Abstract
Background
The Dietary Index for Gut Microbiota (DI-GM) is a novel index reflecting diet quality relative to gut microbiota health.
Objective
The study aims to investigate the relationship between DI-GM and cognitive function in older adults.
Methods
Data were obtained from 2629 participants aged ≥60 years in the National Health and Nutrition Examination Surveys (NHANES) (2011–2014). Cognitive function was assessed using the Consortium to Establish a Registry for Alzheimer's Disease (CERAD), the Animal Fluency Test (AFT), the Digit Symbol Substitution Test (DSST), and a global z score. Multivariable linear regression, restricted cubic splines (RCS), and subgroup analysis were performed. Predictive utility of DI-GM was assessed via the receiver operating characteristic (ROC) analysis against a baseline model. Mediation analysis examined relationships among DI-GM, the Dietary Inflammatory Index (DII), and cognitive outcomes.
Results
Higher DI-GM was associated with higher AFT, DSST, and the global z scores (p < 0.001). After full adjustment, participants with DI-GM (≥ 6) showed higher AFT score (β = 1.11, 95% CI 0.46∼1.75), DSST score (β = 4.95, 95% CI 3.05∼6.86) and z score (β = 0.19, 95% CI 0.10∼0.28), compared to those with DI-GM (0–3). No significant direct association was observed with CERAD (β = 0.45, 95% CI −0.29∼1.18, p = 0.233). RCS indicated linear relationships between DI-GM and cognitive function scores. DI-GM had excellent predictive performances based on the ROC. No significant interactions were detected by subgroup analysis. Furthermore, DII partly mediated the relationship between DI-GM and cognitive function.
Conclusions
The DI-GM showed a linear positive correlation with cognitive function in older adults.
Keywords
Introduction
The incidence of cognitive dysfunction has increased over the past few decades as the global population ages. 1 Cognitive dysfunction, including mild cognitive impairment (MCI) and dementia, constitutes a critical public health challenge, particularly among geriatric demographics. 2 Modifiable lifestyle factors represent promising targets for influencing cognitive health in later life and may potentially contribute to attenuating the probability of progressive cognitive dysfunction and the onset of neurocognitive disorders. Diet, emerges as a key modifiable lifestyle factor, can mitigate the susceptibility to neurocognitive deterioration. 3 Healthy dietary patterns modulate the diversity and function of the intestinal microbiome. 4 The interaction linking diet with gut microbiome significantly influences health outcomes, with nutrient-microbe interactions determining microbiota stabilization or disruption, altering glucose homeostasis, lipid metabolism, inflammatory and cardiovascular pathways.5–7 Accumulating evidence elucidates the pivotal role of the gut microbiome in modulating microglial activation, neuronal plasticity, and the integrity of the cerebrovascular barrier. Furthermore, via the microbiota-gut-brain signaling pathway, these microbes drive neuroinflammatory cascades intimately linked to the pathogenesis of Alzheimer's disease (AD).8,9 Probiotics have exhibited the capacity to improve cognitive flexibility and mitigate psychological strain in healthy older adults, while also promoting positive alterations in gut microbiota composition. 10 These diet-microbiota interactions are increasingly recognized as significant factors in cognitive impairment, and targeted dietary interventions might potentially decrease morbidity risk.
Kase and colleagues recently introduced the Dietary Index for Gut Microbiota (DI-GM), a scoring system derived from an extensive review of 106 investigations, to assess how dietary patterns shape the functional diversity of the intestinal microbiome. 11 The DI-GM scores 14 foods or nutrients that either promote or inhibit microbial health, with higher scores indicating a healthier gut microbiota. As a validated dietary assessment tool, the DI-GM serves as a surrogate marker to assess dietary patterns with the potential to influence the intestinal environment, providing a foundation for future research on microbiome-targeted dietary strategies.
Prior investigations have linked diet and gut microbiota with cognitive impairment onset. 10 Recent findings have demonstrated that anti-inflammatory nutritional regimes, notably the Mediterranean diet (MD) and Dietary Approaches to Stop Hypertension (DASH), confer neuroprotection and modulate the trajectory of cognitive senescence.12–14 However, there remains a paucity of empirical data concerning the specific correlation between the DI-GM and cognitive outcomes. We suggest that DI-GM could be a simple and efficient tool for investigating the association between gut microbiome-related dietary patterns and cognitive performance in research populations. To test this hypothesis, the present study utilized the National Health and Nutrition Examination Survey (NHANES) cohort to examine the association between DI-GM and cognitive function in adults aged 60 and older, contributing to a better understanding of the role of diet-gut interactions in cognitive aging.
Methods
Study design and population
We analyzed data derived from the NHANES dataset spanning the years 2011 to 2014. Administered by the National Center for Health Statistics (NCHS), NHANES employs a stratified, multistage probability sampling design to generate a representative sample of the civilian, non-institutionalized United States population. Data acquisition encompassed structured interviews and clinical examinations covering demographic, socioeconomic, dietary, and health-related parameters. The protocol received institutional ethical approval, and written consent was obtained from all subjects.
The study population was restricted to adults aged 60 and above. Individuals were excluded from the study if they lacked complete records for cognitive performance, DI-GM components, or essential covariates. In total, 2629 eligible participants were included in the final statistical analysis (Supplemental Figure 1).
Dietary Index for Gut Microbiota
Adopting the methodology of Kase et al., 15 we derived the DI-GM index to assess diet-induced microbial modulation. These elements are stratified into two functional categories: protective (n = 10) and detrimental (n = 4). The protective cluster includes fermented dairy, chickpeas, soybeans, dietary fiber, cranberries, whole grains, avocados, broccoli, coffee, and green tea. Conversely, the detrimental cluster comprises red meat, processed meat, refined grains, and high-fat diets (≥40% energy from fat). Dietary data were extracted from NHANES 2011–2014 interviews. Participants received 1 point for each protective component consumed above the sex-specific median. For detrimental components, reverse scoring was applied: 0 points for intake above the median (≥40% for fat) and 1 point for lower intake. The cumulative DI-GM score ranged from 0 to 13 (comprising 0–9 for protective and 0–4 for detrimental factors), with higher values indicating a dietary pattern more favorable to gut microbiota health. The study population was stratified by score quartiles: 0–3, 4, 5, and ≥6. Detailed definitions are provided in Supplemental Table 1.
Cognitive function
Cognitive integrity was evaluated using a comprehensive battery comprising the Consortium to Establish a Registry for AD (CERAD) Word Learning Subtest, the Animal Fluency Test (AFT), and the Digit Symbol Substitution Test (DSST). The CERAD subtest, assessed immediate and delayed verbal memory through three learning trials and a subsequent recall task involving ten distinct words (score range: 0–10). 15 The category-based verbal fluency was assessed via the AFT. 16 Participants were allotted 60 s to enumerate a maximum number of distinct animals; the aggregate count of valid entries constituted the primary metric. The DSST was deployed to quantify psychomotor speed, sustained attention, and working memory. 17 Examinees were first provided with a reference key pairing nine digits with corresponding symbols and were then asked to transcribe the appropriate symbol for each digit in a series of adjacent cells within a 2-min interval. The DSST result was operationalized as the total count of correct symbol–digit matches. The z score was utilized to normalize the scores for CERAD, AFT, and DSST, computed as (individual score − sample mean) divided by the sample standard deviation for each instrument.
Currently, there is no established criterion for defining low cognitive performance on these tests. In accordance with methodologies employed in published literature, we utilized the 25th percentile (lowest quartile) as the cutoff value.18,19
Covariates
Based on prior epidemiological evidence, the statistical models were adjusted for a comprehensive set of confounding factors, comprising age, sex, race and ethnicity, educational attainment, family poverty income ratio, body mass index, total calories intake, smoking status, alcohol use, hypertension, diabetes, coronary heart disease and stroke.20–22 Race / ethnicity was stratified into five discrete cohorts: Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, and other races. Education level was coded into five ordered levels: Less Than 9th Grade, 9–11th Grade, High School Grad, Some College degree, College Graduate or above. Smoking status was coded as positive (“Yes”) for individuals reporting a lifetime consumption of ≥100 cigarettes, whereas those below this threshold were labeled as negative (“No”). Similarly, alcohol use was binary-coded based on a cutoff of 12 drinks annually. In addition, the Dietary Inflammatory Index (DII) was employed to characterize the inflammatory potential of habitual dietary patterns, where positive values denote a pro-inflammatory profile and negative values reflect anti-inflammatory properties. 23
Statistical analysis
To address the complex sampling architecture of NHANES, which involves multi-level stratification and clustering, all statistical procedures incorporated appropriate sample weights to ensure the findings remained unbiased and nationally representative. Continuous variables were reported as means with standard deviations (SD) or medians with interquartile ranges (IQR), whereas categorical data were summarized as weighted proportions. Baseline differences between groups were evaluated utilizing the Wilcoxon rank-sum test for continuous parameters and the Rao-Scott chi-square test for categorical distributions. The association of DI-GM with cognitive outcomes was examined via survey-weighted multivariable linear regression, with effect estimates presented as β coefficients and corresponding 95% confidence intervals (CIs). Model 1 was the unadjusted crude model. Model 2 was adjusted for sex, race and ethnicity, educational attainment, family poverty income ratio, body mass index, total calorie intake, smoking status and alcohol use. Model 3 was further incorporated additional adjustments for hypertension, diabetes mellitus, cardiovascular disease, and stroke. To detect potential non-linear dose–response relationships, we employed restricted cubic splines (RCS). Furthermore, the receiver operating characteristics (ROC) curves were established to assess the predictive value of DI-GM on cognitive function. Calibration curve was illustrated to assess the goodness-of-fit of the model. Model fit and net clinical benefit were verified using calibration plots and Decision Curve Analysis (DCA). Additionally, stratified analyses were performed across age, sex, hypertension, diabetes mellitus, cardiovascular disease, and stroke categories. Effect heterogeneity was evaluated using interaction terms and likelihood ratio testing. The potential mediating role of the Dietary Inflammatory Index (DII) was investigated using a bootstrapping approach with 1000 resamples.
All computations were performed using R (version 4.4.1) and IBM SPSS (version 27). Statistical significance was determined using a two-sided α level of 0.05.
Results
Characteristics of participants at the baseline
2629 participants were included in the final analytic sample (weighted population: 3,067,660; weighted mean age [SD], 69.6 [6.6] years; weighted proportion female, 53.5%). Baseline characteristics stratified by DI-GM quartiles are summarized in Table 1. Across DI-GM categories, statistically meaningful differences were observed for age, sex, race and ethnicity, total calories intake, BMI, smoking status, alcoholic use, diabetes, stroke, CERAD, AFT, DSST, z score, and DII.
Baseline characteristics of US population stratified by DI-GM quartiles, NHANES 2011 to 2014.
All means and SD for continuous variables and percentages for categorical variables were weighted. The DI-GM score comprises BGMS and UGMS, categorized into four groups: 0–3, 4, 5, and ≥6.
The weighted percentages may not reach exactly 100% due to missing data. The DI-GM score comprises BGMS and UGMS, categorized into four groups: 0–3, 4, 5, and ≥6.
DI-GM: dietary index for gut microbiota; NHANES: National Health and Nutrition Examination Survey; BGMS: beneficial to gut microbiota score; UGMS: unfavorable to gut microbiota score; CERAD: the Consortium to Establish a Registry for Alzheimer's Disease Word List Learning Test; AFT: the Animal Fluency Test; DSST: the Digit Symbol Substitution Test; DII, Dietary Inflammatory Index.
Relationship between DI-GM and cognitive function
Table 2 summarizes findings from the weighted multivariable linear regression examinations regarding the link between DI-GM and cognitive function. Overall, elevated DI-GM levels indicated a positive correlation with improved outcomes in specific cognitive domains. In the unadjusted model, participants with the highest DI-GM category (≥6) had higher cognitive scores than those with the lowest category (0–3) (CERAD: β = 0.90, 95% CI 0.17∼1.63; AFT: β = 1.37, 95% CI 0.75∼1.99; DSST: β = 5.60, 95% CI 3.68∼7.52; z score: β = 0.24, 95% CI 0.15∼0.33). As DI-GM increased, cognitive scores suggested a significant upward trend (p for trend < 0.001). After adjusting for all covariates, participants with DI-GM (≥ 6) were found with higher AFT score (β = 1.11, 95% CI 0.46∼1.75), DSST score (β = 4.95, 95% CI 3.05∼6.86) and z score (β = 0.19, 95% CI 0.10∼0.28), compared to those with DI-GM (0–3). However, the association with the CERAD test became non-significant in the fully adjusted model (β = 0.45, 95% CI −0.29∼1.18, p = 0.233). Notably, the positive correlation between BGMS and cognitive function measured by AFT [0.24 (0.11, 0.37)], DSST [1.08 (0.69, 1.46), and z score [0.05 (0.03, 0.07)] remains significant in the model 3, while UGMS was not independently associated with cognitive scores after controlling for covariates (Table 2). Furthermore, the RCS model confirmed a significant positive linear relationship between DI-GM and AFT, DSST, and global z scores after full adjustment for covariates, whereas the linear relationship with CERAD was not statistically significant (Figure 1).
Association between DI-GM and cognitive performance in CERAD, DSST, AFT and z score.
The associations between DI-GM with cognitive performance.
Model 1: No covariates were adjusted. Model 2: age, sex, race and ethnicity, educational attainment, family poverty income ratio, body mass index, total calories intake, smoking status, and alcohol use were adjusted. Model 3: Additionally adjusted for hypertension, diabetes, coronary heart disease and stroke.
Predictive performance
Figure 2 illustrated the predictive utility of DI-GM for cognitive function. The basic model incorporated variables including sex, race and ethnicity, educational attainment, family poverty income ratio, body mass index, total calorie intake, smoking status, alcohol use, hypertension, diabetes mellitus, cardiovascular disease, and stroke. We subsequently added the novel biomarker DI-GM to the basic model and conducted a comparative evaluation of the diagnostic capabilities of both models using ROC curve analysis (Figure 2A). Compared with the basic model, adding the DI-GM optimized the predictive ability by the area under the curve (AUC) (CERAD: 0.777 versus 0.718, p < 0.001; AFT: 0.698 versus 0.738, p = 0.019; DSST: 0.694 versus 0.767, p = 0.019). Calibration curve analysis revealed minimal deviation between the ideal and observed curves. The empirical prediction curve demonstrated excellent alignment with the theoretical ideal (B = 100 bootstrap repetitions; CERAD: mean absolute error = 0.006; AFT: mean absolute error = 0.007; DSST: mean absolute error = 0.008), indicating superior model accuracy (Figure 2B). Furthermore, DCA confirmed the clinical utility of the model, demonstrating a marked net benefit over a range of relevant threshold probabilities and suggesting meaningful utility for clinical decision-making (Figure 2C).
Predictive utility test of DI-GM for cognitive function. A. The area under the receiver operating characteristic (ROC) curve (AUC); B. Calibration curve; C. Decision curve analysis (DCA).
Subgroup analyses
Subgroup analyses were conducted based on covariates including age, sex, hypertension, diabetes, coronary heart disease and stroke. The stratified results for the relationship between DI-GM and cognitive performance are presented in Supplemental Table 2. Using DI-GM 0–3 as the reference category, participants with DI-GM ≥6 exhibited higher AFT, DSST, and standardized z score if the participants were aged 60–70 years and >70 years, both males and females, with and without hypertension, without diabetes, and without coronary heart disease or stroke (Table 3). Elevated DI-GM scores showed a significant positive correlation with better CERAD performance, if participants were aged 60–70 years, with and without stroke, without hypertension, diabetes or coronary heart disease. Notably, diabetes status was identified as a significant effect modifier in the association between DI-GM and DSST (p for interaction < 0.05). Moreover, no other participant characteristics or comorbid conditions showed statistically significant interaction with DI-GM.
Stratified analysis of the association between DI-GM and cognitive function.
Stratified analyses of the association between cognitive performance and DI-GM according to baseline characteristics in z score test. The p value for interaction represents the likelihood of interaction between the variable and DI-GM.
DI-GM: dietary index for gut microbiota; CI: confidence interval; CERAD: the Consortium to Establish a Registry for Alzheimer's Disease Word List Learning Test; AFT: the Animal Fluency Test; DSST: the Digit Symbol Substitution Test.
Mediation analysis
To elucidate the underlying mechanisms, we examined whether the Dietary Inflammatory Index (DII) serves as an intermediary in the DI-GM–cognition relationship. In multivariable linear regression, DII was inversely associated with cognitive function (CERAD: β = −1.75, 95% CI −2.81 ∼ −0.69; AFT: β = −2.21, 95% CI −3.15 ∼ −1.27; DSST: β = −9.52, 95% CI −12.31 ∼ −6.73; z score: β = −1.23, 95% CI −1.62 ∼ −0.84) (Supplemental Table 3). DII mediated 66.32% and 57.30% of the total effect of DI-GM on AFT and DSST scores, respectively (Figure 3B, C), whereas no mediation effect was detected for CERAD test (Figure 3A).
The mediating effect of DII on the relationship between DI-GM and cognitive function (A. CERAD; B. AFT; C. DSST).
Discussion
Utilizing data from the NHANES (2011–2014), this cross-sectional investigation identified a positive correlation between DI-GM and specific domains of cognitive performance within the elderly population. In multivariable-adjusted linear regression models, DI-GM was positively related to DSST, AFT, and the composite z score, although no significant association was established with the CERAD test. In addition, the RCS analysis confirmed a dose-dependent relationship between DI-GM and cognitive performance was observed among older adults (AFT, DSST, and global z score, nonlinear p > 0.05). The DI-GM significantly enhanced the predictive power of basic models for cognitive performance. No significant interactions were detected in the subgroup analysis. Moreover, the associations between DI-GM and both AFT and DSST were mediated by DII. These findings suggest that dietary patterns characterized by higher DI-GM scores may serve as a potential modifiable factor for supporting cognitive health, particularly among adults aged 60 years and older.
Recently, the relationship between cognitive ability and diet has been a widely discussed topic. Diets known for their anti-inflammatory profiles, specifically the MD and DASH, play a potential role in suppressing the neuronal inflammatory responses associated with the progression of AD. 24 These indices serve as valuable tools for evaluating overall dietary quality; however, they lack specificity in examining the links between dietary patterns and gut microbiota. In contrast, DI-GM is a dietary scoring system designed to quantify the intake of specific food components believed to influence the structure and function of the gut microbiota. Rather than directly assessing microbial diversity, short-chain fatty acid production (SCFAs), or phylum-level compositional shifts, DI-GM serves as an exposure indicator. It provides a structured approach to evaluate diet that can be combined with actual microbiome and metabolomic data to investigate diet–microbiota relationships.
Multiple studies have investigated the relationship between gut microbiota and cognitive performance.25–27 In AD, the gut community is commonly described as less diverse and taxonomically altered, with lower relative abundance of Firmicutes and Bifidobacterium and enrichment of Bacteroidetes. 28 Some meta-analyses have proposed that biotic interventions may ameliorate cognitive deficits.29,30
In our analysis, we similarly found that higher DI-GM scores, indicative of gut-friendly dietary patterns, were associated with better cognitive function. DI-GM exhibited positive associations with DSST, AFT, and a composite global z score (all p < 0.001). After full adjustment for covariates, DI-GM still remained a significant predictor of higher DSST, AFT, and overall z score performance. Notably, the association between DI-GM and the CERAD test became non-significant after full covariate adjustment (β = 0.45, 95% CI −0.29 ∼ 1.18, p = 0.233). This discrepancy between CERAD and DSST or AFT outcomes may suggest a domain-specific neuroprotective effect of the diet-gut microbiota axis. The CERAD test primarily assesses episodic memory, a cognitive function predominantly mediated by hippocampal integrity and closely linked to Alzheimer's disease-specific neuropathology, such as neurofibrillary tangles and amyloid-β deposition.31,32 In contrast, the DSST and AFT evaluate executive function and processing speed, which are more sensitive to vascular health, white matter integrity, and metabolic dysregulation—pathways directly influenced by gut microbiota-derived metabolites and systemic inflammation.33,34 Therefore, in this elderly population, the impact of DI-GM on cognitive aging may primarily originate from metabolic and vascular pathways rather than direct alterations in hippocampus-specific pathology. This explains why its association is more pronounced with DSST and AFT.
Our findings align with emerging evidence supporting the association between DI-GM and cognitive outcomes. Several studies have similarly highlighted the neuroprotective potential of dietary patterns conducive to gut microbiota health.35–38 However, the relationship appears complex and may not be uniform across all cognitive outcomes. Notably, Liu et al.[ 37 ] recently reported that a higher DI-GM score was associated with an increased prevalence of AD. They noted that certain foods traditionally considered “gut-healthy”, such as soy products and high fiber intake, may not uniformly benefit cognitive health in older adults. Furthermore, the cross-sectional nature of these studies precludes causal inference. Reverse causation remains a critical consideration, as dietary patterns may be altered by neurobehavioral changes in preclinical AD or by functional limitations associated with cognitive decline.
This study is, to our knowledge, the first to integrate the predictive performance of DI-GM for cognitive function in older adult by incorporating it into established baseline models. Our data indicate that DI-GM significantly improves the predictive accuracy of baseline models for cognitive function, suggesting its potential utility in future cognitive impairment risk stratification systems. Additionally, DII was evaluated as a potential mediator of the association between the DI-GM score and cognitive performance using mediation analysis. Our results support a partial mediation effect of DII in this association, suggesting that the neuroprotective properties of a microbiome-friendly diet are, to some extent, realized through the attenuation of systemic inflammation.
The observed protective correlation between DI-GM and cognitive function may be partially explained by the diet-microbiota-gut-brain axis, including neural, immune, endocrine, and metabolic signaling. Higher DI-GM reflects dietary patterns enriched in fermented products, polyphenol-containing foods, and fiber, which are conducive to SCFA production. 39 SCFAs are considered key effectors in regulating neuroinflammatory processes, including effects on cognition, mood, and other neuropsychiatric phenotypes. 40 SCFAs also interact with receptors on colon cells to indirectly signal to the central nervous system through receptor-mediated actions in the colon that promote secretion of gut hormones such as glucagon-like peptide 1 (GLP1) and peptide YY (PYY), which in turn affects learning, mood and memory.41,42 Fermented foods and dietary fiber, the beneficial component of DI-GM, may be associated with better maintenance of cognitive function in aging adults. 43 Elevated serum brain-derived neurotrophic factor levels were linked to cognitive enhancement following 12 weeks of fermented soybean consumption in older adults. 44 Higher intake of dietary fiber was reported to be related to improvements in specific components of cognitive function in individuals over the age of 60. 45 A high-fat diet, a unfavorable component of the DI-GM, has been observed to change Bacteroidota, increase Bacillota and Pseudomonadota, potentially causing adverse impacts on cerebral function. 46
In this study, we employed a comprehensive and multidimensional strategy to evaluate cognitive function, incorporating four validated neuropsychological tests: the CERAD, DSST, AFT, and a composite global z score. Key confounding variables were also controlled for in multivariable models to mitigate potential confounding. However, several limitations should be acknowledged. Firstly, the cross-sectional design precludes establishing temporal ordering or causal inference between DI-GM and cognitive performance, underscoring the need for prospective cohort studies. Secondly, the impact of unknown confounders or unmeasured variables cannot be entirely ruled out. Additionally, the selection of components in the DI-GM was constrained by existing evidence, while foods related to the gut microbiota that have not been deeply researched were not included in the index. Lastly, reliance on self-reported 24-h dietary recalls may introduce recall bias and exposure misclassification.
Conclusion
DI-GM, a novel diet-quality index developed to reflect dietary patterns linked to gut microbial diversity, was positively associated with cognitive function by utilizing a nationally representative cohort of older adults in the USA. Notably, incorporating DI-GM into the multivariate risk model improved the prediction of cognitive function. Furthermore, DII may partly account for the association between DI-GM and cognitive outcomes. Of course, more research is needed in the future to explain these associations. Our findings support the need for further investigation into gut-microbiota-focused dietary patterns as a potential approach to maintaining cognitive health in aging populations.
Supplemental Material
sj-docx-1-alz-10.1177_13872877261449420 - Supplemental material for Association between the Dietary Index for Gut Microbiota and cognitive function among older adults in the United States
Supplemental material, sj-docx-1-alz-10.1177_13872877261449420 for Association between the Dietary Index for Gut Microbiota and cognitive function among older adults in the United States by Ke Si, Cunwei Sun, Han Guo, Chuanqin Shi and Yangang Wang in Journal of Alzheimer's Disease
Footnotes
Acknowledgements
We thank all participants in the NHANES databases.
Ethical considerations
The NCHS Research Ethics Review Board approved the study protocol.
Consent to participate
Written informed consent was obtained from all NHANES participants.
Consent for publication
Not applicable
Author contribution(s)
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0523500) and the Shandong Provincial Integrated Traditional Chinese and Western Medicine Special Disease Prevention and Treatment Project (S19-0009280000).
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data availability statement
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
