Cognitive impairment, particularly dementia, poses a growing global public health concern, with substantial implications for individuals, families, and healthcare systems (Lee et al., 2023; Reuben et al., 2024). In 2019, an estimated 57.4 million individuals worldwide were affected by dementia, and this number is projected to rise to 152.8 million by 2050 (GBD Dementia Forecasting Collaborators, 2022). Dementia not only substantially diminishes patients’ quality of life but also places a considerable economic burden on families and healthcare systems (Kalaria et al., 2024; Sommerlad et al., 2023). Given the current challenges in dementia treatment, identifying and addressing modifiable risk factors is crucial for developing effective prevention strategies (Budd Haeberlein et al., 2022; van Dyck et al., 2023).

Recent studies increasingly suggest a strong association between dyslipidemia and cognitive decline (Pan et al., 2024). As modifiable biological markers, lipid profiles offer promising targets for interventions aimed at preventing cognitive dysfunction (Petek et al., 2023). The non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio (NHHR) has emerged as a comprehensive lipid indicator that reflects both atherogenic and protective lipoproteins (Zhen et al., 2023). NHHR has been recognized for its utility in cardiovascular risk assessment (Liu et al., 2024). Compared with traditional lipid parameters, NHHR may provide a more holistic reflection of lipid-related vascular risk, yet evidence supporting its role in cognitive health is limited. Therefore, this study aimed to investigate the association between NHHR and cognitive function using data from the Liuyang Jili Hospital Physical Examination Center. We hypothesized that NHHR would exhibit a nonlinear relationship with cognitive function, providing new insights into the potential utility of lipid ratios in strategies for cognitive health maintenance and dementia prevention.

In this cross-sectional study of a community-based population in China, we identified a U-shaped association between the NHHR and cognitive impairment. Breakpoint regression analysis suggested an inflection point at 2.772, above which NHHR was significantly associated with increased risk of cognitive impairment. Machine learning analysis confirmed this U-shaped relationship and indicated that NHHR contributed more strongly to the model than individual lipid markers. Collectively, these findings suggest that NHHR may better reflect lipid-related risk for cognitive impairment than single lipid measures and support its potential role as an alternative index in studies of cognitive health.

NHHR reflects the ratio of NHDL cholesterol to HDL cholesterol and is inversely related to HDL levels while positively correlated with NHDL levels. Both lipoproteins have been shown to be strongly associated with cognitive function, with HDL cholesterol demonstrating a U-shaped relationship with cognitive outcomes. For example, a recent study reported that individuals with HDL cholesterol levels exceeding 80 mg/dL had a 27% higher risk of dementia compared with those with HDL levels of 40–60 mg/dL (Hussain et al., 2024). Similarly, another study found that HDL levels above 65 mg/dL were associated with a 15% increased risk of dementia compared with the median level (53.7 mg/dL), while HDL levels between 11 and 41 mg/dL were linked to a 7% increased risk (Ferguson et al., 2023). Collectively, these findings suggest that both excessively high and low HDL cholesterol levels may increase dementia risk, aligning with the U-shaped association observed in the present study. In contrast, NHDL cholesterol primarily comprises LDL and remnant cholesterol (RC), both of which, when elevated, have been associated with cognitive decline. A large population-based study from the United Kingdom found that a 1-standard deviation increase in LDL cholesterol (39 mg/dL) was associated with a 5% increased risk of dementia in the general population. Additionally, higher concentrations of RC were linked to greater risks of all-cause dementia, Alzheimer’s disease, and vascular dementia. Compared with participants in the lowest quartile of RC, those in the highest quartile had hazard ratios of 1.11 (95% CI, 1.09–1.13) for all-cause dementia, 1.11 (95% CI, 1.08–1.13) for Alzheimer’s disease, and 1.15 (95% CI, 1.09–1.21) for vascular dementia. Notably, a recent cross-sectional study also reported a U-shaped relationship between NHDL cholesterol and cognitive function (Li et al., 2024). Taken together, these findings suggest that the right side of the U-shaped curve observed in the present study may reflect the adverse effects of elevated NHDL cholesterol and reduced HDL cholesterol, while the left side of the curve may be attributable to the detrimental impact of excessively high HDL cholesterol levels.

Our study identified a U-shaped association between NHHR and cognitive function, which may be explained by several potential biological mechanisms. At higher NHHR levels, reflecting an increased burden of NHDL lipoproteins, there is a greater propensity for the development of atherosclerosis, cerebral microvascular injury, chronic cerebral hypoperfusion, blood–brain barrier disruption, intracerebral inflammation, and oxidative stress (Bolanle et al., 2025; Chen et al., 2024; Chung et al., 2023; Dias et al., 2015; Huang et al., 2019; Lara-Guzman et al., 2018). These pathophysiological processes can ultimately result in neuronal damage and cognitive decline. In parallel, a decrease in HDL levels may weaken its protective functions, including anti-inflammatory, antioxidant, and β-amyloid clearance activities, thereby further exacerbating the risk of cognitive impairment (Grao-Cruces et al., 2022; Tziomalos et al., 2019; Zhang et al., 2022). Conversely, low NHHR is often accompanied by elevated HDL levels; however, excessively high HDL does not necessarily confer additional neuroprotection. Emerging evidence suggests that dysfunctional HDL particles may be enriched with lipids, exhibit reduced antioxidant and anti-inflammatory capacity, and display impaired reverse cholesterol transport. Such dysfunctional HDL may result in insufficient cholesterol delivery to the brain, negatively impacting neuronal synaptic plasticity and neurotransmitter synthesis (Kjeldsen et al., 2021, 2022; Suleiman et al., 2020). Additionally, reduced NHDL cholesterol levels may reflect suboptimal metabolic or nutritional conditions, which could impair brain homeostasis through insufficient provision of substrates required for neuronal maintenance and repair processes (An et al., 2019; Gong et al., 2022; Li et al., 2020). Moreover, extreme deviations in NHHR may disrupt reverse cholesterol transport. At very low NHHR levels, although HDL concentrations are high, its functional capacity to accept and transport cholesterol may be diminished. Conversely, at very high NHHR levels, the excessive accumulation of NHDL lipoproteins may overwhelm the clearance capacity of HDL, leading to cholesterol retention within the arterial wall, thereby promoting the progression of atherosclerotic plaque (Kjeldsen et al., 2021; Bhale et al., 2024; Durrington et al., 2025; Tosheska Trajkovska and Topuzovska, 2017). In summary, elevated NHHR likely contributes to cognitive impairment through mechanisms related to vascular injury and inflammation, whereas low NHHR is more likely linked to HDL dysfunction and cerebral cholesterol deficiency. Both extremes may disrupt cognitive health, underscoring the potential importance of maintaining NHHR within an appropriate range for preserving cognitive function.

The present study found that cognitive function appeared to be optimal when NHHR was maintained within the range of approximately 2.8, a finding that is generally consistent with existing lipid management guidelines (Patel et al., 2025). When NHDL cholesterol is <130 mg/dL and HDL cholesterol is >40 mg/dL, NHHR naturally falls within this range. Maintaining NHHR within the range of approximately 2.8 provides clinicians with a simple and intuitive target for lipid management. Compared to focusing solely on LDL or HDL levels, NHHR offers a more comprehensive assessment of lipid metabolism, facilitating the development of more individualized intervention strategies.

This study has several limitations. First, as a cross-sectional study, it can only establish associations rather than causality. Second, although multiple confounding variables were adjusted for, our analysis did not account for potential residual confounders such as APOE genotype and psychological factors including depression and anxiety, which are known to influence cognitive function. Third, lipid levels were assessed based on a single measurement, whereas lipid profiles may fluctuate over time; dynamic lipid monitoring may provide a more accurate reflection of long-term lipid status. Fourth, the study population was derived from a hospital-based sample, which may introduce selection bias and limit the generalizability of findings to the broader community population. Survival bias may also be present, particularly in older adults, as individuals with severe comorbidities or cognitive impairment may be underrepresented due to early mortality or inability to participate in examinations.

In this community-based cross-sectional study, NHHR demonstrated a significant U-shaped association with cognitive impairment, with risk increasing when NHHR exceeded an inflection point of approximately 2.8. Compared with single lipid measures, NHHR appeared to provide superior discrimination of cognitive impairment. These findings suggest that NHHR may serve as a useful lipid-related index in studies of cognitive health, although further longitudinal research is needed to clarify its predictive value and clinical implications.