–1500, >1500 metabolic equivalent of task (MET)‐min/week), lipid‐lowering medication use, and body mass index (<25, 25–29.9, ≥30 kg/m²). This stratification allows for a more nuanced understanding of how different variables may influence the association between n‐3 PUFA intake and mortality outcomes.

To further investigate noncardiovascular mortality, we first described the distribution of specific causes of death. Proportional mortality for major noncardiovascular causes was calculated (Figure S3). Cancer was identified as the leading noncardiovascular cause, accounting for 57.3% of noncardiovascular deaths (Figure S3). We then applied survey‐weighted Cox proportional hazards models to estimate associations between n‐3 PUFA intake (per 10 mg/kg per day) and specific cause mortality. Given that cancer was the predominant noncardiovascular cause, we further examined n‐3 PUFA intake (categorical and continuous) in relation to cancer‐specific mortality. Recognizing the mutually exclusive nature of mortality end points, we performed supplementary competing‐risk analyses using Fine–Gray subdistribution hazard models. ³⁹ Cardiovascular death and cancer death were analyzed as primary events with appropriate competing events. These analyses were conducted without survey weights due to current software limitations.

WQS regression (“gWQS” package implementation) ⁴⁰ evaluated the combined effects of multiple n3‐PUFAs on mortality risk and their constituent contributions. The derived composite WQS index encapsulated the cumulative effect of 5 n3‐PUFAs; its association with mortality directly quantified combined exposure impacts. Index weights delineated relative compound contributions: elevated weights (mean>0.25) signified significant protective effects, whereas minimal weights (≤0.05) indicated negligible biological relevance. ⁴¹ , ⁴² Validation involved stratified random sampling (70%/30% training/validation) with 1000 bootstrap iterations to ensure robustness. For cohort‐specific effect estimation, WQS regression quantified mortality risk associations for ALA, EPA, DPA, DHA, and stearidonic acid (SDA) in CKM stages 1 to 4 (n=26 400) and Ssages 1 to 2 (n=20 521). Models integrated 19 covariates across demographic, lifestyle, clinical, and nutritional domains. Component weights derived from 1000 bootstrap iterations incorporated directionally constrained coefficients (β₁ < 0), isolating protective associations. To address potential limitations of the WQS regression model, particularly its assumption of unidirectional associations, and to assess the robustness of our mixture findings, we implemented a Qgcomp approach as a sensitivity analysis. ⁴¹ In contrast to WQS, Qgcomp does not impose an a priori constraint on the directionality of associations between individual n‐3 PUFAs and the outcome (“qgcomp” package implementation). ⁴³ The method quantifies exposures through quartile coding and evaluates the joint effect of the mixture, with the weights for individual components reflecting their partial contributions to the overall mixture effect, which can be positive or negative. The same set of 19 covariates used in the primary WQS analysis was included for adjustment.

We examined potential mediating roles of eGDR and Klemera–Doubal‐biological age acceleration in the association between n3‐PUFA intake and mortality outcomes using the ‘mediation’ package. Adjusted models incorporating 500 bootstrap samples were analyzed to decompose total effects into direct and indirect pathways via the mediation function.

To ensure the robustness of our findings, we conducted 6 sensitivity analyses: (1) We addressed the issue of missing covariate data through multiple imputation using the ‘mice’ package in R. This approach employed the multivariate imputation by chained equations method, which effectively minimizes bias resulting from incomplete observations (Figure S4). (2) We implemented a restricted definition of CKM, focusing solely on the 2011 to 2018 survey cycles that included Asian ethnicity data while excluding participants aged 80 years and older. This exclusion is justified as NHANES Asian ethnicity data began in 2011, and participant age is top‐coded at 80 years. (3) To further mitigate potential reverse causality concerns, we excluded participants who died within the first 2 follow‐up years. (4) Given the established therapeutic effects of SGLT2 (sodium‐glucose cotransporter‐2) inhibitors and GLP‐1RA (glucagon‐like peptide‐1 receptor agonists) on advanced CKM dysfunction, we conducted sensitivity analyses excluding participants receiving these medications. (5) We also repeated our analyses after excluding participants with missing covariate or mediator data, thus ensuring our analysis was restricted to complete cases. (6) To address potential concerns regarding body weight normalization, we conducted sensitivity analyses using 2 established energy‐adjustment methods: the energy density method (mg/1000 kcal) and the residual method. The energy density method expresses n‐3 PUFA intake relative to total energy consumption (n‐3 PUFA/total energy×1000), whereas the residual method employs energy‐adjusted residuals derived from regressing n‐3 PUFA intake on total energy intake. These approaches follow standard recommendations for energy adjustment in nutritional epidemiology ⁴⁴ , ⁴⁵ and align with causal inference considerations. ⁴⁶ All other analytical parameters, including statistical methods (weighted Cox proportional hazards models) and complex survey design accounting, remained consistent with our primary analysis. (7) Finally, we calculated E‐values to quantify the potential impact of unmeasured confounding on the associations between high and low dietary intake of n‐3 PUFA (Q4 versus Q1) and mortality. These E‐values provide a measure of the minimum strength of association that an unmeasured confounder would need to have with both exposure and outcome in order to explain away the observed hazard ratio (HR).

A P value <0.05 was considered statistically significant in a 2‐tailed test. All analyses were conducted using R software (version 4.4.3) and Free Statistics (version 2.2; Beijing FreeClinical Medical Technology Co., Ltd, Beijing, China). ³⁰ , ³³ , ⁴⁷ , ⁴⁸

RESULTS

Baseline Characteristics

To provide context for the mortality analyses, the Table 1 presents the baseline characteristics of the entire study cohort across all CKM stages (0–4), stratified by survival status during follow‐up. This demonstrates the progressive increase in mortality risk across advancing CKM stages, with stages 3 to 4 showing substantially higher mortality rates compared with earlier stages. The Table 1 delineates baseline characteristics stratified by survival status in this nationally representative cohort of 27 934 US adults (weighted N=134.6 million; weighted mean age 50.62±0.22 years). After a median 9‐year follow‐up, 5150 participants died (weighted N=18.4 million). Nonsurvivors were significantly older (66.46±0.31 versus 48.10±0.22 years; P<0.001), more frequently male (52.62% versus 49.97%; P=0.014), and disproportionately Non‐Hispanic White (80.16% versus 69.74%; P<0.001). They exhibited lower educational attainment (11.89% with <high school versus 4.31%), higher poverty rates (49.34% with PIR 1.0–3.0 versus 34.80%), greater prevalence of living alone (43.91% versus 34.42%), and lower physical activity levels (68.72%<600 metabolic equivalent of task‐min/week versus 45.12%; all P<0.001).

Table 1.

Baseline Characteristics of Participants With Cardiovascular‐Kidney‐Metabolic Syndrome (Stages 0–4) Categorized by Survival Status

Characteristic	Overall participants (N=134 640 413, n=27 934)	Survivors (N=116 221 715, n=22 784)	Nonsurvivors (N=18 418 699, n=5150)	P value*
Age, y	50.62 (0.22)	48.10 (0.22)	66.46 (0.31)	<0.001
Male sex, n (%)	14 154 (50.34%)	11 277 (49.97%)	2877 (52.62%)	0.014
Race and ethnicity, n (%)				<0.001
Non‐Hispanic White	13 436 (71.17%)	10 221 (69.74%)	3215 (80.16%)
Non‐Hispanic Black	5933 (10.93%)	4949 (11.02%)	984 (10.42%)
Mexican American	4442 (7.03%)	3829 (7.60%)	613 (3.43%)
Hispanic and Non‐Hispanic Multiracial	4123 (10.87%)	3785 (11.64%)	338 (5.99%)
Education level, n (%)				<0.001
Less than high school	3122 (5.35%)	2146 (4.31%)	976 (11.89%)
High school or equivalent	10 691 (35.69%)	8374 (34.02%)	2317 (46.23%)
College or above	14 121 (58.96%)	12 264 (61.67%)	1857 (41.88%)
Marital status, n (%)				<0.001
Coupled	17 079 (64.28%)	14 300 (65.58%)	2779 (56.09%)
Living alone	10 855 (35.72%)	8484 (34.42%)	2371 (43.91%)
Family poverty–income ratio, n (%)				<0.001
<1.0	5193 (12.98%)	4212 (12.64%)	981 (15.15%)
1.0–3.0	12 033 (36.79%)	9309 (34.80%)	2724 (49.34%)
>3.0	10 708 (50.22%)	9263 (52.56%)	1445 (35.51%)
Smoking status, n (%)				<0.001
Current smoker	14 481 (51.76%)	12 425 (53.87%)	2056 (38.45%)
Former smoker	7907 (27.97%)	5833 (26.26%)	2074 (38.74%)
Never smoker	5546 (20.27%)	4526 (19.87%)	1020 (22.81%)
Alcohol consumption, n (%)				<0.001
Never drinker	6914 (20.57%)	5410 (19.48%)	1504 (27.44%)
Past drinkers	4077 (11.88%)	2827 (10.27%)	1250 (22.03%)
Mild drinkers	8595 (33.97%)	7083 (34.47%)	1512 (30.84%)
Moderate drinkers	5416 (22.70%)	4828 (24.23%)	588 (13.08%)
Heavy drinkers	2932 (10.88%)	2636 (11.56%)	296 (6.60%)
Physical activity, n (%)				<0.001
<600 MET‐min/wk	14 736 (48.34%)	11 107 (45.12%)	3629 (68.72%)
600–1500 MET‐min/wk	4406 (16.55%)	3737 (17.03%)	669 (13.53%)
>1500 MET‐min/wk	8792 (35.11%)	7940 (37.86%)	852 (17.75%)
Body mass index, kg/m2	29.50 (0.08)	29.56 (0.09)	29.12 (0.14)	0.006
Dietary supplements, n (%)	14 882 (56.40%)	11 873 (55.55%)	3009 (61.71%)	<0.001
n‐3 PUFAs (mg/kg/d)	22.10 (0.15)	22.50 (0.17)	19.55 (0.29)	<0.001
n‐6 PUFAs (mg/kg/d)	198.08 (1.15)	201.95 (1.21)	173.66 (2.43)	<0.001
Vitamin B6 (mg/kg/d)	0.03 (0.00)	0.03 (0.00)	0.02 (0.00)	<0.001
Total energy (kcal/d)	2104.24 (8.57)	2142.26 (8.77)	1864.33 (18.31)	<0.001
Total protein (g/d)	81.96 (0.36)	83.45 (0.37)	72.51 (0.76)	<0.001
Total fat (g/d)	81.23 (0.42)	82.85 (0.43)	71.03 (0.86)	<0.001
Total saturated fatty acids (g/d)	26.54 (0.15)	27.05 (0.16)	23.34 (0.31)	<0.001
Total fiber (g/d)	16.60 (0.13)	16.83 (0.13)	15.12 (0.20)	<0.001
Total carbohydrate (g/d)	251.62 (1.11)	255.52 (1.16)	227.02 (2.33)	<0.001
Self‐reported cancer, n (%)	3176 (11.62%)	2014 (9.60%)	1162 (24.36%)	<0.001
Sodium–glucose cotransporter‐2 inhibitors, n (%)	44 (0.27%)	42 (0.32%)	2 (0.04%)	0.003
Glucagon‐like peptide 1 receptor agonists agonist, n (%)	79 (0.40%)	74 (0.46%)	5 (0.13%)	0.015
Lipid‐lowering medications, n (%)	6366 (21.03%)	4619 (19.10%)	1747 (33.18%)	<0.001
Estimated glucose disposal rate	6.68 (0.03)	6.83 (0.04)	5.77 (0.05)	<0.001
Klemera–Doubal method biological age, y
KDM biological age	51.90 (0.23)	49.38 (0.23)	68.22 (0.38)	<0.001
KDM advance acceleration	1.38 (0.16)	1.34 (0.17)	1.60 (0.35)	0.011
Cardiovascular‐kidney‐metabolic syndrome stage				<0.001
0	1534 (6.85%)	1495 (7.77%)	39 (1.04%)
1	3431 (13.31%)	3268 (14.92%)	163 (3.15%)
2	17 090 (63.13%)	14 627 (65.32%)	2463 (49.28%)
3	1825 (4.62%)	1084 (3.25%)	741 (13.28%)
4	4054 (12.09%)	2310 (8.74%)	1744 (33.26%)

Values were weighted mean (mean±SE) for continuous variables or unweighted numbers (weighted percentage) for categorical variables. N represents weighted numbers, and n represents unweighted numbers. Blood lipid‐lowering drugs include fibrate, statin, ezetimibe, proprotein convertase subtilisin/kexin type 9 inhibitors, omega‐3, and niacin. KDM indicates Klemera–Doubal method; MET, metabolic equivalent of task; and PUFAs, polyunsaturated fatty acids.

^*Wilcoxon rank‐sum test for complex survey samples; chi‐square test with Rao and Scott’s second‐order correction.

Nonsurvivors exhibited 11% higher supplement use (61.7% versus 55.6%; P<0.001), despite consumption of n‐3/n‐6 PUFA being 13% to 14% lower and a 13% reduction in overall energy intake (all P<0.001). Furthermore, nonsurvivors exhibited a 16% lower eGDR (5.77±0.05 versus 6.83±0.04 mg/kg per min; P<0.001) and a 19% acceleration in biological aging (1.60±0.35 versus 1.34±0.17 years; P=0.011). Advanced CKM stages (3–4) were 4‐fold more prevalent among nonsurvivors (33.3% versus 8.7%; P<0.001). Use of conventional lipid‐lowering therapy was 74% higher in nonsurvivors (33.2% versus 19.1%; P<0.001), whereas uptake of cardioprotective antihyperglycemic agents (SGLT2 inhibitors/GLP‐1 receptor agonists) remained negligible (<0.5%). Table S4 stratifies participant characteristics by CKM stage (0–4), revealing stage‐dependent increases in age, metabolic derangements, and clinical complexity. Mortality tracked positively with CKM stage severity and declined dose‐dependently across n‐3 PUFA quartiles (Tables S3 and S4 and Figure S5).

Cross‐Sectional Associations Between Dietary n‐3 PUFA Intake and Progression of CKM Staging From 0 to 4

Figure 1A and Table S5 present the association of n‐3 PUFA intake with CKM stage severity. In continuous analyses, each 10 mg/kg per day increment in n‐3 PUFA intake was significantly associated with lower odds of higher CKM stages across all models. The fully adjusted Model 4 yielded an OR of 0.87 (95% CI, 0.83–0.91; P<0.001). Quartile analyses demonstrated a significant inverse dose–response relationship. Compared with the lowest quartile (Q1), participants in the highest intake quartile (Q4) exhibited markedly associated with a lower odds of advanced CKM stage in Model 4 (OR, 0.52 [95% CI, 0.46–0.60], P<0.001). This protective gradient was consistent across all quartiles (Q2: OR, 0.82; Q3: OR, 0.65) and remained robust through progressive multivariable adjustment. Statistically significant trends (P<0.001 for all models) reinforced the dose‐dependent inverse association between increasing n‐3 PUFA intake and CKM stage severity. These findings were strongly supported by 5 sensitivity analyses, demonstrating the reliability of the results (Figure 1B through 1E, Tables S6 through S10).

Figure 1. Cross‐sectional associations between n‐3 PUFA intake and CKM staging (0–4) using survey‐weighted ordinal logistic regression.

A, Complete‐case analysis excluding participants with missing covariates; B, Sensitivity analysis with multiple imputation; C, Sensitivity analysis with stricter CKM definition; D, Sensitivity analysis after excluding participants who died within 2‐year follow‐up; E, Sensitivity analysis excluding participants treated with SGLT2i or GLP‐1RA; F, Sensitivity analysis after excluding covariate missingness and mediator missingness. Model 1 (unadjusted); Model 2 adjusted for sociodemographic factors (sex, age, education, race, marital status, poverty‐income ratio); Model 3 further adjusted for lifestyle factors (smoking, alcohol consumption, physical activity); Model 4 additionally adjusted for clinical parameters (self‐reported cancer, lipid‐lowering medication use) and nutritional covariates (dietary supplements, vitamin B6, total protein, energy, carbohydrate, fiber, saturated fatty acids). Blood lipid‐lowering drugs include fibrate, statin, ezetimibe, proprotein convertase subtilisin/kexin type 9 inhibitors, omega‐3, and niacin. CKM indicates cardiovascular‐kidney‐metabolic; GLP1 agonist, glucagon‐like peptide 1 receptor agonists; n‐3 PUFA, n‐3 polyunsaturated fatty acids; and SGLT2i, sodium–glucose cotransporter‐2 inhibitors.

Prospective Associations Between Dietary n‐3 PUFA Intake and Mortality in Participants With CKM Syndrome

Figures 2 and Figures S6 present clear dose‐dependent inverse relationships between n‐3 PUFA intake and mortality risk in the overall cohort with CKM. Notably, significant inverse associations were identified for both all‐cause mortality and noncardiovascular mortality across the entire cohort, with particularly strong results observed in the subgroup with CKM stage 1/2. In the full cohort, an increased intake of n‐3 PUFA was associated with a lower risk of all‐cause mortality (Q4 versus Q1 HR, 0.85 [95% CI, 0.73–0.99]; P‐trend=0.023) and noncardiovascular mortality (Q4 versus Q1 HR, 0.80 [95% CI, 0.67–0.94]; P‐trend=0.005). These protective effects were even more pronounced in the stage 1/2 subgroup, which exhibited a 25% reduction in the risk of all‐cause mortality (Q4 versus Q1 HR, 0.75 [95% CI, 0.61–0.92]; P‐trend=0.003) and a 30% reduction in noncardiovascular mortality risk (Q4 versus Q1 HR, 0.70 [95% CI, 0.56–0.87]; P‐trend=0.001). Furthermore, a 10 mg/kg per day increase in n‐3 PUFA intake was shown to significantly decrease noncardiovascular mortality among participants in stages 1 and 2 (HR, 0.93 [95% CI, 0.88–0.99]; P=0.015). Conversely, no significant associations were observed for cardiovascular mortality across any subgroup (all P‐trend >0.05).

Figure 2. Prospective associations between dietary n‐3 PUFA intake and mortality among participants with CKM syndrome (stages 1–4) after excluding participants with missing covariates.

A, Entire population with CKM syndrome, B, subgroup with CKM stages 1/2, C, subgroup with CKM stages 3/4. N represents weighted numbers, and n represents unweighted numbers. P value <0.001: ***; P value <0.01: **; P value <0.05: *; Within each subfigure, these dots mark the reference point (OR, 1.0) on the horizontal line. Analyses were adjusted for: demographic factors (age, sex, race or ethnicity); socioeconomic status (education level, marital status, poverty‐income ratio); lifestyle behavior (smoking status, alcohol consumption, physical activity); body mass index; clinical parameters (self‐reported cancer, lipid‐lowering medication use); and nutritional covariates (dietary supplement use, vitamin B6 intake, total protein, total fiber, total energy, total carbohydrate, and total saturated fatty acids). Blood lipid‐lowering drugs include fibrate, statin, ezetimibe, proprotein convertase subtilisin/kexin type 9 inhibitors, omega‐3, and niacin. CKM indicates cardiovascular‐kidney‐metabolic; CVD, cardiovascular disease; OR, odds ratio; and PUFA, Polyunsaturated Fatty Acids.

Restricted Cubic Spline and Threshold Effect Analysis

Figure 3 presents the results of the smoothing curve fitting, which illustrates the relationships between dietary intake of n‐3 PUFA and mortality outcomes. Restricted cubic spline analyses revealed significant inverse associations between dietary n‐3 PUFA intake and noncardiovascular mortality within the overall cohort with CKM syndrome. Additionally, similar inverse associations were noted for all‐cause and noncardiovascular mortality specifically in the subgroup with CKM stage 1/2 (all P‐overall <0.05; see Figures 3C, 3D, 3F). The nonsignificant P‐nonlinearity values (> 0.05) supported a linear dose–response relationship for these outcomes. Importantly, a nonlinear, L‐shaped relationship was identified between dietary n‐3 PUFA intake and all‐cause mortality in the overall CKM cohort (P‐nonlinear=0.036; see Figure 3A). Threshold regression analysis identified the optimal breakpoint at 22.35 mg/kg per day (95% CI, 21.98–22.73). In the 3‐piecewise Cox regression analysis: Below 22.35 mg/kg per day, each 1 mg/kg per day increase in n‐3 PUFA intake was associated with a 1% lower all‐cause mortality risk (HR, 0.99 [95% CI, 0.98–1.00], P=0.008). Above this threshold, additional n‐3 PUFA intake showed no significant mortality benefit (HR, 1.00 [95% CI, 0.99–1.01], P=0.285) (Tables S11). A formal likelihood ratio test comparing the threshold model against a linear specification demonstrated significantly improved fit (P=0.024), providing strong statistical evidence for a nonlinear relationship. This finding highlights the significant potential benefits of increasing n‐3 PUFA intake for reducing mortality risk, particularly at lower levels of consumption, suggesting that even modest increases in intake could lead to meaningful health benefits.

Figure 3. Restricted cubic spline analyses depicting associations between dietary n‐3 PUFA intake and mortality.

A–C, Entire population with CKM syndrome: all‐cause (A), cardiovascular (B), and noncardiovascular (C) mortality; (D–F) subgroup with CKM stages 1/2: all‐cause (D), cardiovascular (E), and noncardiovascular (F) mortality; (G–I) subgroup with CKM stages 3/4: all‐cause (G), cardiovascular (H), and noncardiovascular (I) mortality. The median n‐3 PUFA intake (dashed vertical line) serves as the reference point. Solid lines represent point estimates; dashed lines indicate 95% CIs. Analyses were adjusted for demographic factors (age, sex, race or ethnicity); socioeconomic status (education level, marital status, poverty‐income ratio); lifestyle behaviors (smoking status, alcohol consumption, physical activity); body mass index; clinical parameters (self‐reported cancer, lipid‐lowering medication use); and nutritional covariates (dietary supplement use, vitamin B6 intake, total protein, total fiber, total energy, total carbohydrate, and total saturated fatty acids). Lipid‐lowering medication include fibrate, statin, ezetimibe, proprotein convertase subtilisin/kexin type 9 inhibitors, omega‐3 and niacin. CKM indicates cardiovascular‐kidney‐metabolic; HR, hazard ratio; and PUFA, polyunsaturated fatty acids.

Kaplan–Meier Survival Curves

Kaplan–Meier curves demonstrated a significantly lower risk of all‐cause mortality for participants in the highest quartile of dietary n‐3 PUFA intake compared with those in the lowest quartile in the entire population with CKM syndrome, as well as in the subgroups with CKM 1/2 stage and CKM3/4 stage (all P<0.05) (Figure 4).

Figure 4. Kaplan–Meier curve analyses for associations between dietary n‐3 PUFA intake and mortality.

All‐cause (A), cardiovascular (B), and noncardiovascular (C) mortality for the entire population with CKM syndrome. All‐cause (D), cardiovascular (E), and noncardiovascular (F) mortality for the subgroup with CKM stages 1/2. All‐cause (G), cardiovascular (H), and noncardiovascular (I) mortality for the subgroup with CKM stages 3/4. CKM indicates cardiovascular‐kidney‐metabolic; and n‐3 PUFA, n‐3 polyunsaturated fatty acids.

Subgroup Analyses

Tables S12 through S14 present subgroup analyses exploring associations between n‐3 PUFA intake and all‐cause, cardiovascular, and noncardiovascular mortality. Analyses were stratified by key demographic and clinical factors: age (<65 or ≥65 years), sex, marital status, PIR (<1.3, 1.3–3.0, ≥3.0), physical activity (<600, 600–1500, >1500 metabolic equivalent of task‐min/week), lipid‐lowering medication use, and body mass index (<25, 25–29.9, ≥30 kg/m²).

A more pronounced inverse association between n‐3 PUFA intake and both all‐cause and noncardiovascular mortality was observed in the overall cohort with CKM, particularly among participants aged <65 years (P‐interaction=0.016) and those with higher physical activity levels (P‐interaction=0.017); Table S12. Similarly, within the subgroup with CKM stage 1/2, stronger inverse associations with all‐cause mortality emerged in younger individuals (P‐interaction=0.039) and those with higher PIR (P‐interaction=0.030); Table S13. Among the subgroup with CKM stage 3/4, higher physical activity levels significantly strengthened the inverse association with noncardiovascular mortality (P for interaction=0.029); Table S14. These associations remained consistent across all other subgroups (P for interaction >0.05), supporting the robustness of the findings.

Specific Cause and Competing Risk Analyses for Mortality

In initial screening using continuous n‐3 PUFA intake, inverse associations with cancer mortality were most prominent, particularly in early CKM stages (Figure S7). Further analysis demonstrated that participants in the highest quartile of n‐3 PUFA intake had significantly lower cancer mortality risk in the subgroup with CKM stages 1/2 (Q4 versus Q1 HR, 0.65 [95% CI, 0.45–0.95]; P‐trend=0.027; Figure S8). To account for competing causes of death, we performed Fine–Gray subdistribution hazard analyses (Figure S9). These results were consistent with the primary Cox models: there was no association between n‐3 PUFA intake and cardiovascular mortality, whereas the inverse association with cancer mortality in the subgroup with CKM stages 1/2 persisted (Q4 versus Q1: subdistribution HR, 0.76 [95% CI, 0.59–0.99]).

Combined Effects of n3‐PUFA on Mortality Outcomes

Plant‐sourced ALA represented the predominant contributor (>90%), with marine‐derived EPA, DHA, and DPA comprising ∼10% collectively (Table S15). SDA intake maintained minimal levels throughout, contrasting markedly with the significant upward trajectory in total n‐3 PUFA consumption. Analysis across 10 NHANES cycles revealed a time‐dependent elevation in total n‐3 PUFA intake (P‐trend <0.01; Figure S10), primarily attributable to increasing ALA consumption (P‐trend <0.01). Although fish oil‐based EPA intake showed stability (P‐trend >0.05), DHA levels demonstrated a downward trajectory (P‐trend <0.01). Post 2009 to 2010 cycles, DPA dietary intake displayed progressive augmentation (P‐trend <0.01).

WQS regression revealed significant protective associations across early CKM stages (Tables S16 and S17). For CKM 1to 4, each quartile increase in the n‐3 PUFA mixture index associated with a reduction in all‐cause mortality by 18.8% (β=−0.188 [95% CI, −0.208 to −0.169]; P<0.001), primarily driven by DPA (52.8%); cardiovascular mortality by 14.2% (β=−0.142 [95% CI, −0.172 to −0.112]; P<0.001), with DPA contributing 67.7% of the effect; and noncardiovascular mortality by 18.4% (β=−0.184 [95% CI, −0.208 to −0.160]; P<0.001), predominantly from DPA (43.6%) and SDA (37.1%). In CKM 1/2 (n=87 502 156), stronger effects emerged: DPA dominated protection against all‐cause mortality (56.4% weight, β=−0.180). Cardiovascular mortality reduction (−14.8%) featured DPA (59.1%) and EPA (10.1%) synergy. Noncardiovascular mortality decreased 17.4% per quartile (β=−0.174), with balanced DPA (51.6%) and ALA (30.1%) contributions. Stage‐stratified weights demonstrated DPA’s consistent primacy (>50% in 5/6 mortality outcomes) and EPA’s stage‐specific cardiovascular role (Figure 5).

Figure 5. The weights assigned to n3‐PUFA from WQS models.

A–C, Entire population with CKM syndrome population: all‐cause (A), cardiovascular (B), and noncardiovascular (C) mortality; D–F, Subgroup with CKM stages 1/2: all‐cause (D), cardiovascular (E), and noncardiovascular (F) mortality. ALA indicates alpha‐linolenic acid; CKM. cardiovascular‐kidney‐metabolic; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid; n‐3 PUFA, n‐3 polyunsaturated fatty acids; SDA, stearidonic acid; and WQS, weighted quantile sum.

The Qgcomp analysis, which relaxes the unidirectional assumption of WQS, confirmed the primary role of DPA in the protective association of the n‐3 PUFA mixture with mortality. As detailed in Tables S18 and S19, each quartile increase in the n‐3 PUFA mixture was associated with significant reductions in all‐cause, cardiovascular, and noncardiovascular mortality in both the overall CKM (stages 1–4) and early CKM (stages 1–2) cohorts (all P<0.001). In these models, DPA consistently carried the largest positive weight (ranging from 28.8% to 40.9%) toward the overall protective effect, followed by DHA and ALA. The weights for EPA and SDA were generally smaller. The concordance between the WQS and Qgcomp results, despite their different modeling assumptions, strengthens the evidence for a particularly potent role of DPA per unit of intake.

Mediation Analyses

Mediation analyses were conducted to examine potential explanatory pathways. Specifically, we used mediation analysis to investigate whether the association of n‐3 PUFA intake with lower all‐cause mortality (CKM stages 1–4) operated through specific pathways (Figure 6). The association was primarily direct (β=−0.14 [95% CI, −0.212 to −0.069]). A small portion was indirect, mediated by higher estimated glucose disposal rate (eGDR, a marker of insulin sensitivity; β=−0.004 [95% CI, −0.007 to −0.001]; P=0.02), which accounted for 2.6% (95% CI, 0.7%–6.3%) of the total association, indicating ancillary mechanism (Figure 6A). Concurrently, decelerated biological aging demonstrated significant explanatory contribution (P=0.01). The overall mortality benefit (β=−0.143 [95% CI,:−0.207 to −0.075]; P<0.001) comprised both substantial direct effects (β=−0.140 [95% CI, −0.206 to −0.072]) and indirect effects through this pathway (β=−0.006 [95% CI, −0.010 to −0.002]), representing 4.0% (95% CI, 1.1%–9.5%) of the protective association (Figure 6B).

Figure 6. Mediation effects of composite indicator on the relationships between exposures with all‐cause mortality in the entire population with CKM syndrome.

CKM indicates cardiovascular‐kidney‐metabolic; DE, direct effect; and IE, indirect effect.

Sensitivity Analyses

The robustness of the associations between dietary n‐3 PUFA intake and CKM disease progression was confirmed through 5 sensitivity analyses: multiple imputation for missing covariates, adherence to stringent CKM criteria, exclusion of cases of premature mortality (<2 years), removal of users of SGLT2 inhibitors/GLP‐1 RA, and elimination of cases with incomplete covariate or mediator data (Figure 1B through 1E, Tables S6 through S10). Notably, cohort‐level associations related to all‐cause mortality were consistently observed across 3 distinct analytical scenarios (Figures S10 through S15). Among participants in stages 1/2, the findings demonstrated robustness across all specified conditions (Figures S7 through S11).

In sensitivity analyses using energy‐adjustment methods, both approaches produced results aligning with primary findings (Figures S16 through S18). For CKM progression, the energy‐density method showed reduced odds in the highest quartile (OR, 0.88 [95% CI, 0.78–1.00], P=0.047), and the residual method indicated similar trends (OR, 0.89 [95% CI, 0.79–1.00], P=0.056). In mortality analyses, energy‐adjusted n‐3 PUFA intake demonstrated favorable associations across population subgroups, with substantial risk reductions in stages 1/2 participants. These consistent findings across methodological approaches reinforce the robustness of our primary results.

Furthermore, the E‐value analysis indicated that the presence of unmeasured confounding is unlikely to account for the observed inverse relationships between higher n‐3 PUFA intake (Q4 versus Q1) and mortality in stages 1/2 CKM (Table S20). To nullify the identified association with all‐cause mortality (HR, 0.75), unmeasured confounders would need to exhibit relative risks of at least 2.00 (or a minimum of 1.38 for nullification). The corresponding thresholds for noncardiovascular mortality (HR, 0.70) were found to be 2.21 and 1.56, underscoring the study’s enhanced resilience to residual confounding.

DISCUSSION

Summary of Key Findings and Theoretical Contributions

Our prospective cohort study demonstrates, for the first time, an inverse dose‐dependent association between dietary n‐3 PUFA intake and both CKM stage severity and mortality across the full CKM spectrum. This relationship is characterized by a distinct threshold at 22.35 mg/kg per day, beyond which additional intake yields no further benefit. Each 10 mg/kg per day increase in n‐3 PUFA intake was associated with a 13% reduction in advanced CKM staging (OR, 0.87 [95% CI, 0.83–0.91]), whereas individuals in the highest quartile exhibited 48% lower odds of severe CKM and 25% lower all‐cause mortality compared with those in the lowest quartile. These effects were primarily observed for noncardiovascular mortality, suggesting that n‐3 PUFAs may influence broader metabolic and inflammatory pathways beyond traditional cardiovascular mechanisms.

To delineate the components of noncardiovascular mortality, we analyzed specific cause death distributions and identified cancer as the predominant driver, accounting for 57.3% of such deaths. Subsequent analysis confirmed a robust inverse association between n‐3 PUFA intake and cancer mortality, which was most pronounced in the subgroup with CKM stages 1/2. Complementary competing‐risks analyses, accounting for the mutually exclusive nature of mortality end points, reinforced the absence of an association with cardiovascular death while validating the protection against cancer mortality in early CKM stages. Notably, these models also uncovered a significant reduction in cancer mortality risk at moderate n‐3 PUFA intake levels (Q3) within the subgroup with CKM stages 3/4—an effect attenuated in standard Cox models, likely due to the high competing risk of cardiovascular death in advanced disease. This pattern suggests that the anticancer properties of n‐3 PUFAs may persist across the CKM spectrum, highlighting a complex, stage‐dependent relationship.

Mechanistically, our mediation analyses indicate that improved glucose disposal (2.7%) and decelerated biological aging (4.0%) partially explain the observed associations with mortality, aligning with dysregulated metabolic aging pathways in CKM. ²⁶ , ⁴⁹ Notably, DPA emerged as the predominant contributor to mortality risk reduction across all CKM stages (43.6%–67.7% weight), extending beyond the traditionally studied EPA and DHA. This finding highlights the underappreciated role of DPA, which may derive its efficacy from superior bioavailability and specialized proresolving mediator production. ⁵⁰ , ⁵¹ , ⁵² In contrast, EPA demonstrated stage‐specific cardiovascular protection (10.1% weight) in early CKM (stages 1–2), reinforcing findings from the REDUCE‐IT trial. ¹²

These results challenge the prevailing “one‐size‐fits‐all” approach to n‐3 PUFA supplementation and suggest the potential utility of stage‐specific dietary strategies tailored to CKM progression.

Comparison With Prior Studies and Resolution of Controversies

Our findings demonstrate a 13% reduction in the risk of advanced CKM staging per 10 mg/kg per day increase in n‐3 PUFA intake, aligning with the progressive pathophysiology from early metabolic dysfunction (stages 1–2) to multiorgan involvement (stages 3–4). These results echo the cardiovascular benefits observed in the REDUCE‐IT trial, where high‐dose EPA significantly reduced atherothrombotic events in high‐risk populations, despite neutral renal outcomes. ¹² This apparent organ‐specific efficacy likely reflects EPA’s pleiotropic actions, including triglyceride lowering, anti‐inflammatory effects, and platelet inhibition. ⁵³ Meta‐analytic evidence further supports inverse associations between n‐3 PUFA levels and risks of incident CKD, total stroke, and ischemic stroke, reinforcing the systemic impact of these fatty acids. ⁵⁴ , ⁵⁵ Our study extends these observations by identifying a nonlinear, L‐shaped dose–response relationship with all‐cause mortality, characterized by a threshold at 22.35 mg/kg per day. This aligns with emerging data suggesting that cardiovascular benefits plateau at ∼1 g/day of n‐3 PUFA intake. ⁵⁶ , ⁵⁷

Notably, we observed stronger associations among younger and more physically active individuals, suggesting synergistic interactions between n‐3 PUFAs and lifestyle factors, such as muscle metabolism and sarcopenia‐related CVD risk in CKM. ⁵⁸ This synergy aligns with evidence demonstrating n‐3 PUFA’s capacity to amplify resistance training adaptations in muscle mass and strength. ⁵⁹ , ⁶⁰ We further reconcile the well‐documented inconsistencies across major clinical trials by considering several interrelated factors. First, dose and formulation are critical; the successful REDUCE‐IT ¹² and GISSI‐P ¹⁴ trials used high‐dose EPA (4 g/day) or an n‐3 PUFA formulation (1 g/day), respectively, whereas neutral trials like OMEGA ¹⁵ used a lower 1 g/day dose of a mixed EPA/DHA formulation. The STRENGTH trial, ¹⁶ which used a high‐dose carboxylic acid formulation of EPA and DHA and was terminated for futility, suggests that not only the dose but also the specific chemical composition (ethyl ester versus carboxylic acid) and the ratio of EPA to DHA may influence efficacy. Second, differences in the study population and background therapy are pivotal. Trials like GISSI‐P enrolled patients in the prestatin era, allowing a greater margin for n‐3 PUFA benefit. In contrast, contemporary trials (OMEGA, STRENGTH, OMEMI ¹⁷ ) enrolled patients receiving comprehensive background cardioprotective therapy (eg, high‐intensity statins), which may diminish the additive relative risk reduction achievable with n‐3 PUFAs. Third, the stage of CKM syndrome may modulate response. Our data suggest benefits are most pronounced in earlier CKM stages. It is plausible that in advanced CKM (stages 3–4), underlying microvascular dysfunction or alterations in membrane transporter systems (eg, MFSD2A, CD36) impair the incorporation of supplemented n‐3 PUFAs into tissues, thereby blunting their clinical effects. ⁶¹ , ⁶² , ⁶³ Therefore, the heterogeneity in trial results likely stems from a combination of these factors rather than a single cause.

The observed inverse association between n‐3 PUFA intake and cancer mortality—particularly pronounced in early CKM stages (1–2)—may stem from interconnected biological mechanisms. n‐3 PUFAs exert direct antitumor effects by inhibiting proliferation and inducing apoptosis through modulation of cell cycle regulators and mitochondrial pathways. ⁶⁴ , ⁶⁵ They further mediate pleiotropic effects within the tumor microenvironment by suppressing proinflammatory mediators (eg, NF‐κB [nuclear factor kappa B], COX‐2 [cyclooxygenase‐2]) while enhancing antitumor immunity via CD8+ T‐cell activation and NK cell cytotoxicity. ⁶¹ , ⁶⁶ , ⁶⁷ The stage‐specific protection in early CKM may reflect enhanced modulation of oncogenic pathways within a less dysregulated metabolic and inflammatory milieu.

Importantly, our WQS analysis identifies DPA as the dominant contributor to mortality risk reduction across all CKM stages, suggesting that current dietary recommendations focusing on EPA and DHA may overlook a critical component of the n‐3 PUFA spectrum. Additionally, plant‐derived ALA and SDA demonstrated significant contributions to all‐cause mortality, challenging the prevailing emphasis on marine‐derived sources. These findings support a broader definition of n‐3 PUFA‐rich diets and suggest that CKM‐specific metabolic pathways may differentially regulate the efficacy of individual n‐3 PUFAs.

Methodological Strengths

Our study leverages a large, nationally representative sample from 10 NHANES cycles (1999–2019), with n‐3 PUFA intake normalized per kilogram of body weight to account for metabolic variability. We applied advanced statistical modeling—including restricted cubic splines, threshold analyses, and WQS regression—to delineate the dose–response and mixture effects of n‐3 PUFAs across the CKM spectrum. WQS regression provided biologically interpretable weights for individual fatty acids, identifying DPA as the dominant contributor to mortality risk reduction. Bootstrapped weights, directional constraints, and internal validation ensured robustness and reproducibility.

To evaluate the reliability of our findings, we conducted 6 prespecified sensitivity analyses: (1) multiple imputation for missing data, (2) restriction to 2011 to 2018 cycles and exclusion of participants ≥80 years, (3) exclusion of early deaths (<2 years) to mitigate reverse causality, (4) adjustment for SGLT2 inhibitors/GLP‐1RA use, (5) complete‐case reanalysis, and (6) E‐value analyses further supported the robustness of our findings, indicating that strong unmeasured confounding would be required to fully explain away the observed associations between higher n‐3 PUFA intake and associated with a reduction in mortality risk, particularly among individuals with early‐stage CKM.

Furthermore, the robustness of our mixture findings was confirmed by a sensitivity analysis using Qgcomp, which yielded consistent results regarding the prominent role of DPA despite its different modeling assumptions.

The consistency of our findings across 3 distinct exposure modeling approaches—body weight normalization, energy density, and residual methods—substantially strengthens their validity. The energy‐adjustment analyses specifically address potential confounding by total energy intake while avoiding theoretical concerns related to body mass. The concordant protective associations observed across these methodologically diverse approaches enhance confidence that our results reflect genuine biological relationships rather than methodological artifacts. This methodological triangulation provides robust evidence supporting the primary findings.

Limitations

Several limitations warrant consideration. First, residual confounding remains plausible despite extensive covariate adjustments, inherent to the observational design. Second, although the use of 24‐hour dietary recalls is a standard method in nutritional epidemiology, reliance on 1 or 2 recalls may not fully capture long‐term habitual intake, particularly for episodically consumed foods that are rich in n‐3 PUFAs, such as fatty fish. This measurement error could potentially attenuate the observed effect sizes toward the null. However, the use of the mean from 2 recalls for most participants, the availability of 30‐day fish consumption data for context, and the large sample size help to stabilize the population‐level estimates. Third, although the use of sampling weights ensures that our findings are representative of the noninstitutionalized US adult population with CKM syndrome, the generalizability of our results to populations in other health care systems or countries, or to individuals without CKM syndrome, may be limited. Fourth, the CKM staging framework, although based on American Heart Association criteria, is complex, and its application to NHANES data requires careful operationalization. Although this method has been used in prior studies, any misclassification of CKM stage is possible. We anticipate that such misclassification would most likely be nondifferential with respect to n‐3 PUFA intake. In this case, the effect would be to bias our observed associations toward the null, suggesting that the true protective effects of n‐3 PUFAs may be stronger than those we reported. Fifth, baseline‐only mediator assessments preclude causal pathway validation, necessitating longitudinal studies for mechanistic confirmation. Sixth, biomarker validation through phospholipid measurements would strengthen dietary recall data. Finally, although our WQS and Qgcomp analyses consistently identified DPA as a key contributor to mortality risk reduction, we acknowledge the limitations of interpreting mixture model weights as direct causal evidence. The high weight for DPA, despite its low absolute dietary intake, could reflect its high biological potency or specific metabolic pathways, but it may also be influenced by factors such as differential measurement error or scaling. The weights represent the relative contribution of each fatty acid within the mixture model and should be interpreted as indicators of relative importance in the context of the observed data, rather than as definitive measures of biological effect size. Future research should (1) validate findings using randomized controlled trials with plasma phospholipid profiling, (2) investigate epigenetic regulation of FADS gene cluster in CKM progression, and (3) develop stage‐specific n3‐PUFA efficacy thresholds using pharmacokinetic modeling.

CONCLUSIONS

Higher n‐3 PUFA intake, particularly DPA and EPA, is associated with less severe CKM stage and lower mortality, with benefits most pronounced in younger, physically active individuals. Collectively, these observational findings position n‐3 PUFAs as a modifiable dietary factor of interest for potentially modifying the CKM continuum. Clinically, targeted n‐3 PUFA supplementation could complement existing therapies, especially in early‐stage CKM. From a clinical perspective, these observational data suggest that dietary n‐3 PUFA intake merits further investigation as a potential complementary factor, especially in early‐stage CKM. This hypothesis requires testing in randomized controlled trials. Mechanistically, the interplay between n‐3 PUFAs, metabolic homeostasis, and biological aging warrants further exploration to optimize personalized interventions. Future randomized controlled trials should prioritize DPA supplementation in early CKM, incorporating biomarker validation and time‐to‐event designs.

Sources of Funding

None.

Disclosures

None.

Supporting information

Supplemental Methods S1

Tables S1–S20

Figures S1–S18