Vascular risk factor variability and cognitive impairment in older adults a community-based retrospective study Authors: Xianyun Wang, Shiyu Tian, Haiqing Xiao, Chen Chen, Feng Ning, Yongfeng Lin Published: 2025 ABSTRACT AbstractThe association between vascular risk factors and cognitive performance has been a topic of considerable interest. Nevertheless, the connection between vascular risk factors and cognitive impairment in the elderly population continues to be ambiguous. From January to April 2022, participants aged 65 years and older were recruited in the community. A retrospective review of data was conducted, those who underwent at least three discrete health assessments between 2018 and 2021 were included in the study. Cognitive function was evaluated in 2022 by Mini-mental state examination (MMSE) in the Chinese version. Logistic regression was employed to explore the association of various vascular risk factors (body mass index [BMI], exercise frequency, smoking status, drinking status, systolic blood pressure [SBP], diastolic blood pressure [DBP], total cholesterol [TC], triglycerides [TG], low-density lipoprotein cholesterol [LDL-C], high-density lipoprotein cholesterol [HDL-C], fasting plasma glucose [FPG]) and their variability (intra-individual variability across at least four measurement occasions) with cognitive impairment. Odd ratios (OR) and 95% confidence intervals (CI) were reported. This study included 1,487 participants. Findings indicated that physical activity and alcohol drinking were inversely related to cognitive impairment risk. Moreover, average DBP, TC, LDL-C, and FPG at follow-up were positively correlated with cognitive impairment risk, while HDL-C was linked to a reduced risk. Variability assessments reflected high variability in DBP, TG, TG, LDL-C, HDL-C, and FPG associated with cognitive impairment development. In this research, we conducted a thorough examination of the relationship between four primary vascular risk factors (11 indicators) and the likelihood of cognitive impairment. Our results indicate that various vascular risk factors are associated with the onset of cognitive impairment in the elderly population. These findings may contribute to the formulation of future guidelines aimed at mitigating the risk of cognitive impairment.Keywords: Alzheimer’s disease, Cognitive impairment, Blood pressure, Lipids, Fasting plasma glucose, Risk factorSubject terms: Medical research, Risk factors, Ageing Introduction Alzheimer’s disease (AD) ranks among the most common neurodegenerative disorders, impacting millions globally, with both prevalence and incidence rates escalating with advancing age1,2. Cognitive impairment not only affects the quality of life of individuals but also imposes substantial economic and social costs3. An expanding body of research indicates that vascular risk factors are crucial in the onset of cognitive impairment, both independently and in conjunction with other pathological mechanisms4. These factors, which encompass unhealthy lifestyle choices, hypertension, dyslipidemia, and dysglycemia, present opportunities for targeted preventive strategies aimed at mitigating the incidence of AD5,6. Longitudinal monitoring of these risk factors can yield insights beyond those obtained from single measurements. The growing attention to the variability within individual vascular risk factors has revealed that fluctuations in various physiological parameters, such as blood pressure (BP) variability, lipid variability, and fasting plasma glucose (FPG) variability, are associated with the risk of cognitive impairment7–9. Lee et al. found there was a significant association between variability in metabolic parameters (specifically total cholesterol [TC], body mass index [BMI], systolic blood pressure [SBP], and glucose) and the risk of incident all-cause dementia, AD, and vascular dementia. Additionally, there was a composite effect between the number of high varying (highest quartile) parameters and the risk of incident all-cause dementia, AD, and vascular dementia10. The evidence indicates that efforts should focus not only on managing the absolute values but also on minimizing fluctuations to enhance health outcomes. This presents a novel objective for the prevention and management of cognitive impairment. This study aims to identify the relationship between four key vascular risk factors (lifestyle, BP, lipid levels, and FPG) and cognitive impairment in community-dwelling older adults. After adjusting for confounding variables, it further investigates the association between variability in blood pressure, lipid, and blood glucose with cognitive function. Result Demographic and clinical characteristics of study population As shown in Table 1, a total of 1487 subjects were included in this study, including 782 (52.59%) with normal cognitive function and 705 (47.41%) with cognitive impairment. Older age, female gender, low education, higher BMI, low frequency of exercise, and non-drinking alcohol were more common in the cognitive impairment group. Furthermore, no statistically significant differences were observed in the baseline levels of BP, lipid and FPG levels between the two groups. Nevertheless, at follow-up, the average SBP, average triglyceride levels, and average FPG concentrations were markedly elevated in the cognitive impairment group compared to the cognitively normal participants, while the average HDL-C levels were significantly diminished relative to the cognitively normal participants. Table 1Descriptive characteristics of the study population.CharacteristicsCognitive impairment p None (n = 782)Impairment (n = 705)Age (yr)72.75 ± 5.6974.01 ± 6.36< 0.001***Female, n (%)391 (50.0)466 (66.1)< 0.001***Education, n (%)< 0.001***Illiterate78 (10.0)236 (33.5)Primary School391 (50.0)324 (46.0)Secondary School282 (36.1)136 (19.3)University or Higher31 (4.0)9 (1.3)LifestyleBMI (kg/m2)25.18 ± 3.2825.59 ± 3.610.025*Exercise frequency, n (%)< 0.001***Never432 (55.2)488 (69.2)< 1 times/week34 (4.3)19 (2.7)≥ 1 times/week55 (7.0)15 (2.1)Everyday261 (33.4)183 (26.0)Smoking status, n (%)0.066Non-smoker693 (88.6)645 (91.5)Smoker89 (11.4)60 (8.5)Drinking status, n (%)< 0.001***Non-drinker669 (85.5)651 (92.3)Drinker113 (14.5)54 (7.7)Blood PressureBaseline†SBP (mm Hg)138.86 ± 17.47138.75 ± 20.100.909DBP (mm Hg)79.02 ± 9.9479.63 ± 11.340.271Follow-up†Average SBP (mm Hg)140.72 ± 15.78142.98 ± 16.490.007**Average DBP (mm Hg)78.63 ± 8.9179.45 ± 9.420.085Blood Lipid FractionsBaseline†Total cholesterol (mmol/L)5.60 ± 1.215.69 ± 1.170.13Triglycerides (mmol/L)1.50 ± 0.971.58 ± 1.090.186LDL-C (mmol/L)3.15 ± 0.903.15 ± 0.840.941HDL-C (mmol/L)1.49 ± 0.391.48 ± 0.350.66Follow-up†Average Total cholesterol (mmol/L)5.56 ± 1.035.60 ± 0.920.44Average Triglycerides (mmol/L)1.44 ± 0.751.57 ± 0.840.002**Average LDL-C (mmol/L)3.13 ± 0.783.21 ± 0.730.048*Average HDL-C (mmol/L)1.56 ± 0.341.52 ± 0.300.017*Blood GlucoseBaseline FPG (mmol/L)†6.25 ± 1.996.36 ± 2.120.323Follow-up average FPG (mmol/L)†6.21 ± 1.466.40 ± 1.660.021*BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; FPG, fasting plasma glucose.† Data are presented as mean ± standard deviation. Baseline data were collected in 2018. Follow-up data were collected over four annual visits (2019, 2020, 2021, 2022). The “Average” values for the follow-up period represent the mean of all available measurements across these four visits for each participant.*p < 0.05, **p < 0.01, ***p < 0.001. Descriptive characteristics of the study population. BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; FPG, fasting plasma glucose. † Data are presented as mean ± standard deviation. Baseline data were collected in 2018. Follow-up data were collected over four annual visits (2019, 2020, 2021, 2022). The “Average” values for the follow-up period represent the mean of all available measurements across these four visits for each participant. *p < 0.05, **p < 0.01, ***p < 0.001. Variability of vascular risk factors between different population A notable disparity was observed in the variability of vascular risk factors between the cognitive impairment group and the cognitively normal participants (Table 2). Specifically, individuals with the cognitive impairment exhibited significantly greater four-year fluctuations in BP, lipid, and FPG compared to those in the cognitively normal participants. Table 2Variability of vascular risk factors between different population.CharacteristicsCognitive impairment p None (n = 782)Impairment (n = 705)Blood PressureSBP (mm Hg)SD12.73 ± 6.9513.93 ± 7.630.004**CV8.95 ± 4.509.71 ± 4.910.006**VIM12.71 ± 6.2113.68 ± 6.760.012*DBP (mm Hg)SD6.68 ± 3.557.57 ± 4.690.001**CV8.53 ± 4.519.52 ± 5.490.003**VIM7.32 ± 4.977.93 ± 5.380.043*Blood Lipid FractionsTotal cholesterol (mmol/L)SD0.53 ± 0.450.58 ± 0.370.001**CV9.58 ± 7.0210.76 ± 7.09< 0.001***VIM0.62 ± 0.650.73 ± 0.74< 0.001***Triglycerides (mmol/L)SD0.40 ± 0.440.52 ± 0.65< 0.001***CV26.14 ± 14.5629.70 ± 18.350.001**VIM0.46 ± 0.280.48 ± 0.280.101LDL-C (mmol/L)SD0.46 ± 0.280.51 ± 0.29< 0.001***CV15.03 ± 9.4716.90 ± 10.11< 0.001***VIM0.46 ± 0.280.51 ± 0.29< 0.001***HDL-C (mmol/L)SD0.19 ± 0.180.22 ± 0.17< 0.001***CV11.96 ± 9.1713.72 ± 8.76< 0.001***VIM0.20 ± 0.140.23 ± 0.14< 0.001***Blood GlucoseFPG (mmol/L)SD0.73 ± 0.810.88 ± 1.08< 0.001***CV10.53 ± 8.4412.55 ± 11.060.001**VIM0.65 ± 0.400.76 ± 0.52< 0.001***BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; FPG, fasting plasma glucose.*p < 0.05, **p < 0.01, ***p < 0.001. Variability of vascular risk factors between different population. BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; FPG, fasting plasma glucose. *p < 0.05, **p < 0.01, ***p < 0.001. Correlation of vascular risk factors with cognitive impairment Binary logistic regression was used to explore the correlation of vascular risk factors with cognitive function. The results showed that individuals who engaged in regular physical activity and alcohol drinking consumption were associated with a lower frequency of cognitive impairment compared to those who did not. Even after adjusting for factors related to age, gender, and education, physical activity continued to serve as a protective factor for cognitive health (Table 3). Table 3Association between lifestyle with cognitive impairment.Model1Model2OR95% CI p OR95% CI p BMIEmaciation2.64(0.89–7.85)0.082.11(0.66–6.68)0.206NormalReferenceReferenceOverweight1.17(0.92–1.49)0.191.21(0.94–1.56)0.143Obesity1.31(0.97–1.76)0.0781.30(0.95–1.79)0.104Exercise frequencyNeverReferenceReference< 1 times/week0.51(0.29–0.91)0.023*0.50(0.27–0.93)0.028*≥ 1 times/week0.24(0.13–0.42)< 0.000***0.26(0.14–0.47)< 0.000***Everyday0.65(0.52–0.82)< 0.000***0.72(0.56–0.93)0.011*Smoking statusNon-smokerReferenceReferenceSmoker0.95(0.64–1.40)0.7821.38(0.91–2.10)0.134Drinking statusNon-drinkerReferenceReferenceDrinker0.53(0.36–0.77)0.001**0.72(0.48–1.10)0.126Model 1: Non-adjusted.Model 2: Adjusted for age, sex, educational level.OR: Odds Ratio; 95%CI: 95% confidence interval.*p < 0.05, **p < 0.01, ***p < 0.001. Association between lifestyle with cognitive impairment. Model 1: Non-adjusted. Model 2: Adjusted for age, sex, educational level. OR: Odds Ratio; 95%CI: 95% confidence interval. *p < 0.05, **p < 0.01, ***p < 0.001. After dividing the follow-up average BP, lipid, and FPG into quartiles, we observed in Model 1 that the OR for individuals in the second quartile (Q2) of low-density lipoprotein cholesterol (LDL-C), the third quartile of HDL-C (Q3), and the highest quartile of FPG (Q4), compared to individuals in the lowest quartile (Q1), were 1.62, 0.72, and 1.35. In Model 2, diastolic blood pressure (DBP) (Q3, Q4), TC (Q3), LDL-C (Q2, Q4), and FPG (Q4) were significantly associated with an increased risk of cognitive impairment. In contrast, HDL-C (Q2, Q3, Q4) was significantly associated with reduced risk of cognitive impairment (Table 4). Table 4Correlation of BP, lipid, and FPG with cognitive impairment.Model1Model2OR95% CI p OR95% CI p SBP (mm Hg)Q1 (125 [110–131])ReferenceReferenceQ2 (137 [132–141])0.95(0.70–1.28)0.7140.95(0.69–1.32)0.779Q3 (147 [142–152])1.04(0.76–1.42)0.8030.87(0.62–1.21)0.41Q4 (160 [153–200])1.16(0.84–1.61)0.3660.98(0.69–1.40)0.926DBP (mm Hg)Q1 (70 [55–73])ReferenceReferenceQ2 (77 [74–79])1.15(0.86–1.54)0.3461.35(0.99–1.83)0.056Q3 (82 [80–85])1.09(0.80–1.47)0.6011.43(1.05–1.95)0.025*Q4 (90 [86–115])1.16(0.83–1.62)0.41.48(1.07–2.04)0.018*Total cholesterol (mmol/L)Q1 (4.49 [2.67–4.93])ReferenceReferenceQ2 (5.24 [4.94–5.35])1.14(0.81–1.59)0.4521.22(0.85–1.73)0.276Q3 (5.81 [5.54–6.17])1.43(0.96–2.14)0.0781.60(1.05–2.42)0.028*Q4 (6.69 [6.18–11.60])1.36(0.83–2.24)0.2231.32(0.79–2.19)0.289Triglycerides (mmol/L)Q1 (0.85 [0.34-1.00])ReferenceReferenceQ2 (1.14 [1.01–1.32])1.09(0.80–1.47)0.590.92(0.67–1.27)0.606Q3 (1.52 [1.33–1.76])0.94(0.69–1.29)0.70.77(0.55–1.08)0.136Q4 (2.19 [1.77–6.45])1.29(0.93–1.81)0.131.09(0.76–1.56)0.638LDL-C (mmol/L)Q1 (2.31 [1.06–2.61])ReferenceReferenceQ2 (2.88 [2.62–3.13])1.62(1.19–2.21)0.002*1.56(1.14–2.13)0.005*Q3 (3.37 [3.14–3.67])1.23(0.86–1.77)0.2541.29(0.94–1.76)0.117Q4 (4.09 [3.68–6.21])1.22(0.79–1.89)0.3791.42(1.03–1.95)0.031*HDL-C (mmol/L)Q1 (1.23 [1.00-1.33])ReferenceReferenceQ2 (1.42 [1.34–1.50])0.74(0.55–1.00)0.0530.70(0.51–0.96)0.030*Q3 (1.60 [1.51–1.71])0.72(0.52–0.99)0.046*0.63(0.46–0.86)0.000*Q4 (1.87 [1.72–3.90])0.73(0.51–1.04)0.0810.71(0.51–0.97)0.030*Fasting glucose (mmol/L)Q1 (5.07 [4.21–5.34])ReferenceReferenceQ2 (5.58 [5.35–5.81])1.2(0.89–1.61)0.2321.26(0.92–1.71)0.154Q3 (6.12 [5.82–6.63])1.02(0.76–1.37)0.9181.03(0.76–1.42)0.834Q4 (7.83 [6.64–16.46])1.35(1.00–1.82)0.050*1.49(1.09–2.03)0.013*Model 1: Non-adjusted.Model 2: Adjusted for age, sex, educational level.OR: Odds Ratio; 95%CI: 95% confidence interval.Each quantile is followed by (median [minimum value - maximum value]).*p < 0.05, **p < 0.01, ***p < 0.001. Correlation of BP, lipid, and FPG with cognitive impairment. Model 1: Non-adjusted. Model 2: Adjusted for age, sex, educational level. OR: Odds Ratio; 95%CI: 95% confidence interval. Each quantile is followed by (median [minimum value - maximum value]). *p < 0.05, **p < 0.01, ***p < 0.001. Association between variability of BP, lipid, and FPG with cognitive impairment The variability in BP, lipid levels, and FPG concentrations over a four-year period was subsequently categorized into quartiles. The findings indicated that no significant associations were detected between SBP, DBP, and triglyceride variability with cognitive impairment. However, high variability in TC (Q4), LDL-C (Q4), HDL-C (Q3, Q4), and FPG (Q4) exhibited significant positive correlations with cognitive impairment (Fig. 1a). In conclusion, increased fluctuations in BP, lipid, and FPG levels within the study cohort over a four-year duration correlates with an elevated risk of cognitive impairment. Similar outcomes were noted when standard deviation (SD) and coefficient of variation (CV) were employed to assess variability (Fig. 1b, c). Association between variability of BP, lipid, and FPG with cognitive impairment. (a) VIM. (b) CV. (c) SD. Model 1: Non-adjusted. Model 2: Adjusted for age, sex, educational level. OR: Odds Ratio; 95%CI: 95% confidence interval. Discussion In this population-based study of community-dwelling older adults conducted over a five-year period (2018–2022), we investigated the association between four key vascular risk factors (11 indicators) and the risk of developing cognitive impairment. Results showed that physical activity and alcohol drinking exhibited a negative correlation with the risk of cognitive impairment. In the analysis of averaged follow-up values, a significant dose-response relationship was observed between higher DBP levels and an increased risk of cognitive impairment. Additionally, significant associations with increased risk were identified for specific higher quartiles of TC, LDL-C, and FPG, while higher levels of HDL-C were consistently associated with a reduced risk. We evaluated variability through three distinct methodologies, and the results generally reflected that high variability in DBP, TC, TG, LDL-C, HDL-C, and FPG were correlated with the onset of cognitive impairment. A multitude of studies has demonstrated that lifestyle determinants, including dietary habits, alcohol drinking, smoking, and physical activity have been extensively investigated, with growing evidence supporting their associations with cognitive impairment and dementia11. A cohort study shows that a healthy lifestyle can help prevent vascular damage or inflammation from the cognitive damage attributed to cardiometabolic diseases12. The evidence clearly demonstrates a robust association between lifestyle, a significant modifiable determinant of vascular risk factors, and cognitive performance. More evidence in the literature shows conclusions consistent with the results of our study. In a large meta-analysis, including 58 cohort and case-control studies and 420,114 individuals, higher physical activity levels were associated with lower incidence of all-cause dementia and Alzheimer’s disease13. In addition, the Chinese Longitudinal Healthy Longevity Survey (CLHLS) showed that moderate alcohol consumption was associated with improved cognition status14. However, the proposed cognition protective actions of alcohol have been lately a matter of debate. A recent dose–response meta-analysis including six prospective studies (n = 4244) found that heavy alcohol intake (>14 drinks/week) was significantly associated with higher risk of progression to dementia in people with MCI, while there was a nonlinear relation between alcohol intake and risk of this progression15. In conclusion, further empirical investigation may be necessary to determine the ideal dosage at which alcohol demonstrates its neuroprotective effects on cognition. A cross-sectional analysis encompassing 46,011 elderly Chinese individuals revealed that modifiable risk factors for dementia include hypertension, hyperlipidaemia, diabetes, among others16. In contrast to the observed facilitative impact of elevated SBP on cognitive impairment in the majority of studies17,18, our research identified a more pronounced detrimental influence of high DBP on cognitive function. Tsivgoulis G found early on that an increment of 10 mmHg in diastolic BP, but not systolic BP, was associated with an increased risk (7%) of cognitive impairment19. A study of community-dwelling older adults in the USA found that across the cognitive spectrum from normal aging to MCI, elevated DBP is associated with reduced cortical volume and perfusion20. In addition, a growing number of studies have found that high blood pressure variability may lead to cognitive impairment21,22. A recent 12-year longitudinal cohort study showed that higher diastolic blood pressure variability in later life was associated with increased risk of subsequent decline in cognition and greater white matter lesion volume23. The aforementioned evidence compellingly indicates that the potential mechanisms through which DBP influences cognitive function have yet to be investigated. The correlation between lipid profiles and cognitive function remains controversial because of conflicting results from different studies. A comprehensive retrospective cohort analysis involving 1.8 million individuals tracked over a 20-year period revealed modest positive correlations between TC, LDL-C, and the risk of developing dementia. This association appeared to differ based on the age at which measurements are taken (mid-life [< 65 years] versus later life [≥ 65 years]) and the duration of follow-up24. Our findings validate this view and are consistent with several other studies and WHO guidelines25–28. It is noteworthy that, in contrast to previous reports, two studies conducted within Chinese populations indicated that elevated serum TG may serve as a protective factor against cognitive impairment in middle-aged male participants29,30. Furthermore, the direction of the association between HDL-C and risk of cognitive impairment has been more controversial in several studies. Many epidemiological studies support the hypothesis that HDL-C protects against cerebrovascular dysfunction in Alzheimer’s disease31,32. A recent meta-analysis reported no significant association between HDL-C and AD risk33. Conversely, a Korean study identified a positive correlation between HDL-C and the risk of AD27. This differential association may be attributed to the deterioration of HDL function due to aging. Holzer M found that HDL exhibited functional deficiencies, including defective antioxidant properties and lower paraoxonase 1 activity, in older individuals34. Rather than HDL quantity, deterioration in HDL functional quality may elucidate the varying effects of HDL on cognitive impairment. In contrast to the aforementioned debates, a significant consensus among researchers indicates that high variability in blood lipids constitutes a risk factor for cognitive performance8,35–37. In our study, a positive correlation between high HDL variability and cognitive impairment was found for all three variability indices. A recent study from the China Health and Retirement Longitudinal Study (CHARLS) found that the highest quartile of VIM for HDL demonstrated a significant correlation an increased risk of cognitive impairment compared to the lowest quartile in all models36 In addition, HDL-C variability was increasingly proposed as a predictor for many adverse outcomes38–41. The mechanisms linking HDL-C variability to cognitive impairment remain unclear. Research indicates that HDL performs at least four distinct functions (reducing Aβ accumulation, mitigating vascular inflammation and neuroinflammation, enhancing nitric oxide production in brain endothelial cells, and delaying Aβ fibrillization) may help prevent Alzheimer’s disease31. Sanjana F even found that HDL-C plays an important independent role in maintaining neuronal composition and integrity42. The negative effects of high FPG on cognitive function have been investigated in extensively studied43–45. While the potential role of FPG variability on cognitive function remains to be explored. A Taiwanese investigation revealed that high variability in fasting plasma glucose (HR = 1.27; 95% CI 1.06–1.52) and HbA1c levels (HR = 1.32; 95% CI 1.11–1.58) are associated with a heightened risk of developing Alzheimer’s disease among patients with type 2 diabetes46. A systematic review and meta-analysis indicate that individuals with glycaemic variability have a 2.65 times higher risk of developing AD compared to those with normal glucose levels. However, due to the limited number of available studies and the confidence interval included the null value of one, the findings are not statistically significant47. Research on glycaemic variability has predominantly targeted diabetic patients. This study on community-dwelling elderly individuals has only partially addresses a gap in understanding glycaemic variability among non-diabetic patients. Further investigation is required to explore the relationship between FPG and cognition. The consistent association observed between higher variability in BP, lipids, and FPG with cognitive impairment necessitates a discussion on potential underlying mechanisms. We posit that these associations are not merely epiphenomena but may reflect a state of underlying physiological dysregulation that directly damages cerebral health. First, excessive fluctuations, particularly in BP, can induce endothelial dysfunction and impair cerebral autoregulation, leading to repetitive hypoperfusion and shear stress on the cerebral microvasculature48,49. Second, visit-to-visit variability itself is an independent predictor of accelerated atherosclerosis, which can manifest as silent cerebral infarcts and white matter hyperintensities—established structural correlates of cognitive impairment37,50. Finally, variability in metabolic parameters like glucose is closely linked to systemic inflammation50. This chronic, low-grade inflammatory state can promote neuroinflammation and exacerbate the pathological processes underlying cognitive impairment. Therefore, vascular variability may serve as a composite marker of physiological instability, whose detrimental effects on the brain are mediated through interconnected pathways of vascular injury and inflammation. Within our community cohort, elevated long-term within-individual variability likely represents a state of impaired physiological homeostasis. This instability may be driven by undiagnosed pathology, ineffective management of known conditions, or a natural decline in regulatory capacity with age. Critically, our findings suggest that irrespective of its origin, such variability is a measurable and independent risk marker for cognitive health. Therefore, beyond average values, tracking the stability of vascular parameters could enhance risk stratification in primary care and community health settings, pinpointing individuals who warrant closer monitoring and preventive actions. The strengths of this research encompass its novel simultaneous investigation of long-term variability in blood pressure, lipids, and fasting plasma glucose within a community-based elderly cohort. The use of three distinct statistical metrics (SD, CV, VIM) enhances the robustness of our findings. Furthermore, we provide insights into the understudied role of glycemic variability in non-diabetic older adults. Notwithstanding these strengths, limitations should be acknowledged. First, the sample size, while valuable for an initial exploration, was relatively modest, which may limit the generalizability of our findings. Future studies will involve proactive community outreach to recruit a larger and more diverse participant pool. Second, the cross-sectional assessment of cognitive function with the MMSE at a single time point precludes any analysis of longitudinal cognitive impairment or causal inference. The establishment of a longitudinal cognitive assessment within our cohort is a critical objective for our subsequent research. Finally, the use of the MMSE, while widely accepted, is a screening instrument rather than a comprehensive diagnostic tool, which may not capture subtler forms of cognitive deficits. In conclusion, this community-based study demonstrated that higher average levels of DBP, TC, LDL-C, and FPG during follow-up were associated with an increased risk of cognitive impairment, whereas physical activity, alcohol consumption, and HDL-C showed protective associations. Increased variability in HDL-C and FPG were associated with a heightened risk of cognitive impairment across all three assessment methods. To enhance our understanding of the relationship between vascular risk factors and cognitive impairment in the elderly population, more comprehensive and prospective studies are warranted. Methods Study design This was a community-based study that combined a retrospective cohort design for the assessment of vascular risk factors and their long-term variability (using data from 2018 to 2021) with a cross-sectional assessment for cognitive status at the study endpoint (in 2022). Participants In this study, individuals aged 65 years and older were recruited in the community from January 2022 to April 2022, and received a health examination and cognitive function assessment at community hospital according to National Basic Public Health Service Norms. We conducted a retrospective review of the health examination data for all participants from 2018 to 2021, Individuals were excluded if they had one of the following: (1) absence of baseline data from 2018; (2) fewer than three health examinations; (3) missing data for one or more key vascular risk factors (e.g., blood pressure, lipid profiles, or fasting plasma glucose) at any of the required time points; (4) a prior diagnosis of stroke, dementia, or any memory-related disorder established before 2018. Ultimately, 1,487 subjects were included in the analysis. This study was conducted in accordance with the principles of the Declaration of Helsinki and ethically approved by the Ethics Committee of Qingdao Center for Disease Control and Prevention. All subjects have signed informed consent before they were included. Data collection and definitions Sociodemographic characteristics, lifestyles, and clinical characteristics were obtained through a standardized questionnaire administered by trained personnel, including age, sex, education, smoking status, drinking status, physical activity levels. Participants engaging in an extensive workout (e.g., brisk walking, jogging, running, cycling, swimming, Tai Chi) are deemed to fulfill the criteria for physical activity or exercise qualified, rather than simply walking around the house. BMI was computed using the formula weight (kg)/height (m)2. Blood pressure readings were obtained in a seated position utilizing a mercury sphygmomanometer, with the mean of three measurements calculated for both SBP and DBP. Trained professionals collected fasting venous blood specimens in the morning after at least 8 h of fasting from the participants, with subsequent clinical analyses conducted at the community hospital laboratory. Definition of variability indices Variability was defined as intra-individual variability in each parameter between four health examinations. Three indices of variability were used: coefficient of variation (CV), standard deviation (SD), and variability independent of the mean (VIM). The VIM was calculated as 100 × SD / meanβ, where β is the regression coefficient, based on the natural logarithm of the SD over the natural logarithm of the mean (10). Assessment of cognition Cognitive function was assessed utilizing the Chinese adaptation of the Mini-Mental State Examination (MMSE), this scale extensively employed in clinical evaluations of cognitive performance. This instrument consists of 11 items that examine multiple cognitive domains, such as orientation, memory, attention, calculation, recall, language, and visuospatial skills, resulting in a maximum score of 30. The evaluation was carried out through direct interviews, with all questions answered by the participants themselves, without the involvement of a proxy. Cognitive impairment was defined as a MMSE score of < 26. Statistical analysis Descriptive characteristics are presented as the mean ± standard deviation or number (percentage). Differences between participants with cognitive impairment and those without were compared using Student’s t-test or Mann-Whitney U test for continuous variables and Pearson’s chi-sqaure test for categorical variables. Logistic regression was used to analyse the odds ratio (OR) and 95% confidence interval (CI) for the associations of various vascular risk factors and their variability with cognitive impairment. Specifically, the independent variable vascular risk factors or variability were divided into quartiles and entered into the model as a categorical variable. Logistic regression was conducted to investigate the association between vascular risk factors or variability and cognitive impairment, using the lowest quartile of vascular risk factors or variability as the reference category. Two multivariable models were established according to covariates. Model 1 was non-adjusted; Model 2 was adjusted for age, sex, educational level. All analyses were conducted using SPSS (version 21.0) and R software (version 4.0.3). p < 0.05 was considered statistically significant. Acknowledgements We would like to thank the community hospital and Qingdao Centers for Disease Prevention and Control for their support in carrying out the study. Author contributions YFL, XYW designed the study. XYW, SYT, CC conducted the investigation and data analysis. FN supervised the project and acquired funding. YFL, HQX drafted the manuscript. XYW, YFL, SYT and CC extensively revised the manuscript. All authors approved the final manuscript. Funding The study was supported by the Qingdao Science and Technology Fund (21-1-4-rkjk-1-nsh). Data availability The datasets used in this analysis are available from the corresponding author. Declarations Ethics approval and consent to participate The study was conducted in accordance with the principles of the Declaration of Helsinki and ethically approved by the Ethics Committee of Qingdao Center for Disease Control and Prevention. All subjects have signed informed consent. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used in this analysis are available from the corresponding author.