For our study, we specifically examined data collected between 2005 and 2014. Subjects were excluded from our study for the following reasons: (1) missing data on the frailty index (FI); (2) missing data on ω-3 and ω-6 intake; (3) aged under 45 years old (Non-Elderly Population); (4) missing covariate data (such as hypertension, hyperlipidemia, diabetes, etc.) (Figure 1).

Building upon the standards for frailty established by Searle (18), Hakeem (19) further expanded the FI to include 49 deficits, encompassing various systems. These systems include cognitive function (related to confusion and memory issues), dependence (difficulties with activities of daily living), depression (assessed with the Patient Health Questionnaire-9), comorbidities (various chronic diseases), hospital utilization, and self-rated health status (healthcare utilization frequency and prescription medication counts compared to the past year), physical performance and anthropometric measurements (grip strength and body mass index), and laboratory values (including complete blood counts and blood glucose levels). The FI is expressed as the number of deficits acquired by participants divided by the total potential deficits (with a numerical range from 0–1) (Supplementary Table S1). For example, an individual with 10 deficits would have an FI score of 0.20 (10/49 = 0.20). Consistent with previous research, we categorized individuals as “robust” (FI ≤ 0.21) and “frail” (FI > 0.21) (19).

Dietary ω-3 and ω-6 intake was based on data obtained from 24-h dietary recall interviews. The primary dietary interview was conducted by trained interviewers at the Mobile Examination Center (MEC) using an automated data collection system. Detailed information on data processing procedures can be found in the NHANES website’s dietary interview component [The examination protocol and data collection methods. https://wwwn.cdc.gov/nchs/data/nhanes/2011-2012/manuals/mec_in_person_dietary_procedures_manual_jan_2012.pdf].

In this study, we aimed to capture as many components of ω-3 and ω-6 as possible. Dietary ω-3 included alpha-linolenic acid (ALA, 18:3), EPA (20:5), docosapentaenoic acid (DPA, 22:5), and docosahexaenoic acid (DHA, 22:6), in addition to other forms [stearidonic acid (SDA), 18:4]. Dietary ω-6 included linoleic acid (LA, 18:2) and arachidonic acid (AA, 20:4) (20). Daily average intake of ω-3 and ω-6 was calculated from dietary intake based on the USDA Food and Nutrient Database for Dietary Studies, and only dietary intake of these fatty acids was considered, with supplements not being taken into account [U.S. Department of Agriculture, Agricultural Research Service. Usda Food and Nutrient Database for Dietary Studies. https://www.ars.usda.gov/northeast-area/beltsville-md-bhnrc/beltsville-human-nutrition-research-center/food-surveys-research-group/docs/wweianhanes-overview/].

Furthermore, the low-fat dietary pattern is defined based on the total lipid intake and total calorie intake of the respondents. Specifically, this dietary pattern is characterized by a total lipid intake of less than or equal to 30% of the daily intake. Since the NHANES questionnaire does not provide a detailed breakdown of all lipid energies, we used a fat energy coefficient for conversion, where approximately 1 g of fat intake equals 9 kcal.

Covariates included age, sex, race/ethnicity (Mexican American, other Hispanic, non-Hispanic white, non-Hispanic black, non-Hispanic Asian, other race), BMI, marital status, smoking status, cotinine, hypertension, hyperlipidemia, diabetes and coronary artery disease (CVD). BMI was divided into three categories: Normal (<25 kg/m²), Overweight (≥25 kg/m², <30 kg/m²), and Obese (≥30 kg/m²). Smoking status was classified as Current Smoker (Defined as having smoked more than 100 cigarettes in a lifetime and still smoking), Former Smoker (Defined as having smoked more than 100 cigarettes in a lifetime but no longer smoking), Never smoke (defined as smoking less than 100 cigarettes in their lifetime). Hypertension was defined according to the American Heart Association/American College of Cardiology (AHA/ACC) 2017 guidelines as systolic blood pressure ≥ 130 mmHg or diastolic blood pressure ≥ 80 mmHg and self-reported diagnosis or use of antihypertensive medication. As per the guidelines set by the Adult Treatment Panel III (ATP 3) of the National Cholesterol Education Program (NCEP), hyperlipidemia is defined by the following criteria: total cholesterol levels equal to or exceeding 200 mg/dL, triglyceride levels equal to or exceeding 150 mg/dL, HDL cholesterol levels below 40 mg/dL for men and below 50 mg/dL for women, or LDL cholesterol levels equal to or exceeding 130 mg/dL (21). Diabete as any of the following: (1) HbA1c levels equal to or greater than 6.5%; (2) serum glucose levels exceeding 200 mg/dL at 2 h after a 75 g glucose load (OGTT); (3) fasting glucose levels equal to or greater than 126 mg/dL; (4) self-reported diagnosis of diabetes; (5) self-reported use of insulin or other diabetes medication. The duration of diabetes was determined by subtracting the participant’s current age from the self-reported age at diagnosis, or zero for individuals diagnosed during the NHANES examination. For CVD, a positive response to any of the following questions was defined as CVD: “Has a doctor or other health professional ever told you that you have congestive heart failure (CHF)/coronary heart disease (CHD)/angina/heart attack/stroke?”

Considering drugs affecting lipid metabolism, we defined lipid-lowering drugs and drug categories (such as statins, fibric acid derivatives, ezetimibe, cholesterol absorption inhibitors, and PCSK9 inhibitors, etc.) using the Multum Lexicon standardized drug codes or therapeutic classification schemes. Specifically, we confirmed the use of prescription drugs in the past month, with primary treatment category drugs classified as “metabolism modifiers” (code “358”) and secondary treatment drug category classified as anti-hyperlipidemic agents (code “19”), and associated them with respondents’ individual identification symbols “SEQN” through the unique identifier “RXDDRGID” (https://wwwn.cdc.gov/Nchs/Nhanes/1999-2000/RXQ_DRUG.htm#Component_Description).

A complex sampling design was implemented to ensure nationally representative estimates. All analyses were adjusted for survey design and weighted variables, with new sample weights calculated as the original 2-year sample weights divided by 2. Dietary ω-3 and ω-6 fatty acids intake were categorized into tertiles (with 1/3, 2/3, and 1 as cut-off points). The ω-6: ω-3 ratio was divided into four groups: recommended (≤4), mildly high (>4, ≤10), high (>10, ≤15), and very high (>15). Continuous variables were expressed as mean ± standard deviation (SD), while categorical variables were presented as counts (N) and percentages (%). Weighted t-tests (for continuous variables) and weighted chi-square tests (for categorical variables) were used to assess differences between robust and frailty subjects. Kruskal-Wallis tests (for continuous variables) or weighted chi-square tests (for categorical variables) were employed to evaluate differences among the three groups based on omega intake.

Initially, a crude model was fitted, followed by stepwise adjustment for covariates. Model 1 adjusted for age, sex, and race; Model 2 further adjusted for BMI, marital status, serum cotinine, and smoking status based on Model 1; Model 3 additionally adjusted for hyperlipidemia, hypertension, diabetes, and CVD based on Model 2. Results were presented as odds ratios (OR) with corresponding 95% confidence intervals (95% CI). Subgroup analyses were conducted for significant results. Furthermore, a logistic regression model was used to assess the significance of the interaction between omega intake and covariates on frailty. Regression models and restricted cubic splines (RCS) flexibly model the relationship between independent and dependent variables, especially in regression analysis. When the relationship between independent and dependent variables is not a simple linear one, RCS can help capture this non-linear relationship. They allow researchers to approximate the relationship using different polynomial functions within different ranges of independent variables, thereby providing a more accurate description of the data’s trend. RCS with three knots, at the 10^th, 50^th, and 90^th percentages, were used to explore the non-linear relationships of Omega intake levels and frailty in the linear terms model.

In addition, the risk magnitude of all variables on frailty was evaluated by constructing a nomogram. Subsequently, calibration curves were plotted to assess the fit between the predicted probabilities from the nomogram and the actual proportions. To further evaluate the sensitivity of the constructed model, the performance of the model was assessed using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve. Decision curve analysis (DCA) was employed to estimate the net benefit at different threshold probabilities, determining the clinical utility of the model (22). Statistical significance was considered at a p-value <0.05, and all reported probability tests were two-sided.

In this study, a total of 12,315 individuals were finally included. Based on the exclusion criteria, 3,568 participants were classified as “Frail.” Compared to the “Robust” group, which consisted of 8,747 individuals, the “Frail” individuals were older, had a higher proportion of females, obesity, divorced individuals, and smokers (smoking history population). Additionally, they had a higher prevalence of underlying conditions such as hypertension and hyperlipidemia. Their intake of ω-3 was notably inadequate, and their intake of ω-6 was slightly lower. However, the ω-6:ω-3 ratio was significantly high at 10.1, indicating a deficiency of ω-3 intake and an imbalance in the ω-6 to ω-3 ratio compared to the recommended values in the human evolutionary diet (Table 1). These findings align with the current health challenges faced by frail individuals, especially in terms of their dietary habits. These risk factors are visualized through a nomogram (Figure 2A). The daily intake of ω-3 was divided into three equal parts using the tertiles method: T1 (≤1.175 g/day), T2 (>1.175, ≤2.050 g/day), and T3 (>2.050 g/day). Similarly, the daily intake of ω-6 was divided into three parts: T1 (≤11.423 g/day), T2 (>11.423, ≤19.160 g/day), and T3 (>19.160 g/day). It was observed that female individuals, divorced individuals, smokers, and those with hypertension were less attentive to ω-3 and ω-6 intake. In contrast, individuals with normal BMI did not pay much attention to the intake of polyunsaturated fatty acids. Additionally, we also investigated the use of ω-3 dietary supplements, with approximately 5.3% of respondents reporting their intake. There was no statistically significant difference in the intake of supplements between the frail and robust groups, indicating that supplements are unlikely to have influenced the results of this study. It’s interesting to note that an increased intake of ω-3 fatty acids often goes hand in hand with an increased intake of ω-6 fatty acids, and the proportion of vulnerable populations is also on the decline. When grouping individuals based on the ω-6: ω-3 ratio, there was no significant difference in the proportion of frail individuals in these groups, and most variables showed no significant differences (Supplementary Tables S2–S5).

Through the construction of multiple linear regression models, we found that for ω-3, using the daily intake level T1 (≤1.175 g/day) as the reference, at the T3 level of ω-3 intake (>2.050 g/day), there was a significant negative correlation with the risk of frailty in all models (p < 0.05). This result remained robust even after adjusting for covariates, especially after adjusting for variables such as age, gender, race, marital status, serum cotinine, smoking, hypertension, hyperlipidemia, CVD, and diabetes (Model 3). The risk reduction was up to 17% (OR: 0.83, 95% CI: 0.70–0.99, p = 0.035). Similarly, for ω-6, using the daily intake level T1 (≤11.423 g/day) as the reference, at the T2 level of ω-6 intake (>11.423, ≤19.160 g/day), there was a significant negative correlation with the risk of frailty in all models. In Model 3, the risk reduction was up to 14% (OR: 0.86, 95% CI: 0.75–1.00, p = 0.044). These results showed little variation across different models (Table 2).

The ROC curves for ω-3 and ω-6 in Model 3 both had an AUC of 0.80, indicating that the constructed Model 3 had good predictive capability (Figures 2B,C). However, no significant association was found between the ω-6: ω-3 ratio and the reduction in frailty risk. Even when using a reference ratio of 4, the OR values in all models remained greater than 1, suggesting that increasing the ratio between the two is beneficial for reducing frailty development, though not reaching statistical significance. Further interaction analysis was performed on the covariates that showed significant differences between the healthy and frail groups in Table 1. Although there were differences between some subgroups, particularly in males, individuals with high BMI, non-smokers, low-fat intake diet, those without hypertension, diabetes, and those with cardiovascular diseases, these covariates showed no interaction with omega intake regarding frailty risk (p > 0.05). This suggests that the influence of covariates on the results of this study was minimal (Table 3). The decision curve analysis (DCA) showed that the net benefit probability for both ω-3 and ω-6 models in Model 3 ranged from 0 to 90%. The net benefit probability consistently remained higher than the group that received no intervention (depicted by the black solid line “None” in the graphs). This indicates that implementing the conditions in Model 3 (higher ω-3 intake and moderate ω-6 intake) is beneficial in reducing frailty risk without causing any side effects (Figures 3A,B). Furthermore, we used the RCS model to fit the relationship between ω-3 and ω-6 and frailty. After adjusting for covariates, a non-linear relationship was observed (p = 0.0085 for ω-3, p = 0.0006 for ω-6) (Figures 3C,D). To further assess the impact of drugs affecting lipid metabolism (primarily some lipid-lowering drugs) on this study, we conducted a sensitivity analysis by further grouping the population using these drugs. The results showed that in this study, there were 4,437 respondents with records of using these drugs in the past month. They still maintained a lower intake of ω-3 in the frail population but had a higher intake of ω-6 (Supplementary Table S6). After adjusting for multiple models, we still maintained the conclusion that a high level of ω-3 and moderate intake of ω-6 are associated with a reduced risk of frailty in middle-aged and elderly individuals (Supplementary Table S7).