The atherogenic index of plasma and triglyceride glucose-body mass index are inversely associated with cognitive impairment in middle-aged and older Chinese adults: A community-based cohort study

Abstract

Background

Although metabolic disorders are associated with cognitive dysfunction, the relationship with composite metabolic indices remains unclear.

Objective

This study aimed to examine the longitudinal associations of the atherogenic index of plasma (AIP) and the triglyceride-glucose body mass index (TyG-BMI) with incident cognitive impairment in a community-based cohort of middle-aged and older adults.

Methods

This prospective cohort study included 1492 cognitively normal adults (age ≥40 years). Baseline AIP and TyG-BMI were calculated from blood samples. Incident cognitive impairment was defined using education-specific Mini-Mental State Examination cut-offs at a 2-year follow-up. Associations were evaluated using multivariable logistic regression (adjusted for age, sex, education, apolipoprotein E ε4, and other confounders) and restricted cubic spline models.

Results

Among participants, 59.4% were female, with a median age of 55 years (IQR, 48–63). Over a median follow-up of 2.02 years (IQR, 1.98–2.06), 133 participants (8.9%) developed cognitive impairment. After full adjustment, higher levels of AIP (OR = 0.21, 95% CI: 0.09–0.50) and TyG-BMI (OR = 0.99, 95% CI: 0.98–0.99) were inversely associated with the risk of cognitive impairment (both p < 0.001). A dose-response relationship was observed, with risk decreasing as index levels increased.

Conclusions

In this middle-aged and older Chinese cohort, higher AIP and TyG-BMI were associated with a reduced short-term risk of cognitive impairment. These findings highlight the complex, context-dependent nature of metabolic-cognitive associations, suggesting these indices may reflect a metabolic profile relevant to early cognitive maintenance.

Keywords

Alzheimer's disease apolipoprotein E atherogenic index of plasma cognitive impairment triglyceride glucose-body mass index

Introduction

The aging of the worldwide populace has resulted in an increasing incidence of cognitive impairment and dementia, imposing substantial social and economic burdens worldwide.¹ Cognitive impairment, a prodromal stage of dementia characterized by deficits in memory, orientation, and executive function, arises from a multifactorial interplay of genetic, vascular, metabolic, and lifestyle determinants.^2,3 Timely recognition of modifiable characteristics is thus vital for the development of effective preventative strategies. Among these, metabolic factors are of particular interest given their potential for intervention and clinical monitoring.

Biochemical conditions, featuring dyslipidemia and insulin resistance, are well-established contributors to cognitive decline.^4,5 Computed as the logarithmic value of the triglyceride (TG) to high-density lipoprotein cholesterol (HDL-C) ratio, the atherogenic index of plasma (AIP) is a robust marker of metabolic dysregulation, linked to cardiovascular disease, atherosclerosis, and cerebrovascular pathology.^6–9 Nevertheless, evidence regarding its relationship with cognitive impairment remains limited and inconsistent, suggesting that the connection between lipid metabolism and neurological health is far more multifaceted than originally conceived.

Similarly, the triglyceride-glucose (TyG) index, a surrogate marker of insulin resistance derived from fasting TG and glucose levels, correlates with cognitive impairment and dementia.^10,11 Once multiplied by body mass index (BMI) to generate the TyG-BMI index, it offers a more integrated measure of metabolic obesity and systemic lipid-glucose homeostasis.¹² However, its association with cognitive outcomes, particularly in longitudinal cohorts, remains inadequately characterized.

To address these evidence gaps, a community-focused, prospective cohort investigation of midlife and senior Chinese individuals was performed in the present research, with the objective of assessing the associations of AIP and TyG-BMI with incident cognitive impairment. The findings refine the understanding of the contribution of these metabolic indices to cognitive decline and may inform early risk stratification, screening, and preventive strategies.

Methods

Study population

Between October 2014 and March 2015, we conducted a community-based prospective cohort study in Qubao Village, near Xi’an, China, using cluster sampling to recruit participants from the general community. Qualified individuals were obligated to satisfy the subsequent inclusion criteria: (1) age ≥40 years; (2) permanent residency in the village for at least three years; and (3) provision of written informed consent. A total of 2158 individuals were initially enrolled, providing a well-defined sample from a rural community in the Xi’an region. All participants completed a standardized face-to-face questionnaire and provided blood samples at baseline.

A follow-up assessment, including repeated cognitive evaluation, was conducted at the 2-year follow-up visit, with a permissible window of ±4 weeks. To enhance the validity of incident cognitive impairment analysis, these subsequent removal standards were implemented: (1) missing baseline data on TG, HDL-C, FBG, or other covariates; (2) incomplete baseline or follow-up cognitive data; (3) presence of mild cognitive impairment (MCI) or dementia at baseline; (4) self-reported history of conditions that could interfere with cognitive assessment, such as severe neurological, psychiatric, or sensory disorders; and (5) diagnosis of serious cardiac, pulmonary, hematologic, hepatic, or renal disease, or any malignant tumor. This screening process yielded a final sample of 1492 participants for analysis (Figure 1).

Figure 1.

Flowchart of study enrollment and exclusion.

Anthropometric measurements and laboratory assessments

Each individual donned lightweight garments excluding footgear during weight measurement. The BMI was calculated via weight/height squared (kg/m²). Blood pressure was recorded by an experienced nurse employing a hand-held liquid-filled sphygmomanometer using standard full-size cuff at a preferred arm between 8:00 and 10:00 AM. Resting heart rate (RHR) was assessed by palpating the radial artery.

Fasted blood specimens, gathered following nocturnal abstinence, were analyzed for FBG, total cholesterol (TC), TG, low-density lipoprotein cholesterol (LDL-C), and HDL-C employing an enzyme-based technique via an automated clinical analyzer (C501, Roche, Sweden). AIP was defined as the logarithm to the base 10 of [TG (mol/L)/HDL-C (mol/L)]. TyG-BMI was defined as the natural logarithm of [TG (mg/dL) × FBG (mg/dL)/2)] × BMI (kg/m²).

Cognitive assessment

Cognitive function was assessed at baseline and follow-up using the Chinese version of the Mini-Mental State Examination (MMSE),¹³ which scores from 0 to 30, with higher scores indicating better cognitive function. All assessments were conducted by trained neurologists and neurology postgraduate students following standardized protocols. To account for the well-established influence of education on MMSE performance,¹⁴ cognitive impairment was defined using education-specific cut-offs validated in the Shanghai Dementia Survey.¹⁵ Specifically, the cut-off values were as follows: ≤17 for individuals with no formal education, ≤20 for those with primary education (≤6 years), and ≤24 for those with secondary or higher education (>6 years). These established cut-offs align with nationwide normative studies in China.¹⁶ Incident cognitive impairment was defined as an MMSE score below the corresponding education-specific threshold at follow-up in participants who were cognitively normal at baseline.^17,18

APOE genotype

APOE genotyping was performed on DNA extracted from peripheral blood leukocytes. Briefly, target polymorphisms (rs7412 and rs429358) were amplified by PCR and analyzed via Sanger sequencing (Sangon Biotech, Shanghai, China) to determine ε2, ε3, and ε4 alleles. Participants were then categorized as APOE ε4 carriers or non-carriers.

Other covariates

Data on demographic characteristics, medical history, lifestyle habits, and medication use were collected by qualified interviewers employing a standardized questionnaire. Demographic information comprised age, gender, and educational attainment (years). Medical records addressed hypertension, diabetes mellitus, dyslipidemia, coronary heart disease, and a record of transient ischemic attack (TIA) or stroke. Lifestyle factors assessed included smoking status, alcohol intake, and physical activity levels. The definitions of medical diseases were aligned with those applied in our previous study.¹⁹

Statistical analysis

Continuous variables were presented as mean (standard deviation) or median (interquartile range), and categorical variables as count (percentage). Between-group differences were assessed using one-way analysis of variance (ANOVA) or Kruskal-Wallis tests for continuous variables and chi-squared tests for categorical variables.

To evaluate the association between metabolic indices and cognitive impairment, participants were stratified into quartiles (Q1–Q4) based on AIP and TyG-BMI for incidence comparison. Logistic regression was used to calculate odds ratios (ORs) with 95% confidence intervals (CIs). Consistent with established analytical practices,^20–23 both indices were analyzed not only as continuous variables but also in quartile form. This integrated approach leverages the complementary strengths of each method: continuous analysis maximizes statistical power to detect linear associations, while quartile-based analysis can reveal non-linear patterns and facilitates clinical interpretation through risk comparisons (e.g., Q4 versus Q1). Three models were fitted: Model 1 was unadjusted; Model 2 was adjusted for age, sex, years of education, and APOE ε4 carrier status; and Model 3 was additionally adjusted for smoking, drinking, physical activity, hypertension, diabetes mellitus, dyslipidemia, coronary heart disease, and history of TIA or stroke.

Restricted cubic spline (RCS) logistic regression using four knots (at the 5th, 35th, 65th, and 95th percentiles) was conducted to further examine nonlinear associations between the AIP and TyG-BMI indices and cognitive impairment.

Subgroup analyses were conducted to explore potential effect modification by age, sex, APOE ε4 carrier status, BMI, and lifestyle factors. Effect modification was tested by adding multiplicative interaction terms between each exposure (AIP, TyG-BMI) and the subgroup variable in the regression models. Sensitivity analyses assessed robustness by (i) excluding participants using lipid-lowering or glucose-lowering medications and (ii) fitting alternative model specifications to check for potential multicollinearity.

All statistical analyses were performed using R software (version 4.2.1; R Core Team, 2022) and IBM SPSS Statistics (version 20.0). Statistical significance was set at a two-sided p of < 0.05.

Results

Baseline characteristics of study population

The final analysis included 1492 participants aged 40–84 years (median age, 55 years; IQR, 48–63), with 886 (59.4%) being women. They were stratified by their AIP into four quartile-based groups. As expected, those in the highest AIP category showed an increased prevalence of hypertension, diabetes mellitus, and dyslipidemia. Across increasing AIP quartiles (Q1 to Q4), significant increases were observed in BMI, DBP, FBG, TG, TC, LDL-C, and TyG-BMI, while HDL-C showed a significant decrease. No significant differences across quartiles were observed for age, sex, educational level, APOE ε4 carrier status, or lifestyle factors (Table 1).

Table 1.

Baseline characteristics of participants by quartiles of the AIP.

Characteristics	Total	AIP level				p
Characteristics	Total	Q1 (←0.16)	Q2 (−0.16∼0.02)	Q3 (0.02∼0.20)	Q4 (>0.20)	p
Number	1492	373	373	373	373
Age, years	55 (48, 63)	53 (48, 62)	55 (48, 63)	56 (49, 63)	54 (48, 63)	0.065
Male, n (%)	606 (40.6)	145 (38.9)	135 (36.2)	157 (42.1)	169 (45.3)	0.065
Education, years	7 (4, 9)	8 (5, 9)	7 (4, 8)	7 (4, 9)	8 (5, 9)	0.154
Hypertension, n (%)	757 (50.7)	135 (36.2)	175 (46.9)	222 (59.5)	225 (60.3)	<0.001
Diabetes mellitus, n (%)	200 (13.4)	24 (6.4)	39 (10.5)	63 (16.9)	74 (19.8)	<0.001
Dyslipidemia, n (%)	975 (65.4)	135 (36.2)	193 (51.7)	274 (73.5)	373 (100.0)	<0.001
Coronary heart disease, n (%)	92 (6.2)	20 (5.4)	17 (4.6)	27 (7.2)	28 (7.5)	0.263
TIA or stroke, n (%)	117 (7.8)	30 (8.0)	30 (8.0)	30 (8.0)	27 (7.2)	0.969
Smoking, n (%)	412 (27.6)	107 (28.7)	101 (27.1)	93 (24.9)	111 (29.8)	0.481
Drinking, n (%)	222 (14.9)	53 (14.2)	55 (14.8)	55 (14.8)	59 (15.8)	0.940
Lack of physical activity, n (%)	256 (17.2)	65 (17.4)	58 (15.6)	60 (16.1)	73 (19.6)	0.470
BMI, (kg/m²)	25.4 (3.1)	24.0 (2.9)	25.1 (3.1)	25.9 (3.2)	26.4 (2.9)	<0.001
RHR, (b.p.m.)	75.4 (9.2)	75.1 (8.9)	75.8 (9.8)	74.7 (8.5)	76.0 (9.5)	0.173
SBP, mmHg	132.1 (18.3)	127.3 (16.9)	130.5 (17.7)	135.5 (20.1)	135.3 (17.0)	<0.001
DBP, mmHg	82.0 (10.1)	78.9 (9.3)	81.1 (9.3)	83.8 (10.7)	84.2 (10.2)	<0.001
FBG, mmol/L	5.40 (5.07, 5.86)	5.35 (5.01, 5.67)	5.37 (5.03, 5.75)	5.45 (5.09, 5.87)	5.55 (5.15, 6.22)	<0.001
TG, mmol/L	1.44 (1.04, 2.01)	0.85 (0.67, 0.99)	1.25 (1.09, 1.41)	1.65 (1.46, 1.85)	2.49 (2.18, 3.20)	<0.001
TC, mmol/L	5.02 (0.98)	4.80 (0.86)	4.94 (0.91)	5.07 (0.94)	5.27 (1.14)	<0.001
LDL-C, mmol/L	3.29 (0.86)	2.96 (0.74)	3.21 (0.81)	3.43 (0.82)	3.57 (0.93)	<0.001
HDL-C, mmol/L	1.40 (0.31)	1.68 (0.29)	1.48 (0.25)	1.30 (0.21)	1.14 (0.20)	<0.001
TyG-BMI	222.50 (34.08)	195.38 (25.28)	216.37 (28.09)	230.58 (30.29)	247.67 (28.88)	<0.001
APOE ε4 carriers, n (%)	226 (15.2)	48 (12.9)	49 (13.1)	58 (15.6)	71 (19.0)	0.068

Data are mean (standard deviation), median (interquartile range) or number (percentage).

AIP: atherogenic index of plasma; TIA: transient ischemic attack; BMI: body mass index; RHR: resting heart rate; SBP: systolic blood pressure; DBP: diastolic blood pressure; FBG: fasting blood glucose; TG: triglyceride; TC: total cholesterol; LDL-C: low-density lipoprotein cholesterol; HDL-C: high-density lipoprotein cholesterol; TyG-BMI: triglyceride glucose-body mass index.

Incidence rate of cognitive impairment

Over a median follow-up of 2.02 years (IQR, 1.98–2.06), 133 participants developed cognitive impairment (cumulative incidence, 8.9%). Incidence declined monotonically with increasing quartiles of both AIP and TyG-BMI (p for trend < 0.001; Figure 2A, B). For AIP, cumulative incidence across quartiles was 12.1% (Q1), 11.0% (Q2), 7.2% (Q3), and 5.4% (Q4). A similar gradient was observed for TyG-BMI: 12.1% (Q1), 10.7% (Q2), 7.5% (Q3), and 5.4% (Q4). In both indices, the fourth quartile (Q4) exhibited the lowest incidence.

Figure 2.

Incidence of cognitive impairment across quartiles of the AIP and TyG-BMI. AIP: atherogenic index of plasma; TyG-BMI: triglyceride glucose-body mass index.

Associations of the AIP and TyG-BMI with cognitive impairment

Multivariable logistic regression results are summarized in Table 2. When modeled continuously, both AIP and TyG-BMI were inversely associated with incident cognitive impairment in the fully adjusted model (Model 3): AIP, OR = 0.21 (95% CI 0.09–0.50; p < 0.001); TyG-BMI, OR = 0.99 (95% CI 0.98–0.99; p < 0.001). Quartile analyses were concordant: relative to Q1, participants in Q3 and Q4 for both indices had significantly lower odds (all p < 0.05). Restricted cubic spline models (Figure 3) indicated an approximately linear, monotonic decrease in risk with increasing AIP and TyG-BMI, with no evidence of nonlinearity (p_non−linear = 0.311 for AIP; p_non−linear = 0.278 for TyG-BMI).

Figure 3.

Restricted cubic spline plots of AIP/TyG-BMI with cognitive impairment risk. Odds ratios (solid lines) and their 95% CIs (dashed lines) are derived from logistic regression models adjusted for age, sex, years of education, APOE ε4 carrier status, smoking, drinking, physical activity level, hypertension, diabetes mellitus, dyslipidemia, coronary heart disease, TIA or stroke. AIP: atherogenic index of plasma; TyG-BMI: triglyceride glucose-body mass index; TIA: transient ischemic attack.

Table 2.

Associations of AIP and TyG-BMI with incident cognitive impairment.

	Model 1		Model 2		Model 3
	OR (95% CI)	p	OR (95% CI)	p	R (95% CI)	p
AIP	0.29 (0.15 ∼ 0.59)	<0.001	0.25 (0.12 ∼ 0.52)	<0.001	0.21 (0.09 ∼ 0.50)	<0.001
Q1	Ref		Ref		Ref
Q2	0.90 (0.57 ∼ 1.41)	0.647	0.78 (0.49 ∼ 1.24)	0.294	0.76 (0.47 ∼ 1.23)	0.270
Q3	0.57 (0.34 ∼ 0.94)	0.027	0.50 (0.30 ∼ 0.83)	0.008	0.45 (0.26 ∼ 0.79)	0.005
Q4	0.41 (0.24 ∼ 0.71)	0.002	0.38 (0.22 ∼ 0.66)	<0.001	0.34 (0.18 ∼ 0.65)	<0.001
p for trend	<0.001		<0.001		<0.001
TyG-BMI	0.99 (0.98 ∼ 0.99)	<0.001	0.99 (0.98 ∼ 0.99)	<0.001	0.99 (0.98 ∼ 0.99)	<0.001
Q1	Ref		Ref		Ref
Q2	0.88 (0.56 ∼ 1.38)	0.565	0.81 (0.51 ∼ 1.30)	0.378	0.81 (0.50 ∼ 1.31)	0.393
Q3	0.59 (0.36 ∼ 0.97)	0.038	0.58 (0.35 ∼ 0.97)	0.039	0.56 (0.33 ∼ 0.96)	0.034
Q4	0.41 (0.24 ∼ 0.71)	0.002	0.37 (0.21 ∼ 0.65)	<0.001	0.35 (0.19 ∼ 0.65)	<0.001
p for trend	<0.001		<0.001		<0.001

Model 1: unadjusted.

Model 2: adjusted for age, sex, years of education, APOE ε4 carrier status.

Model 3: adjusted for age, sex, years of education, APOE ε4 carrier status, smoking, drinking, physical activity level, hypertension, diabetes mellitus, dyslipidemia, coronary heart disease, TIA or stroke.

OR: odds ratio; CI: confidence; Ref: reference; AIP: atherogenic index of plasma; TyG-BMI: triglyceride glucose-body mass index; TIA: transient ischemic attack.

Subgroup and interaction analyses

Subgroup analyses evaluated the consistency of the protective associations across various population strata (Figures 4 and 5). A consistent inverse association between AIP and cognitive impairment was observed across most subgroups. Notably, the association was attenuated and not statistically significant among APOE ε4 carriers and alcohol consumers. Similarly, higher TyG-BMI was inversely associated with cognitive impairment in most subgroups. This association, however, did not reach statistical significance in the following subgroups: adults ≥65 years, women, APOE ε4 carriers, individuals with BMI <24 kg/m², or those reporting alcohol use. Importantly, multiplicative interaction tests for both metabolic indices (AIP and TyG-BMI) with each subgroup variable were not significant (all p for interaction > 0.05), indicating no detectable effect modification of the overall protective associations.

Figure 4.

Subgroup and interaction analysis of the association between AIP and cognitive impairment. AIP: atherogenic index of plasma; BMI: body mass index; OR: odds ratios; CI: confidence intervals.

Figure 5.

Subgroup and interaction analysis of the association between TyG-BMI and cognitive impairment. TyG-BMI: triglyceride glucose-body mass index; BMI: body mass index; OR: odds ratios; CI: confidence intervals.

Sensitivity analyses

A series of sensitivity analyses supported the robustness of the main results. First, after excluding participants taking lipid-lowering or glucose-lowering medications (n = 1353), the inverse associations for both AIP and TyG-BMI remained statistically significant and essentially unchanged in magnitude (Supplemental Table 1). Second, to assess potential multicollinearity, we additionally adjusted the AIP model for BMI and FBG, and the TyG-BMI model for LDL-C and HDL-C, which did not alter the significant associations (data not shown).

Discussion

In this community-based cohort of predominantly middle-aged adults, higher baseline levels of the AIP and the TyG-BMI were inversely associated with the risk of incident cognitive impairment over a 2-year follow-up. This association was approximately linear, robust to multivariable adjustment, and consistent across sensitivity analyses.

The interpretation of these findings is best framed within the specific context of our study design and cohort. The relatively young age and preserved cognitive health of participants at baseline, combined with a short follow-up period, position this study to investigate metabolic correlates of very early cognitive dynamics, potentially relevant to cognitive maintenance, rather than long-term dementia pathogenesis. Consequently, the observed inverse associations may reflect a metabolic-nutritional state pertinent to preserving cognitive health in midlife, a stage where the biological implications of certain metabolic profiles may differ from those observed in older or clinically compromised populations.

In our cohort, a higher AIP was associated with a lower risk of cognitive impairment. This finding stands in contrast to reports from the China Health and Retirement Longitudinal Study (CHARLS), which linked elevated AIP to an increased risk of cognitive decline in an older population.^24,25 This apparent discrepancy likely underscores the context-dependent biological significance of lipid profiles. In our predominantly middle-aged, cognitively healthy community sample, the observed near-linear dose-response relationship may be indicative of a physiological lipidomic state associated with adequate nutritional intake and a specific lipid composition. Supporting this view, certain triglyceride subclasses, such as those rich in long-chain polyunsaturated fatty acids, have been correlated with reduced atrophy in brain regions critical for memory and may exert neuroprotective effects.²⁶ Thus, in this specific context, a higher AIP may partly capture a lipid signature associated with cognitive reserve. Conversely, in the CHARLS cohort—which is nationally representative, older, and carries a higher burden of comorbidities over a longer follow-up period—an elevated AIP is more likely to represent a sustained pro-atherogenic and pro-inflammatory pathological state, driven by metabolic dysfunction. These studies are therefore complementary, illustrating that the association between AIP and cognitive function is dynamic and varies across the health-disease spectrum and the aging trajectory.

Our results are consistent with prior observations derived from this same cohort, which showed an inverse relationship between serum TG and cognitive impairment in males,²⁷ and with other studies suggesting that higher TG, but not necessarily HDL-C, may be linked to better cognitive outcomes in some older adult cohorts.^28,29 Collectively, this evidence points to the possibility that, in relatively healthy midlife populations, the TG component within AIP (rather than the atherogenic ratio per se) might be a relevant factor linked to short-term cognitive preservation.

Similarly, in our cohort, a higher TyG-BMI index was associated with a lower risk of cognitive impairment. This finding differs from the positive association reported between the TyG index alone and cognitive impairment.³⁰ This divergence highlights how composite metabolic indices can encapsulate distinct biological information compared to their individual components.³¹ The integration of adiposity (BMI) expands the index beyond insulin resistance (TyG) alone, potentially reflecting a broader metabolic-nutritional state. Our results align with a recent study demonstrating that a higher TyG-BMI was associated with slower cognitive decline, a more favorable Alzheimer's disease cerebrospinal fluid biomarker profile (higher Aβ₄₂ and lower tau/p-tau), and less neurodegeneration in key brain regions such as the entorhinal cortex and middle temporal lobe.³² This suggests that in specific populations, TyG-BMI may signify a metabolic state with protective effects against AD pathogenesis.

The protective association of TyG-BMI, in contrast to the risk often linked to TyG alone, can be meaningfully interpreted within the conceptual frameworks of the “obesity paradox” and “metabolically healthy obesity” observed in aging.³³ The inclusion of BMI in the composite formula may be key, and its potential neuroprotective role in this context could be explained by several interrelated pathways. First, in a relatively healthy community sample, a higher BMI may partly reflect greater lean body mass and nutritional reserves, which could buffer against metabolic stress during aging.^34,35 Second, adipose tissue secretes adipokines with potential neuroprotective properties, such as leptin and adiponectin^36,37; a higher TyG-BMI might thus correlate with a more favorable adipokine milieu. Furthermore, neuroimaging evidence suggests a possible central mechanism, wherein higher BMI is associated with specific alterations in functional connectivity within the default mode network—a pattern linked to attenuated mental disengagement and potentially adaptive network efficiency.³⁸ Therefore, within the specific context of our cohort, TyG-BMI may identify a composite metabolic-nutritional state—integrating adequate energy reserves, a potentially favorable adipokine profile, and specific lipid-glucose metabolism—that is conducive to short-term cognitive maintenance.

We further examined whether the APOE genotype, the strongest known genetic risk factor for AD, modified the observed associations. Only minor differences were noted between APOE ε4 carriers and non-carriers, and multiplicative interaction terms for both AIP and TyG-BMI were non-significant. This indicates no detectable effect modification by APOE status in our study, suggesting that the protective associations of these metabolic indices may operate independently of APOE-related lipid transport pathways. Nevertheless, larger studies specifically focused on APOE ε4 carriers are warranted to confirm this finding.

Study strengths and limitations

This study has several strengths. Its prospective cohort design establishes temporality between the metabolic indices and cognitive outcomes, providing stronger evidence for potential causality than cross-sectional studies. The rigorous adjustment for key covariates and comprehensive sensitivity analyses enhanced the robustness and validity of the findings. To our knowledge, this is also the first study to investigate the potential moderating role of APOE genotype in the associations between these metabolic indices (AIP and TyG-BMI) and cognitive impairment.

Several limitations should also be considered. First, the relatively young age of the cohort and the short follow-up period confine our observations to early cognitive aging dynamics, limiting insights into the long-term roles of AIP or TyG-BMI in dementia pathogenesis. Second, AIP and TyG-BMI were measured only at baseline; repeated measurements would better capture their temporal dynamics and relationship with cognitive trajectories. Third, as participants were recruited from a single region in northern China, the generalizability of our findings to other populations or ethnic groups requires further validation. Finally, the precise biological mechanisms through which AIP or TyG-BMI might confer neuroprotection remain unclear, necessitating further experimental and clinical investigation.

Conclusion

In this community-based cohort of middle-aged and older adults, higher baseline levels of AIP and TyG-BMI were associated with a lower risk of incident cognitive impairment over a 2-year follow-up. This inverse association suggests that within a relatively healthy population at an early stage of cognitive aging, these composite metabolic indices may reflect a specific metabolic-nutritional profile potentially relevant to short-term cognitive maintenance rather than long-term pathology. Our findings therefore underscore the complex and stage-specific relationship between metabolism and cognitive health, highlighting the necessity of considering population characteristics, health status, and the stage of cognitive trajectory when interpreting metabolic risk markers.

Supplemental Material

sj-docx-1-alz-10.1177_13872877261434985 - Supplemental material for The atherogenic index of plasma and triglyceride glucose-body mass index are inversely associated with cognitive impairment in middle-aged and older Chinese adults: A community-based cohort study

Supplemental material, sj-docx-1-alz-10.1177_13872877261434985 for The atherogenic index of plasma and triglyceride glucose-body mass index are inversely associated with cognitive impairment in middle-aged and older Chinese adults: A community-based cohort study by Yu Jiang, Xiaowen Zhu, Lili Zhao, Yun Du, Heying Wang, Hong Fan, Kaige Ma, Haixia Lu, Dafei Ma, Jian Qiu and Guilian Zhang in Journal of Alzheimer's Disease

Footnotes

Acknowledgements

We sincerely thank all participants for their time and contributions. We also extend our gratitude to the clinical and research staff for their dedicated support.

ORCID iDs

Yu Jiang

Guilian Zhang

Ethical considerations

The study protocol was reviewed and approved by the Ethics Committee of The First Affiliated Hospital of Xi’an Jiaotong University (Approval No. XJTU1AF2014LSK-111). All procedures were conducted in accordance with the ethical standards of the Declaration of Helsinki.

Consent to participate

Written informed consent was obtained from all individual participants. The study's purpose, procedures, potential risks, and benefits were fully explained to them, and their participation was entirely voluntary.

Consent for publication

All participants provided written informed consent for the use of their de-identified data in academic research and subsequent publication.

Author contribution(s)

Yu Jiang: Conceptualization; Data curation; Funding acquisition; Resources; Writing – original draft.

Xiaowen Zhu: Conceptualization; Data curation; Methodology; Visualization; Writing – original draft.

Lili Zhao: Formal analysis; Writing – original draft.

Yun Du: Formal analysis; Writing – original draft.

Heying Wang: Formal analysis; Writing – original draft.

Hong Fan: Formal analysis; Writing – original draft.

Kaige Ma: Investigation; Writing – review & editing.

Haixia Lu: Investigation; Writing – review & editing.

Dafei Ma: Investigation; Writing – review & editing.

Jian Qiu: Investigation; Writing – review & editing.

Guilian Zhang: Conceptualization; Project administration; Supervision; Writing – review & editing.

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 Natural Science Basic Research Program of Shaanxi Province (Grant No. 2023-JC-QN-0919) and the Institutional Fund of The Second Affiliated Hospital of Xi’an Jiaotong University (Grant Nos. YJ(QN)201908 and GKJTSJH-16).

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

The datasets generated during and/or analyzed during the current study are available from the corresponding author on request.

Supplemental material

Supplemental material for this article is available online.

References

GBD 2019 Dementia Forecasting Collaborators. Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019. Lancet Public Health 2022; 7: e105–e125.

Baumgart

Snyder

Carrillo

, et al. Summary of the evidence on modifiable risk factors for cognitive decline and dementia: a population-based perspective. Alzheimers Dement 2015; 11: 718–726.

Juul Rasmussen

Rasmussen

Nordestgaard

, et al. Impact of cardiovascular risk factors and genetics on 10-year absolute risk of dementia: risk charts for targeted prevention. Eur Heart J 2020; 41: 4024–4033.

Yin

Zhu

, et al. Blood cholesterol in late-life and cognitive decline: a longitudinal study of the Chinese elderly. Mol Neurodegener 2017; 12: 24.

Kim

Arvanitakis

. Insulin resistance, cognition, and Alzheimer disease. Obesity (Silver Spring) 2023; 31: 1486–1498.

Kao

Chang

, et al. Atherogenic index of plasma as predictors for metabolic syndrome, hypertension and diabetes mellitus in Taiwan citizens: a 9-year longitudinal study. Sci Rep 2021; 11: 9900.

Kim

Cho

Kim

, et al. Association of the atherogenic index of plasma with cardiovascular risk beyond the traditional risk factors: a nationwide population-based cohort study. Cardiovasc Diabetol 2022; 21: 81.

Huang

Liu

Wei

, et al. The atherogenic index of plasma and carotid atherosclerosis in a community population: a population-based cohort study in China. Cardiovasc Diabetol 2023; 22: 125.

Zheng

, et al. Relationship between the cumulative exposure to atherogenic index of plasma and ischemic stroke: a retrospective cohort study. Cardiovasc Diabetol 2023; 22: 313.

10.

Wang

Ling

, et al. Association between the triglyceride glucose index and cognitive impairment and dementia: a meta-analysis. Front Aging Neurosci 2023; 15: 1278730.

11.

Liu

Yang

Gao

, et al. Association between triglyceride glucose index and cognitive decline: a meta-analysis. J Affect Disord 2024; 359: 70–77.

12.

Cui

, et al. Comparative study of insulin resistance surrogate indices to predict mild cognitive impairment among Chinese non-diabetic adults. Lipids Health Dis 2024; 23: 357.

13.

Zhang

Katzman

Salmon

, et al. The prevalence of dementia and Alzheimer's disease in Shanghai, China: impact of age, gender, and education. Ann Neurol 1990; 27: 428–437.

14.

Crum

Anthony

Bassett

, et al. Population-based norms for the Mini-Mental State Examination by age and educational level. JAMA 1993; 269: 2386–2391.

15.

Katzman

Zhang

Ouang-Ya-Qu , et al. A Chinese version of the Mini-Mental State Examination; impact of illiteracy in a Shanghai dementia survey. J Clin Epidemiol 1988; 41: 971–978.

16.

Jia

Yang

. Mini-Mental state examination in elderly Chinese: a population-based normative study. J Alzheimers Dis 2016; 53: 487–496.

17.

Chen

Feng

, et al. Trends in the prevalence of cognitive impairment at old age in China, 2002-2018. Alzheimers Dement 2024; 20: 1387–1396.

18.

Lin

Zhou

Liu

. Association of adherence to the Chinese version of the MIND diet with reduced cognitive decline in older Chinese individuals: analysis of the Chinese Longitudinal Healthy Longevity Survey. J Nutr Health Aging 2024; 28: 100024.

19.

Jiang

Shang

, et al. Pulse pressure is associated with plasma amyloid-β transport dysfunction. J Hypertens 2018; 36: 569–579.

20.

Huang

Chen

Peng

, et al. Associations between six cumulative insulin resistance-related indices and incident stroke in individuals with different glucose regulation statuses: a national cohort study in China. Lipids Health Dis 2025; 24: 368.

21.

Jiang

Zhu

Song

, et al. Assessment of six insulin resistance surrogate indexes for predicting stroke incidence in Chinese middle-aged and elderly populations with abnormal glucose metabolism: a nationwide prospective cohort study. Cardiovasc Diabetol 2025; 24: 56.

22.

Tao

, et al. Multiple triglyceride-derived metabolic indices and incident cardiovascular outcomes in patients with type 2 diabetes and coronary heart disease. Cardiovasc Diabetol 2024; 23: 359.

23.

Jiang

, et al. Association between insulin resistance indices and outcomes in patients with heart failure with preserved ejection fraction. Cardiovasc Diabetol 2025; 24: 32.

24.

Zhou

Han

Bai

. Association between atherogenic index of plasma and cognitive impairment in middle-aged and older adults: results from CHARLS. Front Aging Neurosci 2025; 17: 1506973.

25.

Lin

. Association between atherogenic index of plasma and dementia: a longitudinal observational study. J Alzheimers Dis 2025; 106: 197–205.

26.

Bernath

Bhattacharyya

Nho

, et al. Serum triglycerides in Alzheimer disease: relation to neuroimaging and CSF biomarkers. Neurology 2020; 94: e2088–e2098.

27.

Zhao

Shang

, et al. The gender- and age- dependent relationships between serum lipids and cognitive impairment: a cross-sectional study in a rural area of Xi'an, China. Lipids Health Dis 2019; 18: 4.

28.

Zhou

Ryan

Tonkin

, et al. Association between triglycerides and risk of dementia in community-dwelling older adults: a prospective cohort study. Neurology 2023; 101: e2288–e2299.

29.

Shofer

Kukull

, et al. Serum cholesterol and risk of Alzheimer disease: a community-based cohort study. Neurology 2005; 65: 1045–1050.

30.

Nayak

Kuriyakose

Polisetty

, et al. Diagnostic and prognostic value of triglyceride glucose index: a comprehensive evaluation of meta-analysis. Cardiovasc Diabetol 2024; 23: 310.

31.

Liu

Luo

Liu

, et al. Triglyceride glucose-related indexes and lipid accumulation products-reliable markers of insulin resistance in the Chinese population. Front Nutr 2024; 11: 1373039.

32.

Zhang

Chen

Sheng

. Association of triglyceride glucose-body mass index with Alzheimer’s disease pathology, cognition and brain structure in non-demented people. Sci Rep 2024; 14: 16097.

33.

Kishore

. Reverse epidemiology of obesity paradox: fact or fiction? Physiol Rep 2024; 12: e70107.

34.

Cui

Mackey

Shaaban

, et al. Associations of body composition with incident dementia in older adults: cardiovascular Health Study-Cognition Study. Alzheimers Dement 2020; 16: 1402–1411.

35.

Fountain

Bopp

Bene

, et al. Metabolic dysfunction and the development of physical frailty: an aging war of attrition. Geroscience 2024; 46: 3711–3721.

36.

Frater

Lie

Bartlett

, et al. Insulin-like growth factor 1 (IGF-1) as a marker of cognitive decline in normal ageing: a review. Ageing Res Rev 2018; 42: 14–27.

37.

Lieb

Beiser

Vasan

, et al. Association of plasma leptin levels with incident Alzheimer disease and MRI measures of brain aging. JAMA 2009; 302: 2565–2572.

38.

Hsu

Voss

Best

, et al. Elevated body mass index and maintenance of cognitive function in late life: exploring underlying neural mechanisms. Front Aging Neurosci 2015; 7: 155.

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.

0.00 MB

0.02 MB