Alzheimer’s disease (AD), a progressive neurodegenerative disorder, manifests clinically as gradual memory loss and cognitive decline (Khan et al., 2020). The widely recognized neuropathological hallmarks include extracellular β-amyloid (Aβ) deposition, intraneuronal neurofibrillary tangles composed of hyperphosphorylated tau protein, neuronal degeneration, and neuroinflammation (Wirth et al., 2013). With accelerating global population aging, AD has emerged as a critical public health challenge. Recent epidemiological data indicate a crude prevalence rate of 9.1% for dementia among Chinese individuals aged ≥ 65 years (Gan et al., 2024). Global dementia cases are projected to increase from 50 million in 2010 to 113 million by 2050 (Scheltens et al., 2021). According to the China Alzheimer’s Disease Report 2024, there were approximately 16.99 million prevalent cases of AD and other dementias in China in 2021, with AD-related deaths reaching 490,000—accounting for 25.2% of global dementia mortality (Wang et al., 2024). These statistics underscore the urgency of identifying modifiable risk factors and implementing early preventive strategies. While multiple etiological factors contribute to AD pathogenesis, mounting evidence suggests metabolic dysregulation may represent a crucial pathophysiological mechanism underlying AD development (de la Monte and Wands, 2008).

Apolipoprotein E (ApoE) gene is the most impactful genetic risk factor for AD. Despite the growing number of identified genetic risk factors, ApoE remains the strongest and most common, influencing over half of AD cases. Specifically, the ApoE ε4 is a major genetic risk factor for AD, increasing the risk in monozygotic twins by up to 15 times in a gene-dosage-dependent manner (Strittmatter et al., 1993). In contrast, ApoE ε2 nearly halves the risk of AD and is considered a protective factor against cognitive impairment (Corder et al., 1994). The ApoE ε4 has been definitively established as a primary genetic risk factor for AD (Corder et al., 1993). ApoE is primarily produced by peripheral and astrocytes in the central nervous system (CNS), playing a key role in cholesterol transport and regulating lipid transport and injury repair in the brain (Liu et al., 2013). Previous studies have shown that ApoE can regulate lipoprotein metabolism and also participate in the pathogenesis of AD by modulating brain insulin signaling, neuroinflammation, and synaptic plasticity (Mahley et al., 2006). Notably, peripheral metabolic abnormalities in ApoE ε4 carriers, such as hypertriglyceridemia and insulin resistance, may exacerbate the development of AD by increasing blood-brain barrier permeability, affecting intracerebral energy metabolism, and creating a vicious cycle of “metabolic-neuroinflammatory” (Cunnane et al., 2016).

IR and body BMI have long been recognized as core metrics for assessing metabolic health. IR primarily manifests as reduced insulin responsiveness in target tissues (Park et al., 2015), with common forms predominantly driven by energy intake-expenditure imbalance rather than genetic predisposition (Michailidis et al., 2022), predisposing individuals to obesity and type 2 diabetes (Sêdzikowska and Szablewski, 2021). Compared to the invasive hyperinsulinemic-euglycemic clamp method, the TyG index serves as a non-invasive and cost-effective alternative for IR screening. Notably, elevated TyG levels demonstrate strong correlations with accelerated cognitive decline in AD cohorts, suggesting its potential role in predicting disease progression (Tian et al., 2023). The relationship between BMI and neurodegenerative disorders exhibits a paradoxical pattern: midlife obesity (BMI ≥ 30 kg/m²) is significantly associated with increased dementia risk, while late-life weight loss may reflect disease-induced catabolic states, serving as a prodromal biomarker of neuropathological burden (Kivimäki et al., 2018).

While current research has advanced our understanding, critical gaps persist regarding the interaction dynamics between ApoE ε4 and metabolic disturbances (TyG index/BMI) in AD populations, particularly their specific cognitive consequences. This study systematically investigates the interplay of ApoE ε4 status with metabolic dysregulation in AD patients, with dual objectives to determine whether ApoE ε4-driven metabolic perturbations are causally linked to disease progression, and to establish a precision intervention framework for early metabolic-targeted strategies in genetically susceptible AD populations.

The research team has conducted a large-scale cross-sectional study on AD patients in northern China, exploring the link between the ApoE ε4 and metabolic factors (BMI and TyG index) and their joint impact on cognitive function. As is known to all, AD patients are often accompanied by insulin resistance, and this study also obtained similar results (Faqih et al., 2021). Results show that ApoE ε4 carriers in the AD population have significantly lower BMI and higher TyG index than non-carriers, a pattern consistent across gender. Age-stratified analyzes reveal the most significant BMI and TyG associations with ApoE ε4 in patients aged 66–74, while only BMI shows a strong link in those aged 75 and above. In terms of cognitive function levels, as measured by MMSE scores, ApoE ε4 carriers with mild to moderate cognitive impairment (MMSE > 20 and MMSE 10–20) exhibit a significantly higher TyG index than non-carriers, with no such difference observed in severe cases (MMSE < 10). Furthermore, mediation analysis suggests that ApoE ε4 may exacerbate cognitive decline indirectly by lowering BMI and raising the TyG index, accounting for 26.6% of the total effect.

As the core genetic risk factor for AD, the ApoE ε4 not only directly contributes to the central pathological mechanisms of AD, such as amyloid-beta (Aβ) deposition and neuroinflammation, but is also strongly associated with peripheral metabolic disorders. Our findings demonstrate significantly lower BMI (P < 0.001) and elevated TyG indices (P < 0.001) in ApoE ε4 carriers vs. non-ApoE ε4 carriers, suggesting ApoE ε4-mediated metabolic reprogramming involving energy homeostasis and IR contributes to AD progression.

The observed BMI reduction in ApoE ε4 carriers may reflect lipid metabolic dysregulation and cerebral bioenergetic insufficiency (Li and Liu, 2014). By disrupting lipoprotein metabolism, ApoE ε4 contributes to impaired peripheral lipid transport and disrupted lipid homeostasis in the brain. This mechanism may result in reduced fat storage and enhanced energy expenditure, ultimately leading to a lower BMI (Kypreos et al., 2009). In addition, research has demonstrated that defects in brain glucose uptake occur in the early stages of AD. Impaired brain glucose metabolism in ApoE ε4 carriers may exacerbate weight loss by promoting compensatory peripheral fat mobilization (Cunnane et al., 2016). Complementary evidence from Zhao et al., reveals ApoE ε4-mediated impairment of insulin signaling responsiveness—HDL cholesterol-induced peripheral IR synergizes with ApoE ε4 to disrupt cerebral insulin transduction (Zhao et al., 2017). Previous research has demonstrated that ApoE ε4 not only facilitates lipid accumulation in hepatocytes, contributing to fatty liver disease, but also suppresses hepatic insulin receptor signaling. These effects result in increased hepatic glucose production and elevated systemic glucose levels (Almeida et al., 2024; El-Lebedy et al., 2016; Samuel et al., 2010). In addition, ApoE ε4 may also activate the sympathetic nervous system or the hypothalamic-pituitary-adrenal (HPA) axis by impairing hypothalamic function or compromising blood-brain barrier integrity. Chronic stress elevates cortisol and epinephrine levels, stimulates lipolysis and free fatty acid release, and exacerbates peripheral insulin resistance (Martens et al., 2022; Montagne et al., 2020).

Notably, our study found that the significant association between BMI and ApoE ε4 persisted in patients aged ≥ 75 years, while the association between TyG and ApoE ε4 was lost, potentially due to the confounding effects of late-life metabolic changes resulting from disease progression (e.g., increased catabolism) or comorbidities (e.g., malnutrition). This finding aligns with the hypothesis that weight loss in late life may serve as a predictor of neurodegeneration, as observed in the context of the “obesity paradox” (Joo et al., 2018).

In addition, the TyG difference was pronounced in the mild-to-moderate stages of cognitive impairment (MMSE > 20 and MMSE 10–20) but became insignificant in the severe stage (MMSE < 10). This may be attributed to the fact that AD patients often experience systemic metabolic failure, such as malnutrition, muscle atrophy, and multiple organ dysfunction, in the advanced stages of the disease. Irrespective of ApoE genotype, metabolic indicators like TyG tend to exhibit persistent abnormalities with reduced fluctuations, thereby diminishing the distinctions between genotypes (Arnold et al., 2018). Furthermore, as the disease advances to the severe stage, extensive neuronal loss, amyloid-beta (Aβ) deposition, and tau pathology become predominant factors. Consequently, the contribution of metabolic abnormalities to cognitive decline may be overshadowed, leading to genotype-related TyG differences becoming statistically non-significant (de la Monte and Wands, 2008).

Metabolic disorders can exacerbate cognitive impairment. Cunnane et al. (2016) found that in AD patients, insulin resistance (IR)-induced hyperinsulinemia and hyperglycemia, combined with the effects of ApoE ε4, accelerates the formation of neurotic plaques (NPs). IR may also directly influence AD pathology through its interaction with amyloid-beta (Aβ) peptides. One study demonstrated a positive correlation between peripheral IR and brain Aβ deposition in the frontal and temporal lobes of AD patients. Midlife HOMA-IR was also shown to predict Aβ aggregation, with this prediction confirmed 15 years later via amyloid PET imaging (Ekblad et al., 2018; Willette et al., 2015). Laws et al. (2017) confirmed that in cognitively healthy adults, increased peripheral insulin resistance is linked to worse cognitive scores and higher cerebrospinal fluid (CSF) phosphorylated/total tau. Low BMI is significantly associated with AD-related cognitive impairment. Epidemiological studies (Grau-Rivera et al., 2021) indicate midlife or late-life low BMI may increase AD risk due to reduced nutritional status or insufficient metabolic reserves. Notably, weight loss can be an early AD biomarker, with BMI declines occurring before diagnosis. This may result from hippocampal atrophy, reduced olfactory function leading to decreased appetite, or abnormal central metabolic regulation (Grau-Rivera et al., 2021). Previous studies indicate that low BMI is linked to reduced leptin levels. This reduction impairs leptin’s ability to suppress Aβ deposition and boost hippocampal synaptic plasticity (Lieb et al., 2009). Moreover, undernutrition increases neuronal oxidative damage, and sarcopenia-induced brain-derived neurotrophic factor (BDNF) reduction further harms cognitive function (Burns et al., 2010). However, some research suggests a U-shaped relationship between BMI and AD risk, with both low and high BMI being risk factors. Also, low BMI might indicate preclinical AD pathophysiology rather than being a direct cause (Deng et al., 2022). Longitudinal studies with biomarkers like Aβ-PET and serum leptin are needed to clarify causality and guide nutritional intervention strategies.

Our study quantified the metabolic pathway’s role in ApoE genotypes’ impact on cognitive function in AD, accounting for 26.6% of the total effect. Consistent with prior studies, Seth Stoykovich et al. found that in ApoE ε4 carriers, insulin resistance (elevated TyG index) correlates significantly with memory and executive function deficits. Notably, TyG’s mediating effect (16.9%) surpasses BMI’s (9.7%), highlighting insulin resistance as a more central pathological mechanism (Cunnane et al., 2016). This aligns with the theory of AD as “type 3 diabetes” (Peng et al., 2024). Meng et al. (2025) showed that insulin resistance drives neuronal metabolic reprogramming in AD progression and emphasized that enhancing insulin sensitivity could be a viable therapeutic strategy. Thus, metabolic regulation plays a crucial role in translating ApoE genetic risk into cognitive impairment, offering new quantitative insights into AD’s gene-metabolism-cognition relationship.

Our study has several limitations: (1) Its cross-sectional design precludes causal inferences. (2) It lacks CSF biomarker measurements. (3) It doesn’t account for metabolic mediators like inflammatory cytokines [interleukin-6 (IL-6), C-reactive protein (CRP)] and adipokines (leptin, adiponectin), leaving residual mediating effects unexplained. (4) It doesn’t control for confounders such as diet and physical activity. Future research should focus on longitudinal metabolic tracking and brain PET metabolic imaging. This would help explore personalized metabolic interventions guided by ApoE genotypes.

This study demonstrates that ApoE ε4 carriers within the AD population exhibit significant peripheral metabolic homeostasis imbalance, characterized by decreased BMI and elevated TyG index, suggesting concurrent lipid metabolism dysregulation and insulin resistance. Furthermore, the ApoE ε4 may exacerbate AD-related cognitive impairment through metabolic pathways characterized by low BMI and elevated TyG index, with particularly pronounced effects observed in individuals aged 66–74 years and those with mild-to-moderate cognitive impairment. These findings provide crucial scientific evidence for precision prevention strategies in AD: early metabolic interventions targeting ApoE ε4 carriers (including dynamic monitoring of TyG index, nutritional status optimization, and insulin sensitivity enhancement) may represent effective approaches for delaying disease progression. Future research should integrate multi-omics data with metabolism-targeted clinical trials to elucidate the molecular mechanisms underlying metabolic regulation and validate their potential value in primary AD prevention.