https://orcid.org/0009-0001-3206-0353
Figures
Abstract
Background
Cognitive impairment has been a major public health challenge. This study aims to estimate the association of hemoglobin glycation index (HGI) with cognitive function in U.S. older adults.
Methods
In this cross-sectional study, data from the National Health and Nutrition Examination Survey (NHANES, 2011–2014) was obtained. Cognitive function was assessed based on the three scales [Digit Symbol Substitution Test (DSST), Animal Fluency Test (AFT), and Consortium to Establish a Registry for Alzheimer’s disease (CERAD)], along with a composite Z-score derived from the sum of the Z-scores of these three assessments. We used weighted univariate and multivariate linear regression to estimate the unstandardized β coefficient and 95% confidence interval (CI) of the association between HGI levels and cognitive function.
Results
Our study comprised a total of 1,406 subjects. As compared with the second tertile of HGI, the lowest tertile had a lower Z-score [β = −0.13 (95% CI: −0.25, −0.01)] and CERAD score [β = −1.07 (95% CI: −1.93, −0.21)] in the fully adjusted model. For the highest tertiles of HGI (vs. second tertile), Z-score and DSST decreased by 0.16 (95%CI: −0.31, −0.01) and 3.09 (95%CI: −6.00, −0.18), respectively. Females in the lowest HGI tertile (vs. second HGI tertile) exhibited a decline in cognitive function scores (Z-score, CERAD, AFT; all P < 0.05), while males in the highest tertile of HGI (vs. second HGI tertile) showed decreased cognitive function scores (Z-score, DSST; all P < 0.05). In addition, Z-score and CERAD scores also decreased (all P < 0.05) in the highest HGI tertile (vs. second HGI tertile) among non-users of antidiabetic drugs.
Conclusion
Both lower and higher HGI levels are associated with cognitive decline. Lower HGI is related to poorer learning and memory, while higher HGI is associated with executive function impairment. Future longitudinal studies are need to verify whether changes in HGI levels can be used as an early warning indicator for cognitive decline.
Citation: Ge L, Guo Y, Xu X, Xu J, Hu M (2026) Association between hemoglobin glycation index and cognitive function: Evidence in the elderly. PLoS One 21(5): e0338613. https://doi.org/10.1371/journal.pone.0338613
Editor: Li Yang, Sichuan University, CHINA
Received: February 18, 2025; Accepted: November 25, 2025; Published: May 15, 2026
Copyright: © 2026 Ge et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Cognitive impairment can be defined as a significant decline or loss in an individual’s cognitive functions, encompassing memory, language, attention, executive function, and visuospatial processing [1]. Deterioration in cognitive abilities is seen as the precursor of Alzheimer’s disease [2] and other dementias [3], which impose a notable effect on the overall well-being, autonomy, and health of senior individuals [4]. Alzheimer’s disease impacts approximately 6.9 million U.S. adults aged ≥ 65 and older, making it the fifth leading cause of death [5]. How to effectively identify and manage age-related cognitive decline has become an urgent public health problem to be solved.
As common metabolic abnormalities in the elderly, impaired glucose metabolism, and dysfunctional insulin activity are involved in the underlying mechanisms of cognitive decline associated with type 2 diabetes mellitus (DM) [6]. Maan et al. [7] found that high glycated hemoglobin (HbA1c)/uncontrolled DM and DM duration resulted in cognitive impairment. Although fasting plasma glucose (FPG) and HbA1c are commonly used in the diagnosis and management of DM in clinical practice, evidence suggests that the HbA1c level is not completely consistent with the FPG level [8,9]. Factors such as the mean lifespan of red blood cells, glucose transmembrane gradients across cell membranes, and enzyme abnormalities may influence HbA1c levels [10,11], indicating HbA1c measurements may not fully reflect glycemic metabolic status. Hemoglobin glycation index (HGI) has been introduced to measure the extent and direction of the variance between actual and expected HbA1c levels in individuals, aiming to characterize individual susceptibility to hemoglobin glycation [8]. Elevated HGI is associated with metabolic abnormalities, inflammation, and oxidative stress [12,13], while a low HGI may result from elevated FPG and low HbA1c due to stress-induced hyperglycemia [14,15]. Previous studies showed a U-shaped relationship of HGI levels with adverse cardiovascular events in patients with diabetes and coronary heart disease [15–17], and an L-shaped association of HGI levels with 30-day and in-hospital mortality in patients with ischemic stroke [18]. Recent findings have shown an association between increased HGI with shorter telomere length [14], which in turn is connected to cognitive decline [19]. However, the relationship of HGI with cognitive function remains unclear.
Our study aims to provide some evidences of the association between HGI levels and cognitive function in older adults. We believe that our findings can lead to better ways of identification and monitoring of cognitive impairment risk in the elderly through the use of readily accessible biomarkers.
2. Methods
2.1. Study population and design
As a representative, long-term program, the National Health and Nutrition Examination Survey (NHANES) database was designed to evaluate the health and nutritional conditions of U.S. citizens. The health data of participants were obtained through home visits and physical examinations. Data acquisition and details, as well as the design of the NHANES program, can be accessed at https://www.cdc.gov/nchs/nhanes/index.htm. The project was approved by the Ethics Review Board of National Center for Health Statistics (Protocol #2011−17, available at https://www.cdc.gov/nchs/nhanes/irba98.htm). Each survey participant signed a consent form.
Participants with cognitive function tests were included from two NHANES cycles (2011–2012 and 2013–2014). Initially, 3,632 participants ≥ aged 60 years were recruited, excluding missing data on FPG (n = 1,995), HbA1c (n = 4), the consortium to establish a registry for Alzheimer’s disease (CERAD) (n = 147), the animal fluency test (AFT) (n = 17) and the digit symbol substitution test (DSST) (n = 63), 1,406 participants were finally enrolled. The flowchart of participant selection is shown in Fig 1. The sensitivity analysis comparing baseline characteristics between included and excluded cohorts was presented in S1 Table in S1 File.
2.2. Definition of HGI
The HGI calculated relied on the availability of HbA1c and FPG data from participants. Blood samples of participants were obtained in the morning after fasting (Fasting for 8–24 hours before blood sampling). Based on the participants enrolled in this study, an equation between FPG and HbA1c was constructed [predicted HbA1c (%) = 0.422 × FPG (mmol/L) + 3.367, R2 = 0.594] (Fig 2). HGI was calculated by the following equation [20]: HGI = measured HbA1c - predicted HbA1c. All participants were grouped by tertiles of HGI: Tertile 1 (HGI < −0.226), Tertile 2 (−0.226 ≤ HGI < 0.119), Tertile 3 (≥ 0.119).
2.3. Cognitive Function
In NHANES 2011–2014, cognitive function tests were performed among adults aged ≥ 60 years. The CERAD assesses participants’ immediate and delayed learning abilities for new verbal information through three sequential learning trials and a delayed recall. In each trial, participants were tasked with reading aloud 10 unrelated words and were required to recall them immediately after the presentation or 10 minutes later. To evaluate verbal fluency and production, the AFT required participants to list as many animals as possible starting with a designated letter within a time limit of one minute. In DSST, participants were required to match different symbols to the corresponding numbers in a limited time to assess information processing speed and task execution ability. Finally, the composite-Z score was calculated from the sum of the Z scores [(individual test score – mean score)/standard deviation] of these three tests (DSST, AFT, and CERAD). For each cognitive function test, higher scores represented better cognitive function.
2.4. Covariates
Potential covariates included age (continuous, years), sex (male, female), race (non-Hispanic White, non-Hispanic Black, Mexican American, other race), poverty income ratio (PIR; < 1, ≥ 1), marriage (married or living with partner, living alone), education level (high school and below, above high school), physical activity [metabolic equivalent of task (MET)/ minutes per week; < 450, 450–750, ≥ 750, unknown], smoking (never, former, current), hypertension (no, yes), antidiabetic drug used (no, yes), dyslipidemia (no, yes), heavy drinking (no, yes, unknown), body mass index (BMI; kg/m2; < 25, ≥ 25), depression (no, yes), chronic kidney disease (CKD; no, yes), stroke (no, yes), antiparkinson agents (no, yes), anxiolytics sedatives hypnotics (no, yes), psychotherapeutic agents (no, yes), total energy (continuous, kcal), carbohydrate (continuous, gm), protein (continuous, gm), total fat (continuous, gm), total sugars (continuous, gm). Please refer to our S1 File for detailed variable definitions.
2.5. Statistical analysis
To address the multistage design, we employed SURVEY procedures which the National Institutes of Health (NCHS) recommended. The procedures incorporate the morning fasting subsample 2-year weights (WTSAF2YR), strata (SDMVSTRA), along primary sampling units (SDMVPSU) into the weighted analyses. The linearity, normality and homoscedasticity of residuals, and residual plots were shown in S1–S4 Fig. All figures (S1A–S4A Figs) showed that the red line was roughly a straight line parallel to the X-axis. That is, the residuals were not affected by the independent variables. In Figs (S1B–S4B), the residuals were normally distributed. The red line was roughly a straight line parallel to the X-axis, indicating that the variance of the residuals was homogeneous (S1C–S4C Figs). No outliers were found in residual plots (S1D–S4D Figs). Mean ± standard error and the weighted one-way analysis of variance (ANOVA) were used in the descriptive analysis of continuous variables. Frequencies (percentage) [n (%)] and the Rao-Scott Chi-square test were used for the categorical variables. The missing data was evaluated by Little’s MCAR test. The result of P-value was 0.157, indicating the missing data conformed to complete random missing. Missing data were handled using the multiple imputation method based on chained equations under the missing at random. Five imputed datasets were generated and analyzed separately, and the results were pooled according to Rubin’s rules. The missing values were presented in Table S2 in S1 File. The sensitivity analysis on comparison of variables before and after data interpolation was shown in Table S3 in S1 File.
All results were analyzed by R software (version 4.3.3), with a significant set at a P-value < 0.05. The restricted cubic spline (RCS) curves were used to assess the nonlinear correlations of HGI levels and cognitive function scores. The initial weighted univariate linear regression model incorporated all potential variables and those with a P < 0.05 were then entered into a backward stepwise regression analysis to identify the relevant covariates (Table S4-S7 in S1 File). The AIC values of Z-score, CERAD, AFT, and DSST were 3050.86, 8760.71, 8437.54, and 11134.10, respectively. The multicollinearity assessment for covariates were presented in Table S8 in S1 File. The results show that there is no collinearity among covariates. Subsequently, the associations of cognitive function (CERAD, AFT, DSST, and cognitive Z-score) with HGI were evaluated by the univariate and multivariable linear regression models. The unstandardized β and confidence interval (CI) was calculated. An in-depth exploration of the association between HGI and cognitive function was conducted using subgroup analysis (grouped participants according to sex and the use of antidiabetic drugs). The interaction terms of subgroups were further tested. Due to multiple testing may increase the risk of Type I error, a Bonferroni correction was considered to assess the results of all analyses. The Bonferroni correction has a significance level of α’ = 0.05/n, where n is the number of subgroups to be compared (n = 4 in the main analyses, and n = 16 in the subgroups), resulting in an adjusted significance level of α’ = 0.0125 or 0.003125. We considered the result in the subgroups still statistically significant only if the P-value <0.003125.
3. Results
3.1. Characteristics of study participants
This cross-sectional study enrolled 1,406 samples. The characteristics of the included participants were presented in Table 1. The HGI was categorized into Tertile 1 (HGI < −0.266), Tertile 2(−0.266 ≤ HGI < 0.119), and Tertile 3 (HGI ≥ 0.119). The mean (±SD) age of the participants was 69.09 (±0.26) years with most participants being female (n = 715, 55.74%) and non-Hispanic White (n = 708, 79.26%). Our results exhibited differences in Z-scores, CERAD scores, AFT scores, and DSST scores among different HGI level groups were observed (all P < 0.05). Among the different HGI level groups, differences in nutrient intake (total energy, carbohydrates, protein, total fat, and total sugars), as well as glycation indices (HbA1c and FPG), were also found (all P < 0.05). The race, PIR, education level, PA, antidiabetic drug use, heavy drinking, BMI, and CKD history were statistical differences among the tree groups (all P < 0.05).
3.2. Association between HGI and cognitive function
Tables S3-S6 in S1 File exhibited the process of covariables selection. In adjusted model (Table 2), participants with the lowest tertile of HGI had lower Z-score scores [β = −0.13 (95% CI: −0.25, −0.01)] and CERAD scores [β = −1.07 (95%CI: −1.93, −0.21)] than those in the Tertile 2 group. The Z-score [β = −0.16 (95% CI: −0.31, −0.01)] and DSST (β = −3.09 (95% CI: −6.00, −0.18)] in the highest tertile of the HGI group were decreased than those in the Tertile 2 group. Additionally, when HGI was included as a continuous variable, HGI levels were not associated with Z-score, CERAD, AFT and DSST (all P > 0.05). The nonlinear correlation of cognitive function scores (Z-score, CERAD, AFT and DSST) with HGI was displayed in Fig 3. The nonlinear correlation was found between HGI and Z-score, DSST (P < 0.05).
A. Z-score, B. CERAD, C. AFT, D. DSST.
3.3. Subgroup analysis
Further exploratory analysis of the association between HGI and cognitive function was evaluated by the stratification of sex and antidiabetic drugs used (Fig 4). In the subgroup of females, the lowest tertile (vs. Tertile 2) of HGI was associated with decreased scores of Z-scores, CERAD, and AFT by 0.21 (95%CI: −0.36, −0.05), 1.19 (95%CI: −2.14, −0.25), and 1.26 (95%CI: −2.28, −0.23), respectively. However, in male participants, 0.23 (95%CI: −0.43, −0.02) and 4.79 (95%CI: −8.37, −1.22) decreases in Z-score and DSST scores were observed for the highest tertile of HGI levels (vs. Tertile 2). Compared with the Tertile 2 of the HGI level, participants who did not use antidiabetic drugs had an increased Z-score [β = −0.13 (95%CI: −0.26, −0.01)] and CERAD scores [β = −1.17 (95%CI: −2.05, −0.28)] in the lowest tertile of the HGI level. For participants using antidiabetic drugs, each increase in HGI was associated with a decrease of 1.71 (95% CI: −3.01, −0.42) in the DSST score.
4. Discussion
This representative cross-sectional study of the U.S. elderly population found associations between lower or higher levels of HGI and cognitive function decline (Z-score). Specifically, lower HGI was associated with impaired learning and memory (CERAD), while higher HGI was associated with impaired executive function (DSST). We also observed lower CERAD scores and AFT scores for females or non-users of antidiabetic drug with lower HGI, and lower DSST scores for males with higher HGI.
Glucose serves as the primary energy source for the brain, and changes in blood glucose concentration can influence individual brain function [21]. A meta-analysis indicated a non-linear association between FPG and cognitive impairment, while elevated 2-hour post-load glucose and HbA1c levels were related to an increased developing dementia risk. [22]. A community-based study of 13,351 adults found associations between diabetes, pre-diabetes, poorly controlled diabetes, and a longer duration of diabetes with a lower Z-score [23]. Moreover, the association between hyperglycemic status and cognitive decline appears to be primarily driven by declines in DSST and AFT [23,24]. Approximately 30 million people in the U.S. might experience a clinical mismatch between their HbA1c and glucose levels, potentially resulting in either misdiagnosis or inappropriate management related to HbA1c [25]. Low HbA1c or low FPG does not necessarily indicate good blood glucose control [26], which may not accurately reflect individual glucose variability and long-term stability. Cognitive decline was associated with both lower and higher levels of HGI in this study. Unlike previous studies, we found that lower HGI was associated with impaired learning and memory, whereas higher HGI was associated with impaired executive function, indicating that the effects of different glycemic states on cognition may involve distinct cognitive domains.
Elevated HGI levels were associated with greater production of advanced glycation end-products (AGEs) [27]. Increased AGEs can induce the production of reactive oxygen species (ROS) and activate stress-related signal transduction molecules (p38 mitogen-activated protein kinase and c-Jun N-terminal kinase) [28], resulting in abnormal hyperphosphorylation of tau, which is also a hallmark of Alzheimer’s dementia [29]. A mouse model found that oxidative stress induced by high glucose or AGEs inhibited mitochondrial autophagy in hippocampal neurons through Kelch-like ECH-associated protein 1-nuclear factor erythroid 2-related factor 2-prohibitin 2 pathway, leading to mitochondrial dysfunction and cognitive decline [30]. Mitochondrial dysfunction caused by hyperglycemia also accelerates the generation of ROS and further exacerbates neuronal oxidative damage [31]. In addition, excessive production of AGEs induces endoplasmic reticulum stress [32,33], and telomere length shortening [12,19] is associated with cognitive decline. Regarding the observed decline in cognitive function within the low HGI group, we speculate that it may be related to stress hyperglycemia [14,15]. Low HGI may be associated with higher insulin resistance and gluconeogenesis caused by counterregulatory hormones and pro-inflammatory cytokines in stress hyperglycemia, and the increase of FPG [14,15]. Frequent changes in glucose levels can hinder the supply of glucose to brain tissue, potentially disrupting cognitive processes [34]. The hippocampus is an important region for memory formation [35]. Reducing scores for immediate memory and delayed recall were found to be associated with several cortical and subcortical sites [36]. These areas include the parahippocampal region, posterior cingulate gyrus, right thalamus, and right hippocampus [36]. An animal study found that adult male mice experiencing chronic social failure stress developed hyperglycemia, which was associated with hippocampus-related spatial memory impairment [37]. Our study highlights the relationship of changes in HGI levels with cognitive function, and the potential benefits of early intervention need further investigation in longitudinal studies.
The association of reduced HGI with cognitive decline (Z-score and CERAD) was found in females. The reason for this may be that older women experience menopause, estrogen changes, which may be related to cognitive changes [38,39]. When the menopausal estrogen is reduced, inflammatory markers (interleukin-6, interleukin-1, and tumor necrosis factor-α) are also increased, and this inflammatory state may result in further aggravation of cognitive impairment [40]. The observed association between high HGI and lower DST scores in males may be due to mechanisms such as oxidative stress or mitochondrial damage caused by high HGI [28–30]. The observed lower cognitive level in the low HGI group among participants not eligible for antidiabetic drugs may be primarily driven by CERAD. However, the specific mechanism of HGI differences in cognitive function among different sex and antidiabetic drug-used participants was not available in this study, and further experimental and animal studies may help to explain our results.
Our study presents, varying associations between different HGI and cognitive functions in different domains in the elderly. Our findings suggest that HGI provides simple and easily available markers to detect cognitive deterioration in older adults. From a clinical perspective, changes in HGI levels may serve as an early warning indicator for cognitive decline, monitoring of HGI provides an effective tool for identifying older adults who may develop cognitive dysfunction. Physicians can timely identify patients with poor glycemic control by regular examination of HGI, and take corresponding interventions, such as diet modification, drug therapy, or lifestyle intervention, to improve metabolic status, which may slow down the decline of cognitive function.
We must acknowledge that several limitations remain. First of all, the cross-sectional nature of this study restricts our ability to reveal temporal or causal inferences related to the observed associations. Secondly, due to the limited cognitive function tests, this study only focused on the NHANES data from 2011 to 2014. Thirdly, the exclusion of approximately 54.9% of eligible participants due to incomplete FPG data may introduce selection bias. Fourthly, there may still be some confounders that could not be considered although adequate controls for potential confounding factors were implemented, which may introduce a degree of bias. Finally, the direction of HGI changes, different study populations may be related to different functions of HGI-induced cognitive function changes. The underlying mechanisms need to be further verified.
5. Conclusion
In the present study, lower HGI or higher HGI may be associated with cognitive decline among older adults in the U.S. In addition, lower HGI may associated with worse learning and memory dimensions, while higher HGI may be associated with impaired executive function. While the associations differ among populations with different sex and antidiabetic drugs used, our results highlight the necessity for continued investigations on the underlying mechanism of HGI and cognitive function. The cross-sectional nature of this study restricts our ability to reveal temporal or causal inferences related to the observed associations. Future studies need to further validate our findings.
Supporting information
S1 Fig. The linearity, normality and homoscedasticity of residuals, and residual plots (Z-score).
A. linearity; B. normality; C. homoscedasticity; D. residual plots.
https://doi.org/10.1371/journal.pone.0338613.s001
(PDF)
S2 Fig. The linearity, normality and homoscedasticity of residuals, and residual plots (CREAD score).
A. linearity; B. normality; C. homoscedasticity; D. residual plots.
https://doi.org/10.1371/journal.pone.0338613.s002
(PDF)
S3 Fig. The linearity, normality and homoscedasticity of residuals, and residual plots (AFT score).
A. linearity; B. normality; C. homoscedasticity; D. residual plots.
https://doi.org/10.1371/journal.pone.0338613.s003
(PDF)
S4 Fig. The linearity, normality and homoscedasticity of residuals, and residual plots (DSST score).
A. linearity; B. normality; C. homoscedasticity; D. residual plots.
https://doi.org/10.1371/journal.pone.0338613.s004
(PDF)
References
- 1. McCollum L, Karlawish J. Cognitive Impairment Evaluation and Management. Med Clin North Am. 2020;104(5):807–25. pmid:32773047
- 2.
Lu X, Zhang Y, Tang Y, Bernick C, Shan G. Conversion to Alzheimer’s disease dementia from normal cognition directly or with the intermediate mild cognitive impairment stage. 2024:1–14.
- 3. Jia L, Du Y, Chu L, Zhang Z, Li F, Lyu D, et al. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross-sectional study. Lancet Public Health. 2020;5(12):e661–71. pmid:33271079
- 4. Aye S, Bouteloup V, Tate A, Wimo A, Handels R, Jean D, et al. Health-related quality of life in subjective cognitive decline and mild cognitive impairment: a longitudinal cohort analysis. Alzheimers Res Ther. 2023;15(1):200. pmid:37968734
- 5. 2024 Alzheimer’s disease facts and figures. Alzheimers Dement. 2024;20(5):3708–821. pmid:38689398
- 6. Gupta M, Pandey S, Rumman M, Singh B, Mahdi AA. Molecular mechanisms underlying hyperglycemia associated cognitive decline. IBRO Neurosci Rep. 2022;14:57–63. pmid:36590246
- 7. Maan HB, Meo SA, Rouq FA, Meo IMU. Impact of glycated hemoglobin (HbA1c) on cognitive functions in Type 2 diabetic patients. Eur Rev Med Pharmacol Sci. 2021;25(19):5978–85. pmid:34661257
- 8. Hempe JM, Gomez R, McCarter RJ Jr, Chalew SA. High and low hemoglobin glycation phenotypes in type 1 diabetes: a challenge for interpretation of glycemic control. J Diabetes Complications. 2002;16(5):313–20. pmid:12200073
- 9. Nathan DM, Kuenen J, Borg R, Zheng H, Schoenfeld D, Heine RJ, et al. Translating the A1C assay into estimated average glucose values. Diabetes Care. 2008;31(8):1473–8. pmid:18540046
- 10. Khera PK, Joiner CH, Carruthers A, Lindsell CJ, Smith EP, Franco RS, et al. Evidence for interindividual heterogeneity in the glucose gradient across the human red blood cell membrane and its relationship to hemoglobin glycation. Diabetes. 2008;57(9):2445–52. pmid:18591386
- 11. Malka R, Nathan DM, Higgins JM. Mechanistic modeling of hemoglobin glycation and red blood cell kinetics enables personalized diabetes monitoring. Sci Transl Med. 2016;8(359):359ra130. pmid:27708063
- 12. Lyu L, Yu J, Liu Y, He S, Zhao Y, Qi M, et al. High hemoglobin glycation index is associated with telomere attrition independent of HbA1c, mediated by TNFα. J Clin Endocrinol Metab. 2022;107(2):462–73. pmid:34562085
- 13. Rajendran S, Mishra S, Madhavanpillai M, G V. Association of hemoglobin glycation index with cardiovascular risk factors in non-diabetic adults: A cross-sectional study. Diabetes Metab Syndr. 2022;16(9):102592. pmid:35998512
- 14. Dungan KM, Braithwaite SS, Preiser J-C. Stress hyperglycaemia. Lancet. 2009;373(9677):1798–807. pmid:19465235
- 15. Lin Z, He J, Yuan S, Song C, Bian X, Yang M, et al. Hemoglobin glycation index and cardiovascular outcomes in patients with diabetes and coronary artery disease: insights from a large cohort study. Nutr Diabetes. 2024;14(1):69. pmid:39191777
- 16. Wang Y, Liu H, Hu X, Wang A, Wang A, Kang S, et al. Association between hemoglobin glycation index and 5-year major adverse cardiovascular events: the REACTION cohort study. Chin Med J (Engl). 2023;136(20):2468–75. pmid:37265382
- 17. Wei X, Chen X, Zhang Z, Wei J, Hu B, Long N, et al. Risk analysis of the association between different hemoglobin glycation index and poor prognosis in critical patients with coronary heart disease-a study based on the MIMIC-IV database. Cardiovasc Diabetol. 2024;23(1):113. pmid:38555454
- 18. Huang X, Huang L, Tao H, Ren M, Yan L. Nonlinear association between hemoglobin glycation index and mortality in ischemic stroke Patients: Insights from the MIMIC-IV database. Diabetes Res Clin Pract. 2025;224:112105. pmid:40096948
- 19. Linghui D, Shi Q, Chi C, Xiaolei L, Lixing Z, Zhiliang Z, et al. The association between leukocyte telomere length and cognitive performance among the american elderly. Front Aging Neurosci. 2020;12:527658. pmid:33192450
- 20. Huang Y, Huang X, Zhong L, Yang J. Glycated haemoglobin index is a new predictor for all-cause mortality and cardiovascular mortality in the adults. Sci Rep. 2024;14(1):19629. pmid:39179628
- 21. Luo W, Wang J, Chen M, Zhou S, Deng D, Liu F, et al. Alterations of cerebral blood flow and its connectivity in olfactory-related brain regions of type 2 diabetes mellitus patients. Front Neurosci. 2022;16:904468. pmid:35898415
- 22. Xue M, Xu W, Ou Y-N, Cao X-P, Tan M-S, Tan L, et al. Diabetes mellitus and risks of cognitive impairment and dementia: A systematic review and meta-analysis of 144 prospective studies. Ageing Res Rev. 2019;55:100944. pmid:31430566
- 23. Rawlings AM, Sharrett AR, Schneider ALC, Coresh J, Albert M, Couper D, et al. Diabetes in midlife and cognitive change over 20 years: a cohort study. Ann Intern Med. 2014;161(11):785–93. pmid:25437406
- 24. Casagrande SS, Lee C, Stoeckel LE, Menke A, Cowie CC. Cognitive function among older adults with diabetes and prediabetes, NHANES 2011-2014. Diabetes Res Clin Pract. 2021;178:108939. pmid:34229005
- 25. Staimez LR, Kipling LM, Nina Ham J, Legvold BT, Jackson SL, Wilson PWF, et al. Potential misclassification of diabetes and prediabetes in the U.S.: mismatched HbA1c and glucose in NHANES 2005-2016. Diabetes Res Clin Pract. 2022;189:109935. pmid:35662612
- 26. Yu J, Hua H, Yin M. U-shaped association between HbA1c and all-cause mortality in CVD patients with diabetes. Sci Rep. 2024;14(1):28386. pmid:39551840
- 27. Felipe DL, Hempe JM, Liu S, Matter N, Maynard J, Linares C, et al. Skin intrinsic fluorescence is associated with hemoglobin A(1c)and hemoglobin glycation index but not mean blood glucose in children with type 1 diabetes. Diabetes Care. 2011;34(8):1816–20. pmid:21636794
- 28. Zhang J, Wang X, Vikash V, Ye Q, Wu D, Liu Y, et al. ROS and ROS-mediated cellular signaling. Oxidative Med Cellular Longevity. 2016;2016:4350965.
- 29. Palmqvist S, Tideman P, Cullen N, Zetterberg H, Blennow K, Alzheimer’s Disease Neuroimaging Initiative, et al. Prediction of future Alzheimer’s disease dementia using plasma phospho-tau combined with other accessible measures. Nat Med. 2021;27(6):1034–42. pmid:34031605
- 30. Xu S, Gao Z, Jiang L, Li J, Qin Y, Zhang D, et al. High glucose- or AGE-induced oxidative stress inhibits hippocampal neuronal mitophagy through the Keap1-Nrf2-PHB2 pathway in diabetic encephalopathy. Sci Rep. 2024;14(1):24044. pmid:39402106
- 31. Rizwan H, Pal S, Sabnam S, Pal A. High glucose augments ROS generation regulates mitochondrial dysfunction and apoptosis via stress signalling cascades in keratinocytes. Life Sci. 2020;241:117148. pmid:31830478
- 32. Liu D, Zhang M, Yin H. Signaling pathways involved in endoplasmic reticulum stress-induced neuronal apoptosis. Int J Neurosci. 2013;123(3):155–62. pmid:23134425
- 33. Passarelli M, Machado UFF. AGEs-induced and endoplasmic reticulum stress/inflammation-mediated regulation of GLUT4 expression and atherogenesis in diabetes mellitus. Cells. 2021;11(1):104. pmid:35011666
- 34. Hasan F, Jirout J, Garzione S, Kranz S. Changes in Learning Outcomes after Dietary Intervention in Preschoolers: A Pilot Study. Nutrients. 2021;13(6):1797. pmid:34070396
- 35. Husain M. The human hippocampus contributes to short-term memory. Brain. 2024;147(8):2593–4. pmid:38978495
- 36. Dos Santos V, Thomann PA, Wüstenberg T, Seidl U, Essig M, Schröder J. Morphological cerebral correlates of CERAD test performance in mild cognitive impairment and Alzheimer’s disease. J Alzheimers Dis. 2011;23(3):411–20. pmid:21116054
- 37. van der Kooij MA, Jene T, Treccani G, Miederer I, Hasch A, Voelxen N, et al. Chronic social stress-induced hyperglycemia in mice couples individual stress susceptibility to impaired spatial memory. Proc Natl Acad Sci U S A. 2018;115(43):E10187–96. pmid:30301805
- 38. Schwarz KG, Vicencio SC, Inestrosa NC, Villaseca P, Del Rio R. Autonomic nervous system dysfunction throughout menopausal transition: a potential mechanism underpinning cardiovascular and cognitive alterations during female ageing. J Physiol. 2024;602(2):263–80. pmid:38064358
- 39. Sochocka M, Karska J, Pszczołowska M, Ochnik M, Fułek M, Fułek K, et al. Cognitive decline in early and premature menopause. Int J Mol Sci. 2023;24(7):6566. pmid:37047549
- 40. Zhang Y, Peng Y, Deng W, Xiang Q, Zhang W, Liu M. Association between dietary inflammatory index and cognitive impairment among American elderly: a cross-sectional study. Front Aging Neurosci. 2024;16:1371873. pmid:38550747