Association of dietary lecithin intake with mild cognitive impairment and cognitive performance in elderly population

Abstract

Background:

The cognitive-enhancing effects of dietary lecithin are gaining increasing attention, but evidence supporting these claims remains limited. This study aims to explore the relationship between dietary lecithin intake and cognitive function, as well as the prevalence of mild cognitive impairment (MCI), in adults aged over 60 years.

Methods:

This cross-sectional study included participants aged over 60 years recruited from Beijing, Jincheng City (Shanxi Province), and Linyi City (Shandong Province) from 2020 to 2023. Cognitive function and dietary intake were assessed using neuropsychological tests and a food frequency questionnaire, respectively. Body composition was measured by bioelectrical impedance analysis, and blood lipids by an automated biochemistry analyzer. Spearman correlation analysis, multiple linear regression, logistic regression, and receiver operating characteristic (ROC) curve analysis were performed.

Results:

Among the 1917 participants enrolled, 1023 (1023/1917, 53.4%) had MCI, of whom 542 (542/1023, 53.0%) were women. Multiple linear regression showed that after dichotomizing lecithin and serum cholesterol indicators according to the median, high levels of dietary lecithin, total cholesterol (TC), and high-density lipoprotein cholesterol (HDL-C) were associated with higher Montreal Cog­nitive Assessment (MoCA, regression coefficients were 0.49, 0.44, and 0.38, respectively); logistic regression showed that, compared with low intake level, the odds ratio (95% confidence interval [CI]) of MCI in high dietary lecithin intake group was 0.80 (95% CI: 0.64, 0.99). ROC analysis of a model incorporating dietary lecithin, energy intake, demographic, and clinical characteristics yielded an area under the curve (AUC) of 0.735 (95% CI: 0.713–0.757). Targeted lipidomic analysis revealed that the MCI group exhibited a lower total phosphatidylcholine (PC) level, a reduced proportion of PC containing polyunsaturated fatty acid (PUFA), and a higher proportion of PC containing saturated fatty acid (SFA).

Conclusion:

Higher dietary lecithin intake was associated with better cognitive performance and a lower prevalence of MCI among the elderly population aged 60 and above.

Introduction

Mild cognitive impairment (MCI) represents a transitional stage between healthy aging and dementia.^[1] It is characterized by a decline in cognitive functions or abilities that does not result in substantial functional impairment, which distinguishes it from dementia.^[2] The projected annual conversion rate from MCI to Alzheimer’s disease (AD) ranges from 10% to 20%.^[3] Cognitive impairment seriously affects the quality of life of the elderly, and the growing socioeconomic burden it imposes has become an issue for modern families, countries, and society worldwide.

As one of the modifiable environmental factors, the role of dietary nutrition in neuroprotection and improving age-related surrounding tissues has been gradually confirmed. In particular, the beneficial effect of dietary lecithin on cognitive function has attracted increasing attention. Lecithin, also known as phosphatidylcholine (PC), is regarded as a “vascular sweeper” and is a major component of cell membranes. It is also the main source of choline and a precursor for the synthesis of the neurotransmitter acetylcholine. Dietary lecithin not only modulates lipid metabolism and enhances vascular function in older adults, but it also releases choline following digestion and absorption, thereby promoting the synthesis of acetylcholine. In this way, after crossing the blood–brain barrier, it may play a crucial regulatory role in cerebrovascular health and cognitive function.^[4–6]

Prior research indicates that lecithin is associated with cognitive function and may contribute to improvements in cognitive performance.^[7] A prospective cohort study involving 2497 middle-aged men in Finland revealed that a higher intake of dietary lecithin was associated with a reduced risk of developing dementia.^[8] Numerous studies have shown that plasma PC levels were markedly lower in patients with AD than in healthy individuals.^[9,10] In addition, our prior research has highlighted differences in phospholipid intake and disruptions in lipid metabolism among individuals with MCI compared to those without such conditions.^[11] In a recent Japanese intervention study, two dosages of soy lecithin supplementation, high (1.2 g/day) and low (0.6 g/day), were administered to middle-aged women (aged 40–60 years) over an 8-week period. The findings revealed that a dosage of 1.2 g/day significantly alleviated menopausal symptoms and decreased cardiovascular diastolic blood pressure as well as the cardio-ankle vascular index.^[12] Furthermore, in a population-based, double-blind, randomized controlled trial, supplementation with soy phosphatidylserine (300 mg/day) over a 6-month period resulted in enhanced delayed verbal recall among Japanese subjects with low neuropsychological test scores.^[13] These results suggested the potential of dietary lecithin to improve cognitive function. However, the role of lecithin in cognitive function development still requires stronger and more comprehensive evidence.

Therefore, this cross-sectional study aims to investigate the association between dietary lecithin intake and the prevalence of MCI in a relatively large community-based population. Furthermore, it seeks to explore whether lecithin intake could be considered a potential indicator for the early identification of individuals at risk of cognitive decline. By leveraging targeted lipidomics, this study also explores the underlying biological mechanisms. We anticipate that our findings will not only reinforce the epidemiological evidence linking lecithin with cognitive health but also provide a foundation for future longitudinal and interventional studies aimed at developing public health nutrition strategies for healthy cognitive aging.

Methods

Participants

We recruited adults aged 60 and above who had undergone a health examination as part of the registered cohort study (ChiCTR2100054969, https://www.chictr.org.cn), which was conducted from 2020 to 2023 across multiple community health centers and hospitals in Beijing, Jincheng City (Shanxi Province), and Linyi City (Shandong Province), China. All participants were enrolled from community population censuses or village registries and had resided in the study area for at least one year prior to the survey date. The methodology and criteria used in this study were consistent with those utilized in prior research.^[14] The study was approved by the Ethics Committees of Capital Medical University (No. Z2019SY052), and written informed consent was obtained from all participants. Finally, 1917 participants were included in this study. All participants underwent face-to-face interviews, including the collection of demographic information, medical history of chronic diseases, cognitive assessment, dietary surveys, and body composition measurements.

Cognitive assessment

Cognitive function was assessed using the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE).^[15–17] Participants were diagnosed with MCI through screening followed by clinical diagnosis.^[16,17] First, a preliminary screening for potential MCI patients was conducted according to the MoCA scores. The total score for MoCA is 30 points and the cutoff scores for MCI were 13/14 for illiterate individuals, 19/20 for individuals with 1 to 6 years of education, and 24/25 for individuals with 7 or more years of education. Second, participants suspected of having MCI based on their MoCA scores underwent further examination by neurologists to establish the final diagnosis. The clinical diagnosis of MCI by neurologists was conducted in accordance with the core clinical criteria for MCI as recommended by the National Institute on Aging-Alzheimer’s Association (NIA-AA) work groups.^[18] This comprehensive assessment included the following key elements: (1) Concern regarding a change in cognition: A report of subjective cognitive decline from the participant or a knowledgeable informant. (2) Impairment in one or more cognitive domains: Objective evidence of lower performance in one or more cognitive domains, based on the MoCA/MMSE results and clinical judgment. (3) Preservation of independence in functional abilities: Participants generally maintained independence in basic activities of daily living, with any noted difficulties not constituting a significant impairment in social or occupational functioning. (4) Not demented: The cognitive changes were judged to be sufficiently mild and not meeting the diagnostic threshold for dementia.

Dietary assessment and body composition measurement

Dietary information was collected using the food frequency questionnaire (FFQ), according to the China National Nutrition and Health Survey.^[19] Energy and nutrient intake were determined using the China Food Composition Database (Version 6).^[20]

Body composition was estimated by bioelectrical impedance analysis (BIA) (InBody770 body composition analyzer, InBody Co., Ltd., the Republic of Korea), as described in previous studies.^[14]

Blood sample collection

Blood samples were collected from participants who had abstained from food intake since 8 PM the previous evening. Serum levels of total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) were quantified using an enzymatic assay. This analysis was performed using an automated biochemistry analyzer (Olympus AU480, Tokyo, Japan). Non-high-density lipoprotein cholesterol (non-HDL-C) was calculated by subtracting HDL-C from TC. Furthermore, the ratio of LDL-C to HDL-C was calculated based on the measured blood lipid levels.

Targeted lipidomic measurement

Fifteen blood samples were randomly selected from each of the MCI and non-MCI groups for targeted lipidomic measurement. The selected participants did not differ significantly from the non-selected participants in terms of age or gender (both P >0.05). The experimental method used in this study was the same as that used in previous studies.^[21] We used a modified lipid extraction method, and the details of this method are provided in the Supplementary Material, http://links.lww.com/CM9/C988, including details about the reagents, processes, and acquisition software.

Statistical analysis

We validated the collected questionnaires and raw data by performing a double data entry check utilizing Epidata3.1 (The EpiData Association, Odense, Denmark). Statistical analyses and graphical presentations were performed using R software version 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria), SPSS 25.0 (IBM Corporation, Armonk, NY, USA), and GraphPad Prism 9.5.0 (GraphPad Software, San Diego, CA, USA). All quantitative data exhibiting abnormal distribution were reported as median with interquartile range (IQR), specifically presented as M (P25, P75). Intergroup comparisons were conducted utilizing the Mann–Whitney U test. Qualitative data were represented as counts (percentages) and analyzed via the chi-squared test. To descriptively compare participant characteristics across different levels of dietary lecithin intake, lecithin intake was divided into quartiles based on its natural distribution, ensuring equal sample sizes across groups. The cut-off values were determined by the 25th, 50th, and 75th percentiles of the observed data (68.8, 88.0, and 108.6 mg/day, respectively), ensuring equal sample sizes across groups for descriptive purposes. The relationships between lecithin and other indicators were preliminarily assessed using Spearman correlation analysis. Following the transformation of lecithin into categorical variables using the median, multivariable linear regression models and binary logistic regression models were used to examine the associations between lecithin and the MoCA score, as well as between lecithin and the risk of MCI, respectively. All covariates included in the models were checked for multicollinearity, with variance inflation factors (VIF) all below 10, indicating no significant multicollinearity existed. Least absolute shrinkage and selection operator (LASSO) regression was used to select key factors for the final predictive model. A nomogram was developed to visually represent the model. The discriminatory power of the model was further assessed by utilizing receiver operating characteristic (ROC) curve analysis. A calibration curve was used to determine the consistency between predicted probabilities and observed outcomes. Predetermined stratified analyses were conducted to examine the relationships between lecithin levels and MoCA score, as well as between lecithin and the risk of MCI, categorized by age (<70 years old or ≥70 years old), gender (female or male), education level (with or without a junior high school degree), ethnicity (whether Han Chinese), and chronic diseases conditions (with or without hypertension, diabetes, hyperlipidemia, atherosclerosis, stroke, coronary heart disease, and arthritis). All statistical analyses were conducted using two-sided tests, with a statistical significance threshold established at P <0.05.

Results

Demographic and clinical characteristics of the participants

A total of 1917 individuals participated in this investigation, including 1023 patients with MCI (1023/1917, 53.4%), alongside 542 women (542/1023, 53.0%); the average dietary lecithin intake was 86.7 (67.9, 106.0) mg/day. The clinical characteristics of the 1023 MCI patients and 894 healthy control subjects are detailed in Table 1. Notably, the dietary lecithin intake among MCI patients was substantially lower than that in the control group. Moreover, the proportion of individuals with at least a junior high school education in the MCI group exceeded that of the control group. Results from body composition assessments and serum cholesterol analyses indicated that body fat mass (BFM), percent body fat (PBF), visceral fat area (VFA), and fat mass index (FMI) were significantly lower in the MCI group than in the control groups, and TC and HDL-C were markedly lower in the MCI group. Additionally, the incidence of stroke among the MCI group was significantly higher than in the control group.

Table 1.

Baseline characteristics of the total population in the cross‑sectional study.

Variables	Total (N = 1917)	Control (N = 894)	MCI (N = 1023)	Z/χ2	P value
Demographic characteristics
Age (years)	70 (67, 73)	70 (67, 73)	69 (67, 74)	–0.81	0.416
Female	1108 (57.8)	566 (63.3)	542 (53.0)	20.87	<0.001
Han Chinese	1095 (57.1)	550 (61.5)	545 (53.3)	13.25
Education				288.81	<0.001
Illiterate	387 (20.2)	285 (31.9)	102 (10.0)
Primary school	638 (33.3)	358 (40.0)	280 (27.4)
Junior high school	720 (37.6)	173(19.4)	547(53.5)
High school and above	172 (9.0)	78 (8.7)	94 (9.2)
MoCA	21 (18, 24)	23 (20, 25)	19 (16, 22)	–18.15	<0.001
Dietary intakes
Energy (kcal/day)	1756 (1450, 2145)	1770 (1470, 2190)	1750 (1440, 2110)	–1.81	0.070
Lecithin (mg/day)	88.0 (68.8, 108.6)	89.7 (70.7, 110.0)	86.7 (67.9, 106.0)	–2.04	0.042
Body composition
BMI (kg/m2)	26.1 (24.0, 28.5)	26.3 (24.1,28.6)	26.0 (23.9, 28.4)	–1.82	0.068
BMR (kcal/day)	1266 (1155,1402)	1260 (1160,1380)	1270 (1150, 1410)	–1.06	0.288
BFM (kg)	23.2 (18.6, 28.3)	23.8 (19.2, 28.8)	22.7 (18.0, 28.0)	–3.29	0.001
PBF (%)	35.8 (28.5, 41.4)	36.6 (29.5, 41.9)	34.8 (28.0, 40.5)	–3.82	<0.001
WHR	0.90 (0.85, 0.94)	0.90 (0.86, 0.95)	0.90 (0.85, 0.94)	–1.85	0.064
VFA (cm2)	121.0 (87.0, 155.1)	126.0 (92.5, 158.0)	117.0 (83.4, 150.0)	–3.94	<0.001
SLM (kg)	38.6 (33.6, 45.8)	38.6 (33.6, 45.3)	38.6 (33.6, 45.8)	–0.48	0.629
FFMI (kg/m2)	16.6 (15.7, 18.2)	16.5 (15.7, 18.1)	16.8 (15.7, 18.3)	–1.95	0.052
FMI (kg/m2)	9.2 (7.1, 11.5)	9.5 (7.4, 11.7)	9.1 (6.9, 11.2)	–3.24	0.001
SMI (kg/m2)	7.3 (6.5, 8.0)	7.2 (6.5, 7.9)	7.3 (6.6, 8.1)	–2.28	0.023
Serum cholesterol
TC (mmol/L)	4.71 (4.01, 5.39)	4.76 (4.04, 5.41)	4.66 (3.96, 5.36)	–2.16	0.031
TG (mmol/L)	1.33 (0.96, 1.89)	1.35 (0.97, 1.94)	1.32 (0.95, 1.86)	–0.93	0.351
HDL-C (mmol/L)	1.26 (1.08, 1.48)	1.28 (1.10, 1.49)	1.25 (1.05, 1.46)	–2.21	0.027
LDL-C (mmol/L)	2.92 (2.32, 3.52)	2.93 (2.34, 3.51)	2.90 (2.30, 3.52)	–0.71	0.475
Non HDL-C (mmol/L)	3.40 (2.73, 4.06)	3.44 (2.78, 4.08)	3.35 (2.68, 4.04)	–1.77	0.077
LDL-C/HDL-C	2.29 (1.76, 2.89)	2.27 (1.79, 2.89)	2.31 (1.75, 2.89)	–0.24	0.812
Chronic diseases
Hypertension	1361 (71.0)	636 (71.1)	725 (70.9)	0.02	0.896
Diabetes	627 (32.7)	287 (32.1)	340 (33.2)	0.28	0.598
Hyperlipidemia	342 (17.8)	160 (17.9)	182 (17.8)	0.09	0.952
Atherosclerosis	64 (3.3)	31 (3.5)	33 (3.2)	0.01	0.769
Stroke	187 (9.8)	74 (8.3)	113 (11.0)	4.15	0.042
Coronary heart disease	812 (42.4)	377 (42.2)	435 (42.5)	0.02	0.876
Arthritis	295 (15.4)	141 (15.8)	154 (15.1)	0.19	0.664

Data are shown as n (%) or median (Q1, Q3). BMI: Body mass index; BMR: Basal metabolic rate; BFM: Body fat mass; FFMI: Fat-free mass index; FMI: Fat mass index; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low-density lipoprotein cholesterol; PBF: Percent body fat; SLM: Soft lean mass; SMI: Skeletal muscle index; TC: Total cholesterol; TG: Triglyceride; VFA: Visceral fat area; WHR: Waist–hip ratio. Intergroup comparisons were conducted utilizing the chi-squared test and the Mann–Whitney U test.

Comparison of the participant characteristics based on different lecithin subgroups

Participants were divided into four groups according to quartiles of dietary lecithin intake: Quartile 1 (lecithin <68.8 mg/day); Quartile 2 (68.8 ≥lecithin <88.0 mg/day); Quartile 3 (88.0 ≥lecithin <108.6 mg/day); Quartile 4 (lecithin ≥108.6 mg/day). Supplementary Table 1, http://links.lww.com/CM9/C988 presents the characteristics of these subgroups. The results showed that MoCA score, energy, basal metabolic rate (BMR), soft lean mass (SLM), fat-free mass index (FFMI), and skeletal muscle index (SMI) levels gradually increased with the increase of lecithin from Quartile 1 to Quartile 4. The levels of PBF, waist–hip ratio (WHR), VFA, and FMI gradually decreased with the increase of lecithin intake from Quartile 1 to Quartile 4. HDL-C level increased from Quartile 1 to Quartile 3 but slightly decreased in Quartile 4, whereas the LDL-C/HDL-C ratio decreased from Quartile 1 to Quartile 3 but increased in Quartile 4. The proportion of MCI was lower in Quartile 3 and Quartile 4, compared with Quartile 1 and Quartile 2.

Spearman correlation analysis between dietary lecithin and other variables

The Spearman correlation analysis was used to investigate the relationship of lecithin with various other parameters. As shown in Supplementary Figure 1, http://links.lww.com/CM9/C988, the findings indicated a positive association between lecithin and the MoCA score, BMR, SLM, FFMI, and SMI (correlation coefficients were 0.15, 0.16, 0.10, 0.15, and 0.24, respectively, all P <0.05), while negative correlations were observed with BFM, PBF, VFA, and FMI (correlation coefficients were –0.05, –0.13, –0.11, and –0.10, respectively, all P <0.05).

Association between dietary lecithin and MoCA score

Lecithin and serum cholesterol levels were categorized as high or low based on the median values of the overall population. As demonstrated in Table 2, after adjusting for variables including age, sex, education, energy, body composition metrics, serum cholesterol indicators, and chronic diseases, a positive correlation was observed between high lecithin levels and the MoCA score (regression coefficient = 0.49, 95% confidence interval [CI]: 0.09–0.89). Even when serum cholesterol was combined, this positive association between lecithin and the MoCA score remained evident (regression coefficient = 0.53, 95% CI: 0.13–0.93). Furthermore, it was noteworthy that high TC level, high HDL-C level, and a low LDL-C/HDL-C ratio were positively associated with the MoCA score (regression coefficients were 0.44, 0.38, and 0.39, respectively). However, the correlation between low LDL-C levels and the MoCA score did not reach statistical significance.

Table 2.

Association of dietary lecithin and MoCA score among the elderly study population.

Variables	Model 1		Model 2		Model 3
	B (95% CI)	P value	B (95% CI)	P value	B (95% CI)	P value
Lecithin high*	0.49 (0.08, 0.89)	0.018	0.47 (0.07, 0.88)	0.022	0.49 (0.09, 0.89)	0.017
Lecithin high	0.52 (0.12, 0.92)	0.010	0.50 (0.10, 0.91)	0.014	0.53 (0.13, 0.93)	0.010
TC high	0.39 (0.02, 0.76)	0.036	0.42 (0.05, 0.79)	0.025	0.44 (0.07, 0.80)	0.020
Lecithin high†	0.49 (0.09, 0.89)	0.016	0.48 (0.07, 0.88)	0.020	0.50 (0.09, 0.90)	0.016
HDL-C high†	0.33 (<0.01, 0.70)	0.085	0.37 (–0.01, 0.74)	0.057	0.38 (<0.01, 0.76)	0.049
Lecithin high‡	0.49 (0.09, 0.90)	0.016	0.48 (0.08, 0.88)	0.020	0.50 (0.10, 0.90)	0.015
LDL-C low‡	0.05 (–0.31, 0.40)	0.800	0.04 (–0.32, 0.40)	0.831	0.03 (–0.33, 0.39)	0.863
Lecithin high	0.49 (0.09, 0.89)	0.017	0.47 (0.07, 0.88)	0.021	0.49 (0.09, 0.89)	0.017
LDL-C/HDL-C low	0.35 (–0.01, 0.71)	0.058	0.37 (<0.01, 0.73)	0.045	0.39 (0.02, 0.75)	0.037

BMI: Body mass index; BMR: Basal metabolic rate; CI: Confidence interval; FMI: Fat mass index; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low-density lipoprotein cholesterol; MoCA: Montreal Cognitive Assessment; SMI: Skeletal muscle index; TC: Total cholesterol; TG: Triglyceride. Multivariable models were adjusted for: model 1: sex+age+education+ethnicity+energy+TG; model 2: model 1+BMI+BMR+FMI+SMI; model 3: model 2+hypertension+diabetes+hyperlipidemia+atherosclerosis+stroke+coronary heart disease+arthritis. *Adding the variable HDL-C, non-HDL-C, and LDL-C/HDL-C. ^†Adding the variable non-HDL-C. ^‡Adding the variable HDL-C.

Association between dietary lecithin and MCI

The association between lecithin levels and the prevalence of MCI was further investigated, and the findings are presented in Table 3. After adjusting for variables such as age, sex, education, energy, body composition metrics, serum cholesterol levels, and chronic health conditions, high dietary lecithin intake exhibited a significant inverse association with MCI prevalence (odds ratio [OR] = 0.80, 95% CI: 0.64–0.99). When serum cholesterol was incorporated alongside lecithin in the analytical model, high dietary lecithin intake remained inversely associated with MCI prevalence (OR = 0.79, 95% CI: 0.63–0.98), indicating that increased dietary lecithin intake may mitigate the likelihood of MCI. Exploratively, the dose-response relationship between dietary lecithin and MCI prevalence was analyzed using restricted cubic splines combined with spline function with logistic regression, in which dietary lecithin intake was a continuous variable, and four knots (P5, P35, P65, and P95) were taken according to percentile distributions. As shown in Supplementary Figure 2, http://links.lww.com/CM9/C988, the abscissa represents the continuous change in dietary lecithin, the ordinate represents the corresponding predicted OR value, and the upper and lower ranges represent the 95% CI. Despite the apparent “S” shaped curve, the results showed no statistically significant difference in P-nonlinear. Visually, the OR for MCI remained below 1.0 across a dietary lecithin intake range of approximately 88 to 163 mg/day, suggesting a potential inverse association within this interval.

Table 3.

Logistic regression analysis of MCI by dietary lecithin median.

Variables	Model 1		Model 2		Model 3
	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
Lecithin high*	0.81 (0.65, 1.01)	0.061	0.80 (0.65, 1.00)	0.051	0.80 (0.64, 0.99)	0.045
Lecithin high	0.81 (0.65, 1.00)	0.054	0.80 (0.64, 0.99)	0.044	0.79 (0.63, 0.98)	0.037
TC high	0.90 (0.74, 1.09)	0.272	0.87 (0.71, 1.06)	0.170	0.87 (0.71, 1.06)	0.166
Lecithin high†	0.82 (0.66, 1.01)	0.065	0.81 (0.65, 1.01)	0.053	0.80 (0.64, 0.99)	0.046
HDL-C high†	0.89 (0.73, 1.09)	0.251	0.85 (0.69, 1.04)	0.115	0.83 (0.67, 1.02)	0.081
Lecithin high‡	0.81 (0.65, 1.01)	0.060	0.80 (0.64, 1.00)	0.051	0.80 (0.64, 0.99)	0.044
LDL-C low‡	1.02 (0.84, 1.24)	0.842	1.02 (0.85, 1.24)	0.795	1.02 (0.84, 1.24)	0.849
Lecithin high	0.82 (0.66, 1.01)	0.066	0.81 (0.65, 1.00)	0.055	0.80 (0.64, 0.99)	0.049
LDL-C/HDL-C low	0.90 (0.74, 1.07)	0.277	0.88 (0.72, 1.07)	0.200	0.86 (0.70, 1.05)	0.133

BMI: Body mass index; BMR: Basal metabolic rate; CI: Confidence interval; FMI: Fat mass index; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low-density lipoprotein cholesterol; MCI: Mild cognitive impairment; OR: Odds ratio; SMI: Skeletal muscle index; TC: Total cholesterol; TG: Triglyceride. Multivariable models were adjusted for: model 1: sex+age+education+ethnicity+energy+TG; model 2: model 1+ BMI+BMR+FMI+SMI; model 3: model 2+hypertension+diabetes+hyperlipidemia+atherosclerosis+stroke+coronary heart disease+arthritis. ^*Adding the variable HDL-C, non-HDL-C, and LDL-C/HDL-C. ^†Adding the variable non-HDL-C. ^‡Adding the variable HDL-C.

Predictive model of dietary intake, demographic, and clinical characteristics in MCI participants

To avoid overfitting and improve model simplicity, LASSO regression was used for the preliminary selection of risk factors [Supplementary Figure 3A, http://links.lww.com/CM9/C988]. Subsequently, the selected variables including dietary intake (dietary lecithin and energy intake), demographic, and clinical characteristics were integrated into the predictive model. A nomogram was developed to visually represent the predictive model [Supplementary Figure 3B, http://links.lww.com/CM9/C988]. Next, the discriminatory performance of the model was evaluated using the area under the curve (AUC) of the ROC curve. This model showed good discrimination performance for MCI in the study population, with an AUC of 0.735 (95% CI: 0.713–0.757) [Figure 1A]. The calibration curve indicated a high degree of consistency between the predicted probabilities and the actual frequencies of MCI [Figure 1B].

Figure 1.

Predictive model for MCI incorporating dietary intake, demographic, and clinical characteristics. (A) ROC curve of the predictive model; (B) Calibration curve of the predictive model. The model included the following variables: age, sex, education, energy intake, lecithin intake, basal metabolic rate , body fat mass, and stroke. AUC: Area under the curve; ROC: Receiver operating characteristic.

Subgroup analyses

Subgroup analyses were performed to explore the potential modifying effects of various factors on the association between dietary lecithin and cognitive outcomes. The results of the multiple linear regression (MoCA score as outcome) are presented in Figure 2. Crucially, tests for interaction were not statistically significant for most subgroups, including age, sex, and the presence of most chronic conditions (e.g., hypertension, diabetes, stroke, coronary heart disease; all P for interaction >0.05). This indicates that the association between lecithin and MoCA score did not differ significantly across these strata. Notably, a statistically significant interaction was observed for educational attainment (P for interaction <0.001), suggesting that the association between higher lecithin intake and a higher MoCA score was significantly stronger among participants with education level below junior high school.

Figure 2.

Subgroup analyses based on multiple linear regression model. The model was adjusted for sex, age, education, ethnicity, energy, TG, HDL-C, non-HDL-C, LDL-C/HDL-C, BMI, BMR, FMI, SMI, hypertension, diabetes, hyperlipidemia, atherosclerosis, stroke, coronary heart disease, and arthritis. Significance was accepted as P for interaction <0.0045 (α = 0.05/11) after Bonferroni correction. BMI: Body mass index; BMR: Basal metabolic rate; CI: Confidence interval; FMI: Fat mass index; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low-density lipoprotein cholesterol; SMI: Skeletal muscle index; TG: Triglyceride.

Targeted lipidomics analysis

Based on the significant role of dietary lecithin in cognitive function and the risk of MCI, erythrocyte membrane lipids were analyzed in 15 randomly selected subjects from each group. As illustrated in Figure 3, the MCI group exhibited a significant reduction in total erythrocyte membrane PC compared to the control group (P <0.001). Further stratification of PC subclass based on fatty acid composition revealed significant decreases in PC species containing saturated fatty acid (SFA), unsaturated fatty acid (UFA), and polyunsaturated fatty acid (PUFA) within the MCI cohort (P <0.001 for all comparisons). However, no intergroup differences were observed in PC containing monounsaturated fatty acid (MUFA). Proportional analysis of fatty acid-specific PC subclasses demonstrated that the MCI group had a significantly higher percentage of PC containing SFA relative to total PC compared to controls (P <0.05). Conversely, the proportions of both PC containing UFA and PC containing PUFAs relative to total PC were significantly lower in the MCI group (both P <0.05). The percentage of PC containing MUFAs showed no statistically significant difference between the two groups. Furthermore, orthogonal partial least squares-discriminant analysis (OPLS-DA) was performed to further identify discriminative lipid profiles associated with MCI. The score plot demonstrated a separation between the MCI and control group [Supplementary Figure 4, http://links.lww.com/CM9/C988]. Subsequently, six differential lipid profiles were identified based on P value (<0.05) and variable importance in projection (VIP) values (>1.00) [Supplementary Table 2, http://links.lww.com/CM9/C988]. Notably, PC(16:0/18:2), which contains linoleic acid, was identified as one of the differential lipids, consistent with our earlier observation of a reduced PC containing PUFA in MCI [Figure 3].

Figure 3.

Targeted lipidomic analysis of PC profiles in control and MCI groups (n = 15 per group). (A) Total PC levels; (B) PC containing SFA; (C) PC containing UFA; (D) PC containing MUFA; (E) PC containing PUFA; (F) PC containing SFA/total PC; (G) PC containing UFA/total PC; (H) PC containing MUFA/total PC; (I) PC containing PUFA/total PC. ^*P <0.05, ^†P <0.001. MCI: Mild cognitive impairment; MUFA: Monounsaturated fatty acid; PC: Phosphatidylcholine; PUFA: Polyunsaturated fatty acid; SFA: Saturated fatty acid; UFA: Unsaturated fatty acid.

Sensitivity analysis

To assess the robustness of our primary findings, sensitivity analyses were performed by categorizing dietary lecithin and serum cholesterol indicators into tertiles instead of using the median-based dichotomy [Supplementary Tables 3 and 4, http://links.lww.com/CM9/C988]. When analyzed by tertiles, higher dietary lecithin intake (Tertile 2 and Tertile 3) remained significantly associated with a higher MoCA score compared to Tertile 1, the lowest intake group (Tertile 2: B = 0.51, 95% CI: 0.07–0.95; Tertile 3: B = 0.58, 95% CI: 0.07–1.09). Similarly, the highest lecithin tertile was inversely associated with MCI prevalence (OR = 0.73, 95% CI: 0.55–0.96). For serum cholesterol indicators, the highest HDL-C tertile showed a significant positive association with MoCA score (B = 0.56, 95% CI: 0.10–1.03), suggesting a dose-response trend consistent with the association observed in the primary analysis. Notably, the highest tertile of HDL-C showed a significant inverse association with MCI risk (OR = 0.74, 95% CI: 0.57–0.96), although no significant association was observed in the prior analysis using median-based stratification. This suggests that a more refined categorization may reveal protective associations at the highest HDL-C levels.

Discussion

In this study, a high level of dietary lecithin was found to be positively correlated with MoCA score, and positive associations were also found between MoCA score with high levels of TC and HDL-C, as well as low levels of LDL-C/HDL-C ratio. Logistic regression analysis showed that after adjusting for multiple confounding factors, a high level of dietary lecithin was associated with a reduced prevalence of MCI. ROC curve analysis further supported this association by demonstrating fair discriminatory power for distinguishing current MCI cases from non-cases in a model incorporating lecithin intake, energy intake, demographic, and clinical characteristics. Targeted lipidomic analysis revealed that the MCI group exhibited lower total PC level, a reduced proportion of PC containing PUFA, and a higher proportion of PC containing SFA. Although these results cannot establish causality due to the nature of the cross-sectional design, they provide preliminary insights that may inform future strategies for the early detection and nutritional management of MCI.

First, we focused on differences in demographic and clinical characteristics between MCI and healthy controls. The results showed that the proportion of female patients with MCI was lower. The proportion of patients with a junior high school education or above was higher in the MCI group than that of the control group. Previous studies have reported mixed results regarding sex differences in MCI prevalence. Liu et al^[22] found that the incidence of MCI was more pronounced among elderly individuals within a survey of 1325 relatively healthy residents of Shanghai community aged 60 years and older, with a notably higher prevalence observed in women compared to men. A systematic review of the prevalence of MCI in China also indicated that women were more susceptible to developing MCI compared to men in this cohort, a finding that contrasts with several studies conducted in other countries.^[23] Research conducted on Finnish, German, and Italian populations revealed no significant differences attributable to sex.^[24,25] Petersen et al^[26] observed that the incidence of MCI was consistently higher in men than in women across all age demographics. The observed higher prevalence of MCI among women in preliminary Chinese studies could be attributed to historical disparities in educational access, where women faced greater barriers compared to their male counterparts in earlier decades.^[27] The lower proportion of women in the MCI group in this study may be due to the broader acceptance of basic education, which has narrowed the educational gap between men and women. Previous research has demonstrated that an elevated BMI is associated with an increased risk of developing MCI.^[28] However, our findings showed no significant difference in BMI between the two groups. Additionally, the MCI group exhibited lower body composition indicators reflecting fat content, whereas the indicators reflecting muscle mass were higher than those in the control group. Considering the gender distribution and daily living habits, we speculate that this may be related to the higher proportion of men in the MCI group, many of whom may continue farming or engage in other secondary occupations, as well as partake in basic physical activities after the age of 60 years. The MCI group also displayed lower levels of TC and HDL-C, indicating that reduced levels of TC and low HDL-C might be associated with cognitive decline. This finding is consistent with a cross-sectional study of 2000 community residents from a longevity region in China and a cohort study involving 180,000 participants in the United States.^[29,30] Previous studies have also confirmed a close association between stroke and cognitive impairment.^[31] Regarding chronic diseases in the present study, the prevalence of stroke in the MCI group was significantly higher than that in control group, suggesting a possible interactive relationship between stroke and MCI. This also implies that changes in cerebrovascular function may be related to the occurrence and progression of MCI, which warrants further exploration in our future study.

Based on univariate analysis, we found that dietary lecithin intake was significantly lower in the MCI group, which was also consistent with our previous results.^[11] These results suggest that insufficient lecithin intake may contribute to cognitive decline and an increased likelihood of MCI. Next, the participants were divided into four groups according to the quartiles of dietary lecithin intake. It is worth noting that MoCA score, SLM, FFMI, and SMI levels gradually increased with increasing lecithin intake from Quartile 1 to Quartile 4, whereas PBF, BFM, VFA, and FMI levels gradually decreased with increasing lecithin intake from Quartile 1 to Quartile 4. This result was also supported by the heat map, suggesting that increased lecithin intake may be associated with improved cognitive function and muscle mass. With increasing lecithin intake, HDL-C levels showed an upward trend, whereas the LDL-C to HDL-C ratio showed a downward trend. Although this trend did not reach statistical significance in the heat map analysis, it may suggest a potential pattern of lipid profile modulation associated with lecithin intake that warrants further investigation.

In our study, we investigated the association between dietary lecithin levels and MoCA scores. The results showed a significant positive association between high dietary lecithin intake and MoCA score, suggesting that higher lecithin intake in this population was associated with improved cognitive function. Further analyses incorporating serum cholesterol indicators showed that the positive correlation between high lecithin level and MoCA score remained stable. Furthermore, in the logistic regression analysis, high dietary lecithin intake was associated with a lower prevalence of MCI, and this association remained stable after incorporating serum cholesterol indicators. In a comparative study, Oresic et al^[32] observed that, at baseline, patients diagnosed with AD exhibited reduced serum levels of PC; another study based on untargeted metabolomics also found a decrease in PC in people with MCI.^[33] Our findings are consistent with these studies and provide further insights into the relationship between lecithin and MCI. Our subsequent ROC curve and calibration curve analyses indicated a satisfactory predictive performance of the model incorporating lecithin intake, energy intake, demographic factors, and clinical characteristics. Taken together, these results suggest a potential protective role of higher lecithin levels against MCI, highlighting their relevance in strategies aimed at mitigating cognitive decline.

Furthermore, when we categorized serum cholesterol indicators into tertiles, the highest tertile of HDL-C was found to be associated with a higher MoCA score and a lower MCI prevalence, which is consistent with existing studies.^[34,35] Particularly, in the Copenhagen General Population Study, data from a total of 75,708 individuals indicated a significant reduction in HDL-C levels among patients diagnosed with dementia.^[36] Additionally, our study revealed a positive correlation between high TC levels and MoCA scores, coupled with the observation of lower TC levels within the MCI group, despite the absence of a significant association between TC and MCI prevalence. Currently, the role of cholesterol in the pathogenesis of cognitive impairment remains unclear. Most studies argued that higher TC levels increase the risk of cognitive impairment by increasing the β-amyloid production and deposition in the brain.^[37] In the meantime, an increasing number of studies have also suggested that cholesterol may have a protective effect on cognition in the elderly population, possibly due to its role in maintaining the structural integrity and modulating the fluidity of neuronal cells.^[29,38] Furthermore, the majority of participants in this study had TC levels within the appropriate range or borderline elevated range, with only 7.9% of them exhibiting significantly elevated TC (≥6.2 mmol/L), which also supports our finding that higher TC levels within a reasonable range may exert a certain protective effect on cognitive function. We believe future research focusing on the effects of specific cholesterol subtypes will help more accurately elucidate the mechanisms by which cholesterol contributes to the onset and progression of cognitive impairment.

In the subgroup analysis, the most statistically robust finding was the significant interaction with educational attainment. The association between higher lecithin intake and better cognitive outcomes (higher MoCA scores) was significantly stronger in individuals with lower education levels. Based on this, educational level may act as a modifier in the relationship between dietary factors and cognitive function. One possible explanation is that individuals with lower educational attainment may have lower cognitive reserve, which refers to the brain’s resilience to neuropathological damage. As a result, their cognitive performance may be more vulnerable to metabolic disturbances but also more responsive to dietary factors such as lecithin. In contrast, those with higher education may benefit from greater lifelong cognitive stimulation and enriched environments, which could buffer the effects of suboptimal lipid profiles on cognition. These findings imply that targeted dietary strategies—such as enhancing lecithin intake through food or supplementation—may yield greater cognitive benefits among individuals with lower education levels. In public health planning, educational background could therefore serve as an important factor for identifying vulnerable subgroups who might gain the most from dietary interventions aimed at preserving cognitive function. This stratified approach may help improve the efficiency and effectiveness of nutritional strategies in preventing cognitive decline in aging populations.

The erythrocyte lipid profile can reflect long-term dietary lipid intake. According to a metabolomic study of residents in the Province of Huelva, Spain, AD patients exhibited elevated serum levels of PC containing SFA and reduced levels of PC containing PUFA.^[39] To further elucidate differences in PC metabolism between individuals with cognitive dysfunction and healthy populations, this study used targeted lipidomics to analyze erythrocyte samples. The results demonstrated a significant reduction in total PC levels in the MCI group. Further stratification based on fatty acid composition revealed marked declines in PC containing SFA, PC containing UFA, and PC containing PUFA in the MCI group. Although PC containing SFA levels were significantly lower in the MCI group, the proportion of PC containing SFA was notably higher in MCI patients compared to healthy controls. Conversely, the proportions of PC containing UFA and PC containing PUFA were significantly reduced in the MCI group. These findings align with existing research conclusions. Given that the fatty acid composition of membrane lipids influences their biophysical properties—including fluidity, permeability, and charge—deficiencies in key unsaturated fatty acids highly enriched in neurons, such as docosahexaenoic acid (DHA) and arachidonic acid (AA), may contribute to AD pathogenesis by inducing membrane damage.^[40] Studies have reported decreased levels of PUFA-containing phospholipids in the brains and plasma of AD patients,^[41–43] with a pronounced reduction in PC levels.^[44,45] The existing evidence mentioned above suggests that characteristic alterations in PC metabolism may serve as potential biomarkers for early MCI identification. Pending future validation through serological biomarkers, our findings suggest that simple assessment of dietary lecithin intake may offer a feasible and low-cost approach for preliminary screening of elderly individuals at risk of cognitive decline in community and public health settings. To translate these findings into practical and specific dietary advice, our results highlight the importance of ensuring sufficient dietary lecithin intake; the established dietary reference intakes (DRIs) for choline can serve as a reference. According to the DRIs for Chinese residents, the adequate intake (AI) for choline is 450 mg/day for males and 380 mg/day for females,^[46] which is readily achievable through a balanced diet including lecithin-rich foods including eggs, legumes, and lean meats.^[47] In addition, adherence to healthy dietary patterns that naturally include these foods, such as the Mediterranean diet, has been consistently linked to better cognitive health in the elderly.^[48,49] Therefore, incorporating these foods into regular dietary patterns represents a simple, evidence-based approach to potentially support cognitive health.

This study has several limitations. First, a causal relationship between lecithin and MCI cannot be established due to the cross-sectional nature of the study; prospective cohort studies are necessary for a more in-depth evaluation. Second, although the sample size was substantial, recruitment from a single region may limit the generalizability of our findings to other populations with different genetic backgrounds, dietary habits, or socioeconomic contexts. Third, although we adjusted for a range of confounders, residual confounding cannot be entirely ruled out, as factors such as household income and size—which may influence both diet and cognitive health—were not available for adjustment. Fourth, our targeted panel did not include separate quantification of individual PC species containing the long-chain PUFAs DHA and AA, which are of particular interest in neuroscience areas and warrant targeted analysis in future research. Moreover, although MCI was diagnosed through a two-step process incorporating clinical evaluation, the absence of a comprehensive neuropsychological battery remains a limitation, and the possibility of outcome misclassification cannot be entirely excluded. Finally, dietary lecithin intake was assessed using an FFQ, which is inherently subject to recall bias and measurement error. Nevertheless, the FFQ remains a well-validated and practical instrument for ranking participants’ dietary intake, particularly in large-scale epidemiological studies in which relative comparisons are more critical than precise quantification.

In conclusion, higher dietary lecithin intake was associated with better cognitive performance and a lower prevalence of MCI in this elderly population. Therefore, the assessment of dietary lecithin intake holds promise as a cost-effective component of initial cognitive risk screening in community settings. Our findings also reinforce the association between a favorable blood lipid profile, characterized by higher HDL-C and a lower LDL-C/HDL-C ratio, and better cognitive performance. Given that MCI is a multifactorial condition, a small reduction in multiple risk factors may substantially decrease the overall risk. As part of a broader strategy, ensuring adequate lecithin intake through the consumption of lecithin-rich foods or dietary supplements may be a feasible and cost-effective approach to help preserve cognitive health and reduce the overall risk of MCI.

Acknowledgments

We express our gratitude to all the individuals who participated in this research study. We extend our gratitude to Professor Chen Rui, Vice President of Capital Medical University, for his guidance on the overall research design and key technical aspects. We thank Mr. Endi Hodaj for his professional assistance in revising the language of this paper.

Funding

This research was supported by grants from the National Natural Science Foundation of China (No. 82273620), Beijing Municipal Natural Science Foundation (No. 7252179), and Chinese Nutrition Society (CNS) Nutrition Science Foundation-nutrilite Plant Functional Ingredients and Health Research Fund (No. CNS-NCL2025-217).

Conflicts of interest

None.

Supplementary Material

cm9-139-2938-s001.pdf ^{(510.2KB, pdf)}