Over Half of the United States Population Had an Undesirably Low Omega-3 Index Based on Erythrocyte Membrane Measurements: Results From the Cross-Sectional NHANES From August 2021 to August 2023

Abstract

The omega-3 index (O3I; defined as the sum of red blood cell concentrations of eicosapentaenoic acid and docosahexaenoic acid) is strongly correlated with the risk for sudden cardiac death. We aimed to provide nationally representative data on the prevalence of undesirable (<4%), intermediate (4%–8%), and optimal (>8%) O3I categories in the United States population aged ≥6 y and to determine associations with selected sociodemographic, lifestyle, intake, and health variables among adults. Fifty-four percent of United States persons aged ≥6 y (n = 7213) had an undesirable O3I; the mean O3I was 4.12%, and it increased with age. Among adults ≥20 y (n = 5721), the O3I was higher in females, non-Hispanic Asians, persons above the poverty line, those with education beyond high school, nonsmokers, PUFA supplement users, cholesterol-lowering drug users, and persons with a healthy weight or certain heart health conditions. These first data demonstrate that the United States population falls short of dietary recommendations required to achieve an optimal O3I.

Introduction

Among PUFA, the omega-3 (n–3) class is essential for human health and requires dietary intake because of inefficient endogenous synthesis. The protective effects of n–3 PUFA on cardiovascular health are well documented, including lowering triglycerides and modulating inflammation [[1], [2], [3], [4], [5], [6]]. Assessment of dietary intake of n–3 PUFA is insufficient to accurately reflect n–3 status [7]. Red blood cell (RBC) membrane fatty acids are a strong indicator of long-term n–3 status and reflect habitual intake [8]. Both dietary intake and RBC n–3 PUFA levels are inversely associated with the risk of cardiac arrest and acute coronary syndrome, suggesting that higher n–3 PUFA levels in cell membranes may reduce coronary artery disease (CAD) mortality [[9], [10], [11], [12], [13]]. Heart disease remains the leading cause of death among most racial and ethnic groups in the United States [14].

The omega-3 index (O3I), defined as the sum of EPA and DHA as a percentage of total fatty acids in RBC membranes [9], is a validated indicator of primary cardiac arrest risk [11] and can reflect an individual’s heart n–3 fatty acid content [15]. The O3I is typically categorized as undesirable, intermediate, and optimal, with the latter being associated with a lower risk for cardiovascular disease and type 2 diabetes [9,16,17]. Randomized clinical trials, observational, and cohort studies indicate that the mean O3I in United States adults (excluding those in Alaska) is well below the optimal category, ranging from 4.37% to 5.95% [18,19]; yet nationally representative data on RBC O3I are lacking. This report aims to evaluate the prevalence of undesirable, intermediate, and optimal O3I in a nationally representative sample of the United States population aged ≥6 y and determine the associations with sociodemographic, lifestyle, intake, and health variables in adults. These first estimates allow for the identification of subgroups with undesirable O3I.

Methods

Basic methods information is provided below, with additional information presented in Supplementary Text 1.

Study design, population, and data collection

The NHANES is a cross-sectional survey of the civilian, noninstitutionalized United States population conducted by the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention (CDC). Data are collected continuously and released in 2-y cycles to assess health and nutritional status in a nationally representative sample. Participants receive an in-home interview, followed by a health examination in a mobile examination center (MEC) where biological specimens are collected. We used data from the cycle conducted during August 2021−August 2023. Dietary intake assessment (24-h dietary recall) and 30-d dietary supplement (DS) data collection were conducted via telephone interviews 3−7 d after participants’ MEC examination. The unweighted response rates for the interview and examination components among adults were 32.9% and 25.5%, respectively [20]. Participants >18 y provided consent, parents provided consent for those <18 y, and youth aged 7–17 y assented. The NCHS Ethics Review Board approved the NHANES survey protocol.

Biomarker measurements

The MEC collected, washed, and diluted RBCs (1:1) with saline; specimen aliquots were stored and shipped frozen. The CDC laboratory measured concentrations of 21 RBC cis-fatty acids via acid/base hydrolysis, hexane extraction, pentafluorobenzyl bromide derivatization, and negative chemical ionization GC-MS detection [21]. Individual fatty acid molar concentrations were calculated from multipoint calibration curves including 12 isotopically labeled internal standards and converted to weight concentrations (mg/L); these were used to calculate weight percentages of each RBC fatty acid relative to total fatty acids. All fatty acid concentrations, including those reported as below the limit of detection (LOD), were included in the calculation of total fatty acids so that percentage values are based on the complete fatty acid profile. If a specific fatty acid had a concentration below the LOD, its individual percentage was reported as missing.

Study variables

We categorized demographic variables as follows: 5 age groups (6–11, 12–19, 20–39, 40–59, and ≥60 y) and adults ≥20 y, sex (males and females), and 4 race and Hispanic origin groups (all Hispanic, non-Hispanic Asian [NHA], non-Hispanic Black, and non-Hispanic White). Among adults ≥20 y, we examined sociodemographic, lifestyle, and health risk variables previously shown to be associated with RBC fatty acids and the O3I [17]: family poverty income ratio (PIR) (below the poverty line, PIR <1; yes/no), education level (≤high school and >high school), BMI (kg/m²; underweight <18.5; healthy weight 18.5–24.99; overweight 25–29.99; and obesity ≥30), dietary intake (quartiles based on total PUFA intake), and total PUFA DS use during the previous 30 d. Additional variables, such as recent smoking (past 5 d), any DS use, blood pressure, heart disease, blood cholesterol, cholesterol-lowering medication use, diabetes, and prediabetes, were classified using questionnaire data (yes/no).

Statistical analyses

Statistical analyses incorporated the complex multistage sampling design of the NHANES. Two-year phlebotomy weights were applied to account for unequal selection probabilities and survey nonresponse. Weighted arithmetic means, percentiles, and corresponding 95% confidence intervals were used to report weight percentages of fatty acids and O3I. Wald F tests were used to assess differences among the weighted means across selected covariates.

We categorized the O3I based on CAD risk zones as undesirable (<4%, high CAD risk), intermediate (4%–8%), and optimal (>8%, low CAD risk) [9,16,17] and calculated weighted prevalence estimates across these categories overall and for selected covariates. Differences in the distribution of O3I categories across selected covariates were assessed using the Rao–Scott χ² test of independence.

To identify factors independently associated with DHA, EPA, and O3I among adults, we developed weighted linear regression models separately for DHA, EPA, and O3I using backward elimination and included covariates that demonstrated significant associations with either DHA or EPA in the final models. The final adjusted models included sex, age, race and Hispanic origin, education (>high school versus ≤high school), obesity status (≥30 versus <30 kg/m²), current smoking status (yes/no), cholesterol-lowering medication (yes/no), and PUFA DS use (yes/no). Because O3I is defined as the sum of DHA and EPA, the direction and magnitude of O3I associations can be approximated by the summed DHA- and EPA-specific beta coefficients. Analyses were conducted using SAS for Windows, version 9.4 (SAS Institute, Inc) and SAS-callable SUDAAN, version 11.0.4 (RTI International).

Results

Box-and-whisker plots of the distribution of 21 RBC fatty acid and O3I weight percentages for the United States population ≥6 y show increases in IQRs with increasing fatty acid weight percentages (Figure 1). However, the relative interindividual variability (relative IQR) was greater among fatty acids with smaller weight percentages (Supplementary Table 1). Variability was not consistent across lipid classes, with PUFA and MUFA showing greater heterogeneity than SFA in the United States population ≥6 y (range of relative IQR: 13%–54%, 13%–44%, and 7.9%–28%, respectively). The mean, median, IQR, and reference intervals for 21 RBC fatty acids and O3I are shown in Supplementary Table 1. The central 95% reference interval (2.5th and 97.5th percentiles) for O3I for the United States population ≥6 y and ≥20 y was 1.56%–7.79% and 1.53%–8.00%, respectively, with the upper end reaching the optimal O3I category. The mean O3I in the United States population ≥20 y was 4.25%, at the low end of the intermediate category, whereas in children 6–11 y (3.61%) and adolescents 12–19 y (3.60%), it was in the undesirable category (Supplementary Table 2).

FIGURE 1.

Distribution of fatty acid weight percentages by lipid class for the United States population ≥6 y, NHANES August 2021–August 2023. Fatty acids are sorted by their relative abundance (weight %) and color coded by their lipid class [MUFA in orange, PUFA) in purple, SFA) in green]. Each box represents the interquartile range with the median indicated by the central line and the mean indicated by an “x”; whiskers extend the full range of the data. Boxes around the O3I, DHA, and EPA box-and-whisker plots indicate the fatty acids included in the O3I calculation and the product of the calculation. ALN, α-linoleic acid; ARA, arachidonic acid; AR1, arachidic acid; DA1, docosanoic acid; DP3, docosapentaenoic acid (n–3); DP6, docosapentaenoic acid (n–6); DTA, docosatetraenoic acid; ED1, eicosadienoic acid; EN1, eicosenoic acid; GLA, γ-linoleic acid; HGL, homo-γ-linoleic acid; LNA, linoleic acid; LG1, lignoceric acid; MR1, myristic acid; NR1, nervonic acid; OL1, oleic acid; O3I, omega-3 index; PL1, palmitoleic acid; PM1, palmitic acid; ST1, stearic acid.

Overall, 97.9% of the United States population ≥6 y and 97.6% of adults had an O3I <8% (Supplementary Table 3). All prevalence estimates in the optimal category (>8%) were suppressed, as they were considered unreliable due to small sample sizes. Among all age groups, persons ≥60 y had an intermediate O3I prevalence, ∼13 to 31 points higher than the other age groups. Among adults, females and NHA had an intermediate O3I prevalence, ∼9 to 22 points higher compared with their respective counterparts. Adults below the poverty line, those with a high school education or less, and current smokers had an undesirable O3I prevalence 11 to 27 points higher than their respective counterparts. Adults taking PUFA-containing DS had ∼33 points higher intermediate O3I prevalence compared with those not taking a PUFA-containing DS, whereas adults consuming any DS showed a 17-point difference with non-DS users. Adults with self-reported coronary heart disease, those with high cholesterol, and those who were told they had prediabetes showed an intermediate O3I prevalence ∼10 points higher than their respective counterparts.

Unadjusted bivariate means for DHA, EPA, and O3I among adults showed significant group differences for all factors, except for diabetes status (Supplementary Table 2). We assessed which factors were independently associated with DHA, EPA, and O3I and visualized the decomposition of the DHA and EPA beta coefficients from regression analysis to illustrate the relative contribution of each component to the O3I, showing only the final selected variables in the model (Figure 2). After adjusting for all variables, PUFA DS use remained the strongest predictor of higher O3I. Conversely, smoking and obesity were associated with lower O3I.

FIGURE 2.

Independent associations of selected covariates with DHA and EPA and their contributions to the O3I based on covariates retained in the final DHA- and EPA-specific models among adults ≥20 y. Because O3I is defined as the sum of DHA and EPA, stacked bars illustrate the decomposition of DHA- and EPA-specific beta coefficients and their combined contribution to the O3I. Error bars represent the 95% CI for O3I. Chol-meds, cholesterol-lowering medication; DS, dietary supplement; NHA, non-Hispanic Asian; NHB, non-Hispanic Black; NHW, non-Hispanic White (non-Hispanic Other persons not analyzed separately but included in other subgroup comparisons); O3I, omega-3 index. Source: NHANES August 2021–August 2023.

Discussion

Most O3I research is based on observational cohorts, randomized controlled trials, and clinical samples from Asia (Japan and South Korea), Europe (Italy, Spain, Germany, and the United Kingdom), North America (United States), and the Middle East (Palestinian territories), with only 1 national cross-sectional survey (the Canadian Health Measures Survey) [13,16,19,[22], [23], [24]]. Similar to the Canadian population, nearly 98% of the United States adult population had an O3I in the undesirable and intermediate categories. The United States mean O3I among adults (4.25%) was similar to that of the Canadian adult population (4.5% [17]) and to that from United States adult randomized controlled trials, observational studies, and cohort studies (4.37%–5.54% [13,19]).

Intakes of fish and DS high in n–3 PUFA have a direct impact on O3I and CAD mortality [1,[22], [23], [24], [25]]. Intakes of EPA and DHA required to bring about increases in blood concentrations that may be heart protective are believed to be achievable in the general population [1]. The American Heart Association recommends 2 servings of fatty fish/wk for individuals without CAD and 1 g/d of EPA and DHA for those with established CAD [26]. During the August 2021–August 2023 survey, the intake of seafood at least twice/wk was only 24.3% for United States adults and showed a positive correlation with increasing family income [27]. Unsurprisingly, factors typically associated with a healthy lifestyle or higher socioeconomic status (i.e., DS use and education beyond high school) were associated with higher O3I, whereas smoking and obesity, independent CAD risk factors, showed negative associations after adjusting for covariates consistent with findings reported in the International Society for the Study of Fatty Acids and Lipids statement [7]. This is of public health significance, as ∼40% of United States adults have obesity [28], and 16 million Americans live with a disease caused by smoking [29].

The direct measurement of fatty acids in erythrocyte membranes is a preferred approach to assessing O3I and a major strength of our study [7,8,16]. However, equations are available to convert plasma, serum, and whole blood fatty acid data to RBC equivalents [19]. Although we examined the association between total PUFA dietary intake and RBC fatty acids, dietary intake data on fatty acid sources are available to further explore this relationship. Future analyses can also assess the relationship between RBC fatty acids and health outcomes reported in NHANES. Our data are limited to a single survey cycle, preventing us from presenting findings for the optimal O3I category due to small sample sizes. Declining response rates may increase the risk of nonresponse bias, although no major sources of bias were found in an analysis of the August 2021–August 2023 cycle [30].

These first nationally representative RBC fatty acids and O3I profiles provide baseline data for the United States population that inform public health strategies, guide nutrition policy, and lay the groundwork for future research on the role of fatty acid status in population health and disease prevention. Our findings demonstrate that the United States population falls short of dietary recommendations to achieve an optimal O3I.

Author contributions

The authors’ responsibilities were as follows—CDP and MRS: developed the study concept and design; MRS: performed the statistical data analysis; CDP, EMM, and CMP: conducted data interpretation; CDP and EMM: wrote the initial draft, which was modified based on critical review from all coauthors; CDP: has primary responsibility for all content; and all authors: read and approved the final manuscript.

Data availability

The NHANES data are publicly available at https://wwwn.cdc.gov/nchs/nhanes/default.aspx.

Declaration of Generative AI and AI-Assisted Technologies in the Writing Process

No generative AI and AI-assisted technologies were used in the writing process.

Funding

The authors reported no funding received for this study.

Conflict of interest

The authors report no conflict of interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article: