Correlation of the Serum Fatty Acids with Cognitive Function: An NHANES 2011–2014 and Multivariate Mendelian Randomization Analysis

Abstract

Background:

The relationship between serum fatty acids and cognitive function has been the subject of extensive study.

Objective:

To analyze the relationship between serum fatty acids composition and cognitive function by NHANES database and multivariate Mendelian randomization (MR) analysis.

Methods:

A sub-cohort of 1,339 individuals with serum fatty acids and Digit Symbol Substitution Test (DSST) examinations from the 2011–2014 wave of the NHANES were analyzed using fully adjusted multiple linear regression models for associations between serum hydrolyzed fatty acid levels and cognitive function. Univariable and multivariable MR was used to analyze the correlation between 98 exposures related to serum fatty acids and cognitive function. Results from different database sources were combined using meta-analysis.

Results:

The fully adjusted regression analysis showed that linoleic acid (LA), Omega 6, fatty acids (FAs), and LA/FAs were positively correlated with DSST. 27 exposures were included for univariate MR analysis. Ultimately, only 2 traits had IVW test p-values ranging between 0.0019 and 0.05, both of which were LA/FAs. The meta-analysis of univariate MR revealed that LA/FAs was positively associated with cognitive function (β: 0.040, 95% CI = 0.013–0.067, p = 0.0041). In multivariate MR analysis, after adjusting for education, ischemic stroke, and age, LA/FAs was positively independently associated with cognitive function (IVW β: 0.049, 95% CI = 0.021–0.077, p = 0.0006). The results of MVMR are well in line with the univariate results.

Conclusions:

Both the Cross-sectional observational analyses and MR-based studies supported a suggestive causal relationship between the serum ratio of Linoleic acid in fatty acids and cognitive function.

Keywords

Alzheimer’s disease cognitive function fatty acids linoleic acid Mendelian randomization NHANES

INTRODUCTION

The aging global population presents a pressing challenge to public health systems, emphasizing the critical importance of preserving cognitive function in older adults.1,2, 1,2 Cognitive decline, including conditions such as Alzheimer’s disease, not only affects individuals but also places a substantial burden on healthcare systems and families worldwide. 3 As such, understanding the factors that influence cognitive health in aging populations is of paramount importance.

Fatty acids, particularly omega-3 and omega-6 fatty acids, are essential components of the human diet and play diverse physiological roles, extending beyond mere energy storage. 4 These fatty acids are integral to brain health and cognitive function. They are crucial for the structure and function of neuronal membranes, synaptic plasticity, neurotransmitter synthesis, and the regulation of neuroinflammation. 5 The imbalance between omega-3 and omega-6 fatty acids, often observed in modern diets, has been associated with increased inflammation, oxidative stress, and impaired cognitive function. 6 Conversely, studies have highlighted the beneficial effects of omega-3 fatty acids, such as eicosapentaenoic acid and docosahexaenoic acid (DHA), in supporting cognitive performance, reducing the risk of cognitive decline,7 –9 and potentially mitigating the onset of neurodegenerative diseases.10,11, 10,11

Despite the recognized physiological roles of fatty acids and the importance of cognitive function in aging populations, there is a gap in understanding the causal relationship between serum fatty acid levels and cognitive health. This gap presents an opportunity for investigation, particularly through the use of large-scale population-based surveys such as the National Health and Nutrition Examination Survey (NHANES). NHANES, a program of the National Center for Health Statistics, provides comprehensive health and nutrition data from a representative sample of the U.S. population. Leveraging NHANES data, this study aims to explore the association between serum fatty acid levels and cognitive function in an aging population. Moreover, the principles of Mendelian randomization (MR) offer a unique approach to address causal questions about how modifiable exposures influence different outcomes. 12 MR uses genetic variation as a natural experiment to investigate the causal relations between potentially modifiable risk factors and health outcomes in observational data.

By combining NHANES database analysis with MR research, this study seeks to contribute to the growing body of knowledge on the interplay between fatty acids and cognitive health, with a specific focus on aging populations. The findings from this research could have significant implications for developing strategies to support cognitive health and prevent cognitive decline in older adults.

MATERIALS AND METHODS

Study population

A cross-sectional study was conducted on the adult participants of the 2011–2014 National Health and Nutrition Examination Survey (NHANES). NHANES examines the health and nutritional status of a representative group of children and adults in the USA by conducting household interviews, laboratory and physical examinations.

Cognitive function assessment

This study used Digit Symbol Substitution Test (DSST) to assess cognitive function, DSST relies on processing speed, sustained attention, and working memory. 13 The DSST has been used in large screenings, epidemiological and clinical studies, and was administered during the household interview to participants 60 years and over during NHANES 1999–2002. The exercise is conducted using a paper form that has a key at the top containing 9 numbers paired with symbols. Participants have 2 minutes to copy the corresponding symbols in the 133 boxes that adjoin the numbers. The range of DSST was 0 to 105, lower test scores indicate poorer cognitive function.

Determination of fatty acids

Serum samples in NHANES were processed, stored and shipped to the Division of Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA for testing. The esterified fatty acids are hydrolyzed primarily from triglycerides, phospholipids and cholesteryl esters using sequential treatment with mineral acid and base in the presence of heat. Total fatty acids are hexane-extracted from the matrix (100μL serum or plasma) along with an internal standard solution containing eighteen stable isotopically-labeled fatty acids to account for recovery. The extract is derivatized with pentafluorobenzyl bromide in the presence of triethylamine to form pentafluorobenzyl esters. The reaction mixture is injected onto a capillary gas chromatograph column to resolve individual fatty acids of interest from other matrix constituents. Fatty acids are detected using electron capture negative-ion mass spectrometry within 34 minutes. Eleven saturated, six monounsaturated, and thirteen polyunsaturated fatty acids (thirty fatty acids in total) are measured using selected ion monitoring. 14 Quantitation is accomplished by comparing the peak area of the analyte in the unknown with the peak area of a known amount in a calibrator solution. Calculations are corrected based on the peak area of the internal standard in the unknown compared with the peak area of the internal standard in the calibrator solution. (The specific method for testing fatty acids can be found at the following link. https://wwwn.cdc.gov/Nchs/Nhanes/2013-2014/FAS_H.htm).

Selection of covariates

We incorporated the covariates connected to cognitive function from earlier research and excluded collinearity issues, including age (years), gender (male/female), education level (below high school/high school/above high school), marital status (cohabitation: married/living with someone/solitude: widowed/divorced/separated/unmarried), BMI (body mass index, kg/m²), smoking status (yes/no), alcohol consumption (yes/no), stroke (yes/no) and diabetes (yes/no). All information was acquired by standardized questioning, physical examination and laboratory tests, and questionnaires were provided by qualified staff in the medical field.

Statistical analysis

All continuous variables are reported in the form of mean and standard deviation, while categorical variables are reported in the form of numbers and percentages. All analyses were conducted using a multivariate linear regression model to investigate the association between cognitive function as the independent variable and the dependent variable. Four models were used in the linear regression analysis, in model 1 no covariates were adjusted; age, education, gender and stroke were adjusted in model 2; age, education, gender and stroke were adjusted in model 2; and age, education, gender, stroke, hba1c, 25-hydroxyvitamin d, LDL-c were adjusted in model 3. Unadjusted and adjusted associations are reported in the form of regression coefficients and their corresponding two-tailed 95% confidence intervals (95% CI). When estimating adjusted associations, confounding variables pre-determined based on biological plausibility and evidence were considered. Interaction terms were included in the model to account for potential effect modifications. Stepwise analysis was avoided, and instead, all a priori determined covariates were included. In all regression analyses, the complex survey design of NHANES was considered, and survey weights, primary sampling units, and strata were incorporated into the survey design using the survey package in R version 4.1.3.

MR research design

We used a two-sample MR approach to analyze the causal relationship between serum fatty acid levels and cognitive function. First, we assessed the causal relationship between all fatty acid-related exposures and cognitive function and identified statistically significant exposures. Next, we used a multivariate Mendelian randomization study (MVMR) to investigate the effects of fatty acid-related exposures on cognitive function after multivariate adjustment. Finally, we performed meta-analysis on the results of the same exposure factor from different database. Figure 1 outlines the key steps of the MR study investigating the interplay between fatty acids and cognitive function.

Fig. 1

Flow chart of participants selection and Mendelian randomization analysis. NHANES, National Health and Nutrition Examination Survey; MR, Mendelian randomization analysis.

Data source and tool variable selection

A total of 98 exposure files involving fatty acids were identified in the IEU OpenGWAS, FinnGen and UKbiobank databases (refer to Supplementary Table 1). The outcome GWAS files ebi-a-GCST006572 were acquired from the GWAS catalog. Criteria for selection of instrumental variables were (1) genome-wide significance threshold p < 5×10^–8; (2) exclusion of SNPs with chain disequilibrium r² > 0.001 over a window size of 5,000 kb; and (3) assessment of associations of SNPs with exposure factors using the F-statistic (F = b²/se²), whereby SNPs with a low statistic would be excluded F < 10.15 –17 no proxy SNPs were used in the MR analysis.

MR and meta statistical analysis

The random-effects inverse-variance-weighted (IVW) method was used as the primary analysis method for univariable Mendelian Randomization (UVMR), also 4 additional MR methods were employed to enhance result robustness: Simple mode, MR-Egger, Weighted median, and Weighted mode. 18 MR-PRESSO was used to detect outliers. If outliers were identified, the analysis was re-conducted after their removal. Subsequently, a leave-one-out sensitivity analysis was performed, wherein each SNP was sequentially excluded, and the IVW analysis was re-executed with the remaining SNPs to assess the influence of individual SNPs on the IVW results. MR-Egger’s intercept was employed to assess the presence of horizontal pleiotropy, with a p-value<0.05 indicating its existence. Cochran’s Q test was utilized to analyze the heterogeneity of instrumental variables. Finally, a funnel plot was employed to assess potential biases in the study results. All MVMR statistical analyses were performed using R version 4.1.3 and the TwoSample MR 0.5.10 software package.

As an extension of two-sample MR, MVMR estimates the causal effects of various risk factors on cognitive function by incorporating all exposures into a single model. To demonstrate the direct impact of serum fatty acid levels on cognitive function independent of educational attainment, ischemic stroke, and 25-Hydroxyvitamin D, and that this effect was not mediated through other exposures, significant SNPs are first extracted. These SNPs were then combined with existing exposure instrumental variables. After excluding these overlapping SNPs, the effects and corresponding standard errors of each SNP were obtained from the exposures and outcomes. Here, the multivariable IVW, MR-Egger, weighted median and least absolute shrinkage and selection operator (LASSO) analysis methods were applied to remove multiple covariate SNPs in the presence of various exposures, enhancing the accuracy of the results. 19 Four models were used in the MVMR: no covariates were adjusted in Model 1; Educational attainment was were adjusted in Model 2; Educational attainment and ischemic stroke were adjusted in Model 3; Educational attainment, ischemic stroke and 25-Hydroxyvitamin D were adjusted in Model 4.

The Bonferroni method was employed for multiple test corrections in univariable analysis. A p-value < 0.0019 (0.05/27 exposures) was considered statistically significant for causality. Given the conservative nature of the Bonferroni correction, a p-value between 0.05 and 0.0019 was also regarded as potentially causal.

MR estimates for each outcome from different sources were combined by the fixed-effects meta-analysis method. The I² statistic was calculated to assess the heterogeneity of each outcome from different data sources, and the I² values < 25%, 25–75%, and >75% were considered to indicate low, moderate, and high heterogeneity, respectively.

The MVMR (v0.4) and meta (v 7.0) within the R package (v.4.1.3) facilitated major statistical analysis and graphical representation. β and the accompanying 95% confidence interval (CI) gauged the extent of risk alteration for each additional standard deviation of exposure factors. Statistical significance was set at p < 0.05.

RESULTS

Participant characteristics

A subset of 1,339 individuals with the DSST and serum fatty acids examination from the 2011 to 2014 NHANEs datasets were included. The age range of the participants included in the study was 60 to 80 years, 655 were males while 684 were females.

The median DSST level was 46 (1∼105). Notably, participants in quartiles1 (Q1) were more likely to be elder, had a greater proportion of males, lower educational level, smoking, alcohol use, stroke, diabetes mellitus, CAD history compared to higher quartiles (Q2–Q4), and reported lower levels of 25-hydroxyvitamin D and fatty acids. The characteristics of all participants are presented in Table 1.

Table 1

Basic characteristics of participants by DSST quartile among US adults

	Q1 (DSST < 33)	Q2 (33≤DSST<46)	Q3 (46≤DSST<58)	Q4 (58≥DSST)	P
n		318	342	344	335
Age (y)		71.99 (6.81)	71.01 (6.72)	69.40 (6.37)	66.24 (5.74)	<0.001
60–70 n (%)		122 (38.4)	159 (46.5)	182 (52.9)	248 (74.0)	<0.001
70–80 n (%)		196 (61.6)	183 (53.5)	162 (47.1)	87 (26.0)
Gender n (%)	Male	167 (52.5)	201 (58.8)	165 (48.0)	122 (36.4)	<0.001
	Female	151 (47.5)	141 (41.2)	179 (52.0)	213 (63.6)
Education n (%)	Less than 9th grade	110 (34.6)	35 (10.2)	9 (2.6)	2 (0.6)	<0.001
	9–11th grade	80 (25.2)	63 (18.4)	36 (10.5)	10 (3.0)
	High school graduate/GED	69 (21.7)	102 (29.8)	86 (25.0)	64 (19.1)
	Some college or AA degree	43 (13.5)	88 (25.7)	122 (35.5)	115 (34.3)
	College graduate or above	16 (5.0)	54 (15.8)	91 (26.5)	144 (43.0)
Marriage n (%)	unmarried	13 (4.1)	16 (4.7)	20 5.8)	21 (6.3)	0.009
	married/living with someone	174 (54.7)	224 (65.5)	219 (63.7)	221 (66.0)
	widowed/divorced/separated	131 (41.2)	102 (29.8)	105 (30.5)	93 (27.8)
Smoking history n (%)	No	169 (53.1)	151 (44.2)	172 (50.0)	201 (60.0)	0.001
	Yes	149 (46.9)	191 (55.8)	172 (50.0)	134 (40.0)
CAD history n (%)	No	287 (90.3)	296 (86.5)	304 (88.4)	321 (95.8)	<0.001
	Yes	31 (9.7)	46 (13.5)	40 (11.6)	14 (4.2)
Stroke history n (%)	No	284 (89.3)	318 (93.0)	321 (93.3)	320 (95.5)	0.021
	Yes	34 (10.7)	24 (7.0)	23 (6.7)	15 (4.5)
COPD history n (%)	No	295 (92.8)	314 (91.8)	317 (92.2)	315 (94.0)	0.700
	Yes	23 (7.2)	28 (8.2)	27 (7.8)	20 (6.0)
Diabetes mellitus n (%)	No	209 (65.7)	268 (78.4)	271 (78.8)	291 (86.9)	<0.001
	Yes	109 (34.3)	74 (21.6)	73 (21.2)	44 (13.1)
Alcohol use n (%)	No	207 (65.1)	147 (43.0)	146 (42.4)	95 (28.4)	<0.001
	Yes	111 (34.9)	195 (57.0)	198 (57.6)	240 (71.6)
BMI (Kg/m²)		29.48±6.88	28.92±6.58	29.15±5.89	28.92±6.41	0.648
FBG (mmol/L)		6.76±2.43	6.44±1.95	6.18±1.41	6.09±1.56	<0.001
LDL-c(mmol/L)		2.71±0.98	2.72±1.01	2.83±1.00	2.96±1.02	0.004
TG (mmol/L)		1.37±0.83	1.31±0.76	1.39±0.80	1.40±0.86	0.528
TC (mmol/L)		4.73±1.17	4.75±1.19	4.90±1.16	5.15±1.13	<0.001
HbA1c (%)		6.39±1.42	6.04±1.04	5.90±0.93	5.84±0.81	<0.001
25-hydroxyvitamin D (nmol/L)		73.07±32.81	73.53±29.62	78.54±30.74	80.63±31.08	0.002
Uric acid (mg/dL)		341.00±96.82	338.95±95.54	337.65±91.41	328.22±87.30	0.296
LA (μmol/L)		3,416.23±1,029.13	3,467.37±908.02	3,600.49±935.74	3,756.57±1,010.72	<0.001
Omega 3 (μmol/L)		399.51±183.77	408.47±218.16	421.62±180.66	449.80±209.38	0.007
Omega 6 (μmol/L)		4,613.96±1,171.49	4,630.34±1,066.14	4,764.86±1,097.09	4,936.22±1,148.04	0.001
FA (μmol/L)		11,586.58±3,253.67	11,601.98±3,142.50	11,986.37±3,103.67	12,334.49±3,353.83	0.007
Omega 6/Omega 3		13.07±4.57	13.14±5.42	12.79±5.96	12.39±4.00	0.204
Omega 3/FAs (%)		3.47±1.35	3.49±1.35	3.56±1.37	3.69±1.57	0.184
Omega 6/FAs (%)		40.31±4.53	40.45±4.75	40.14±4.23	40.56±4.35	0.644
LA/FAs (%)		29.62±4.46	30.22±4.57	30.24±4.26	30.74±4.24	0.013

DSST, Digit Symbol Substitution Test,; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FBG, fasting blood glucose; LDL-C, low-density lipoprotein; TC, total cholesterol; TG, total triglyceride; VD25, hydroxyvitamin D; UA, uric acid; LA, linoleic acid; FAs, fatty acids.

Results of unadjusted and adjusted multiple linear regression analyses of association between serum fatty acids and cognitive function

In unadjusted regression models, linoleic acid (LA), Omega 3, Omega 6, fatty acids (FAs), Omega 6/Omega 3, Omega 3/FAs, and LA/FAs were associated with DSST (β= 0.0022, 0.0082, 0.0016, 0.0005, –0.2098, 0.7581, 0.0486, 0.3117, respectively). LA, Omega 6, FAs, and LA/FAs showed relevant associations with DSST (β= 0.0015, 0.0011, 0.0003, 0.2292, respectively), after adjusting for age, education, gender and stroke. The fully adjusted regression analysis by age, education, gender, stroke, HbA1c, 25-Hydroxyvitamin D and LDL-c showed that LA, Omega 6, FAs, and LA/FAs were significantly and positively correlated with DSST (β= 0.0017, 0.0012, 0.0004, 0.1801). The detailed results of the unadjusted and adjusted regression analyses are presented in Table 2.

Table 2

Unadjusted and adjusted regression analyses between fatty acids with Cognitive function

Model		β	95% CI	p
Model 1	LA	0.0022	0.0012∼0.0031	<0.001
	Omega 3	0.0082	0.0035∼0.0128	0.001
	Omega 6	0.0016	0.0008∼0.0024	<0.001
	FAs	0.0005	0.0002∼0.0007	0.002
	Omega 6/Omega 3	–0.2098	–0.3932∼–0.0264	0.025
	Omega 3/FAs	0.7581	0.1013∼1.4149	0.024
	Omega 6/FAs	0.0486	–0.1597∼0.2569	0.648
	LA/FAs	0.3117	0.1009∼0.5225	0.004
Model 2	LA	0.0015	0.0008∼0.0023	<0.001
	Omega 3	0.0032	–0.0005∼0.0069	0.093
	Omega 6	0.0011	0.0004∼0.0018	0.001
	FAs	0.0003	0.0001∼0.0005	0.006
	Omega 6/Omega 3	–0.0290	–0.1744∼0.1164	0.696
	Omega 3/FAs	0.0438	–0.4766∼0.5642	0.869
	Omega 6/FAs	0.0313	–0.1317∼0.1943	0.707
	LA/FAs	0.2292	0.0644∼0.394	0.007
Model 3	LA	0.0017	0.0009∼0.0025	<0.001
	Omega 3	0.0021	–0.0016∼0.0059	0.268
	Omega 6	0.0012	0.0005∼0.0019	0.001
	FAs	0.0004	0.0001∼0.0006	0.003
	Omega 6/Omega 3	0.0014	–0.1441∼0.1469	0.985
	Omega 3/FAs	–0.1655	–0.6886∼0.3577	0.535
	Omega 6/FAs	–0.0139	–0.1798∼0.1519	0.869
	LA/FAs	0.1801	0.0088∼0.3515	0.040

Model 1: no covariates were adjusted. Model 2: Age, Education, Gender and Stroke were adjusted. Model 3: Age, Education, Gender, Stroke, HbA1c, 25-Hydroxyvitamin D, LDL-c were adjusted. SE, standard error.

Univariate MR

The search yielded a total of 98 exposures related to fatty acids (involving 74 types of fatty acids and their derivative indicators), showed in Supplementary Table 1. After instrumental variable selection, there were 27 exposures were chosen for univariate MR while the vast majority of exposures were not associated with cognitive function, as indicated by p-values > 0.05 (shown in Supplementary Table 2). Figure 2 illustrates the p-values of the five MR methods for the 27 exposures. It is notable that only two exposures had p-values between 0.05 and 0.0019 calculated using the IVW method, and both of these had ratio of LA to total fatty acids (LA/FAs). The meta-analysis showed a suggestive causal relationship between the LA/FAs on cognitive function (β: 0.040, 95% CI = 0.013–0.067, p = 0.0041).

Fig. 2

The heatmap displaying significant correlations in forward MR. The illustration employs distinct colors to signify varying p values within individual blocks, transitioning from blue to red to denote ascending p values. The inside axis encompasses five distinct outcomes: Weighted mode, Simple mode, inverse variance weighted (IVW), weighted median, and MR Egger.

Multivariable MR

In the multivariable analysis, the inverse variance weighted (IVW) and weighted median methods were primarily utilized to assess the effects of correction in the confounding factors. Supplementary Table 3 presents the comprehensive estimated values and confidence intervals (Educational attainment, ischemic stroke, and 25-Hydroxyvitamin D) for the 4 models included in the multivariable analysis. All 4 models yielded MR Egger results > 0.05, signifying model stability and the absence of pleiotropic effects. The results of the multifactorial MR analysis are presented in Fig. 3, where it can be seen that there was a significant independent correlation between the presence of LA/FAs and cognitive function after correcting for education, ischemic stroke, and age (IVW β: 0.049, 95% CI = 0.021–0.077, p = 0.0006). Figure 4 illustrates the results of the meta-analysis of the four methods in Model4, showing that the MVMR results for both exposures are highly consistent across the four MVMR methods (all I² = 0%).

Fig. 3

Forest plot of univariate and multivariate IVW results on the relationship between LA/FAs and cognitive function. Model 1: no covariates were adjusted. Model 2: Educational attainment was were adjusted. Model 3: Educational attainment and ischemic stroke were adjusted. Model 4: Educational attainment, ischemic stroke and 25-Hydroxyvitamin D were adjusted. IVW inverse-variance weighted method.

Fig. 4

Forest plot of MVMR results of model 4 on the relationship between LA/FAs and cognitive function by 4 methods. IVW, inverse-variance weighted method.

DISCUSSION

Cognitive impairment is a universal health concern that is influenced by a complex interplay of biological and environmental factors. 20 While some studies have highlighted the significance of omega-3 intake in brain health, emphasizing its role as a crucial risk factor in the development and progression of certain cardiovascular and neuropathological conditions.21,22, 21,22 Two-sample MR is a powerful statistical tool that can provide additional supporting evidence for associations observed in cross-sectional studies and help alleviate potential confounding and biases inherent in such studies. Furthermore, two-sample MR plays a crucial role in confirming causal relationships and has been widely applied in biological and clinical research. 23 The core objective of this study is to investigate the interrelationship between circulating fatty acid levels and cognitive function through the analysis of NHANES data and MR analysis, aiming to uncover potential mechanisms and provide new evidence for improving cognitive function. In this study, we analyzed the association between serum fatty acids and cognitive function, taking into account common risk factors and demographic characteristics.

Results from the NHANES study indicate that after adjusting for covariates related to cognitive function such as age, education, gender, stroke, as well as common serum markers (HbA1c, 25-Hydroxyvitamin D, and LDL-c), LA, Omega 6, FAs, and LA/FAs show a high degree of independent correlation with cognitive function. However, in conducting the MR analysis, the univariate MR analysis results revealed that only LA/FA exhibited a potential causal relationship with cognitive function, while the vast majority of other fatty acid-related exposures showed no significant causal relationship with cognitive function. In the multivariable MR analysis of LA/FA-related exposures from 2 datasets, after adjusting for education, ischemic stroke, and 25-hydroxyvitamin D levels, the causal relationship between LA/FA and cognitive function remained statistically significant. Furthermore, meta-analysis demonstrated high homogeneity in the univariate and multivariable MR results for the two traits.

Our study results indicate a potential causal relationship between the ratio of LA to total fatty acids and cognitive function. Several mechanisms may underlie this finding, one proposed mechanism is the involvement of LA in maintaining central nervous system cell homeostasis. 24 LA is a precursor for the synthesis of bioactive metabolites, such as arachidonic acid and DHA, which are vital components of neuronal membranes. 25 These metabolites play crucial roles in neuronal signaling, synaptic plasticity, and neuroprotection.26,27, 26,27 By providing the necessary building blocks for these processes, LA may contribute to optimal cognitive function.

Additionally, LA has been found to possess anti-inflammatory properties. Chronic inflammation has been implicated in cognitive decline and neurodegenerative diseases. LA can modulate inflammatory pathways by influencing the production of pro-inflammatory and anti-inflammatory mediators. 28 By reducing neuroinflammation, LA may protect against cognitive impairment and promote cognitive health. 29

The brain is highly susceptible to oxidative damage due to its high metabolic activity and abundance of polyunsaturated fatty acids. LA, as an antioxidant, can scavenge free radicals and mitigate oxidative stress. 30 Oxidative stress is known to contribute to cognitive decline, and by counteracting this process, LA may exert neuroprotective effects and preserve cognitive function. 31 Furthermore, the results of Alarcon-Gil have demonstrated that LA acts as a potent neuroprotective and anti-inflammatory agent in these Parkinson’s disease models. 27 We also observed that LA stimulates the biogenesis of lipid droplets and improves the autophagy/lipophagy flux, which resulted in an antioxidant effect in the in vitro Parkinson’s disease model.

LA’s impact on cardiovascular health may also be a relevant mechanism underlying its effects on cognitive function. Given the well-established link between cardiovascular health and cognitive performance, LA’s potential to improve vascular function and reduce cardiovascular risk factors could indirectly benefit cognitive function by preserving cerebral blood flow and vascular integrity.32,33, 32,33

Furthermore, LA’s interaction with other nutrients, such as vitamin D, and its influence on the gut microbiota composition are emerging areas of interest that may provide additional insights into its effects on cognitive function. 34 Brain derived neurotrophic factor is a protein that promotes the survival, growth, and differentiation of neurons. It plays a crucial role in synaptic plasticity and neurogenesis, which are essential for learning and memory. LA has been shown to increase brain derived neurotrophic factor, suggesting a potential mechanism through which it may enhance cognitive function. 35

Limitation

The study faces several limitations. The observational nature of NHANES data restricts the ability to establish causal relationships, making the findings susceptible to residual confounding and unmeasured variables that may influence the observed associations between LA exposure and cognitive function. Moreover, MR relies on the assumption that genetic variants are associated with the exposure of interest and are independent of confounding factors. Yet, genetic instruments may be subject to pleiotropy or linkage disequilibrium, potentially biasing the causal estimates. Additionally, the reliance on self-reported dietary assessments in NHANES introduces recall bias and measurement error, impacting the accuracy of LA intake estimation and potentially affecting the precision of the exposure variable. Furthermore, the exclusive reliance on the DSST cognitive test score in NHANES analysis limits the breadth of cognitive assessment, potentially overlooking other dimensions of cognitive function. Lastly, while NHANES provides a nationally representative sample, the generalizability of the findings to specific subpopulations or diverse demographic groups may be limited due to factors such as cultural dietary differences and genetic heterogeneity.

Conclusion

We found a suggestive causal relationship between the serum ratio of LA to total fatty acids and cognitive function. Further research aimed at elucidating these mechanisms is essential for a comprehensive understanding of the role of LA in cognitive health and for the development of targeted interventions to promote cognitive well-being.

AUTHOR CONTRIBUTIONS

Huimin Zhao (Funding acquisition; Investigation; Writing – original draft); Changlin Yang (Supervision; Validation; Writing – review & editing); Fangkai Xing (Data curation; Investigation; Methodology; Software; Visualization).

Footnotes

ACKNOWLEDGMENTS

We sincerely thank the NHANES survey and openGWAS project in this analysis for providing public datasets and managing summary statistics.

FUNDING

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

Data and respective datasets are displayed at the NHANES (https://www.cdc.gov/nchs/nhanes/Index.htm) and openGWAS ().

The supplementary material is available in the electronic version of this article: .

References

Sziraki

, Lu

, Lee

, et al. A global view of aging and Alzheimer’s pathogenesis-associated cell population dynamics and molecular signatures in human and mouse brains. Nat Genet 2023; 55: 2104–2116.

, Lin

, Wu

, et al. Exploring successful cognitive aging: insights regarding brain structure, function, and demographics. Brain Sci 2023; 13: 1651.

Brown

, Gruijters

, Kotz

. Prediction in the aging brain: merging cognitive, neurological, and evolutionary perspectives. J Gerontol B Psychol Sci Soc Sci 2022; 77: 1580–1591.

de Carvalho

, Caramujo

. The various roles of fatty acids. Molecules 2018; 23: 2583.

Sánchez-Alegría

, Arias

. Functional consequences of brain exposure to saturated fatty acids: From energy metabolism and insulin resistance to neuronal damage. Endocrinol Diabetes Metab 2023; 6: e386.

Djuricic

, Calder

. Beneficial outcomes of omega-6 and omega-3 polyunsaturated fatty acids on human health: an update for 2021. Nutrients 2021; 13: 2421.

Tseng

, Zeng

, Suen

, et al. Efficacy and acceptability of anti-inflammatory eicosapentaenoic acid for cognitive function in Alzheimer’s dementia: A network meta-analysis of randomized, placebo-controlled trials with omega-3 fatty acids and FDA-approved pharmacotherapy. Brain Behav Immun 2023; 111: 352–364.

Wei

, Li

, Dong

, et al. The relationship of omega-3 fatty acids with dementia and cognitive decline: evidence from prospective cohort studies of supplementation, dietary intake, and blood markers. Am J Clin Nutr 2023; 117: 1096–1109.

Xia

, Yang

, Huang

, et al. Omega-3 polyunsaturated fatty acid eicosapentaenoic acid or docosahexaenoic acid improved ageing-associated cognitive decline by regulating glial polarization. Mar Drugs 2023; 21: 398.

10.

Kousparou

, Fyrilla

, Stephanou

, et al. DHA/EPA (Omega-3) and LA/GLA (Omega-6) as bioactive molecules in neurodegenerative diseases. Int J Mol Sci 2023; 24: 10717.

11.

Petermann

, Reyna-Jeldes

, Ortega

, et al. Roles of the unsaturated fatty acid docosahexaenoic acid in the central nervous system: molecular and cellular insights. Int J Mol Sci 2022; 23: 5390.

12.

Bowden

and Holmes

. Meta-analysis and Mendelian randomization: A review. Res Synth Methods 2019; 10: 486–496.

13.

Fellows

and Schmitter-Edgecombe

. Symbol digit modalities test: regression-based normative data and clinical utility. Arch Clin Neuropsychol 2019; 35: 105–115.

14.

Lagerstedt

, Hinrichs

, Batt

, et al. Quantitative determination of plasma c8-c26 total fatty acids for the biochemical diagnosis of nutritional and metabolic disorders. Mol Genet Metab 2001; 73: 38–45.

15.

Burgess

, Davies

and Thompson

. Bias due to participant overlap in two-sample Mendelian randomization. Genet Epidemiol 2016; 40: 597–608.

16.

Burgess

and Thompson

. Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol 2011; 40: 755–764.

17.

Verbanck

, Chen

, Neale

, et al. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet 2018; 50: 693–698.

18.

Palmer

, Lawlor

, Harbord

, et al. Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat Methods Med Res 2012; 21: 223–242.

19.

Rasooly

and Peloso

. Two-sample multivariable mendelian randomization analysis using R. Curr Protoc 2021; 1: e335.

20.

Walsh

, Yang

, Oi

, et al. Age profiles of cognitive decline and Dementia in Late Life in the Aging, Demographics, and Memory Study. J Gerontol B Psychol Sci Soc Sci 2022; 77: 1880–1891.

21.

DiNicolantonio

and O’Keefe

. The importance of marine omega-3s for brain development and the prevention and treatment of behavior, mood, and other brain disorders. Nutrients 2020; 12: 2333.

22.

Román

, Jackson

, Gadhia

, et al. Mediterranean diet: The role of long-chain ω-3 fatty acids in fish; polyphenols in fruits, vegetables, cereals, coffee, tea, cacao and wine; probiotics and vitamins in prevention of stroke, age-related cognitive decline, and Alzheimer disease. Rev Neurol (Paris) 2019; 175: 724–741.

23.

Skrivankova

, Richmond

, Woolf

, et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA 2021; 326: 1614–1621.

24.

Hamilton

and Klett

. Linoleic acid and the regulation of glucose homeostasis: A review of the evidence. Prostaglandins Leukot Essent Fatty Acids 2021; 175: 102366.

25.

Cocchi

, Minuto

, Tonello

, et al. Linoleic acid: Is this the key that unlocks the quantum brain? Insights linking broken symmetries in molecular biology, mood disorders and personalistic emergentism. BMC Neurosci 2017; 18: 38.

26.

da Costa Souza

, Grodzki

, Morgan

, et al. Oxidized linoleic acid metabolites regulate neuronal morphogenesis in vitro. Neurochem Int 2023; 164: 105506.

27.

Alarcon-Gil

, Sierra-Magro

, Morales-Garcia

, et al. Neuroprotective and anti-inflammatory effects of linoleic acid in models of Parkinson’s disease: the implication of lipid droplets and lipophagy. Cells 2022; 11: 2297.

28.

Kim

, Park

, Kim

, et al. Anti-inflammatory effects of fermented lotus root and linoleic acid in lipopolysaccharide-induced RAW 264.7. Cells. Life (Basel) 2020; 10: 293.

29.

Zhou

, Dong

, Hu

, et al. Staged suppression of microglial autophagy facilitates regeneration in CNS demyelination by enhancing the production of linoleic acid. Proc Natl Acad Sci U S A e 2023; 120: 2209990120.

30.

Morvaridzadeh

, Estêvão

, Morvaridi

, et al. The effect of conjugated linoleic acid intake on oxidative stress parameters and antioxidant enzymes: a systematic review and meta-analysis of randomized clinical trials. Prostaglandins Other Lipid Mediat 2022; 163: 106666.

31.

Barzegarzadeh

, Hatami

, Dehghan

, et al. Conjugated linoleic acid-curcumin attenuates cognitive deficits and oxidative stress parameters in the ethidium bromide-induced model of demyelination. Neurotox Res 2021; 39: 815–825.

32.

Santurino

, López-Plaza

, Fontecha

, et al. Consumption of goat cheese naturally rich in omega-3 and conjugated linoleic acid improves the cardiovascular and inflammatory biomarkers ofoverweight and obese subjects: a randomized controlled trial. Nutrients 2020; 12: 1315.

33.

, Hennebelle

, Sahlas

, et al. Soluble epoxide hydrolase-derived linoleic acid oxylipins in serum are associated with periventricular white matter hyperintensities and vascular cognitive impairment. Transl Stroke Res 2019; 10: 522–533.

34.

Fleck

, Hucke

, Teipel

, et al. Dietary conjugated linoleic acid links reduced intestinal inflammation to amelioration of CNS autoimmunity. Brain 2021; 144: 1152–1166.

35.

Chung

, Cui

, Sumiyoshi

, et al. Associations of fatty acids with cognition, psychopathology, and brain-derived neurotrophic factor levels in patients with first-episode schizophrenia and related disorders treated with paliperidone extended release. J Psychopharmacol 2017; 31: 1556–1563.