Association of Hemoglobin Glycation Index With All-Cause Mortality, Cardiac Mortality, and Cardiovascular Mortality in the General Population: A Retrospective Cohort Study of NHANES Data

Abstract

Background:

The hemoglobin glycation index (HGI) presents a discrepancy between observed and predicted glycosylated hemoglobin (HbA1c) and fasting blood glucose values. Meanwhile, compared to the HbA1c values, the HGI provides a more comprehensive reflection of blood glucose variability across populations. However, no studies have examined the association between the HGI and all-cause, cardiac, and cardiovascular mortalities in the general population. Hence, this study aimed to investigate these relationships using data from the National Health and Nutrition Examination Survey (NHANES) database.

Methods:

Participants were stratified into four groups based on the HGI quartiles. Weighted multivariable Cox proportional hazards models were used to assess the associations between HGI and all-cause, cardiovascular, and cardiac mortality. Kaplan–Meier survival analysis based on the HGI quartiles and log-rank tests were employed to compare differences in primary and secondary endpoints. Additionally, restricted cubic spline (RCS) curves were used to explore nonlinear relationships between the HGI and endpoints, identifying inflection points. Subgroup analyses and interaction tests were conducted to assess the robustness of the findings.

Results:

In comparing the baseline characteristics of endpoints across all-cause mortality, cardiac mortality, and cardiovascular mortality, significantly higher mortality rates were observed in the high HGI quartile group (Q4) compared to the other three groups (Q1, Q2, and Q3) (p < 0.05). Kaplan–Meier curves demonstrated increased mortality risks in the high HGI group across all endpoints (p < 0.05). Multivariable Cox proportional hazards models indicated that high HGI levels were associated with all-cause mortality (Q4: hazard ratio (HR) (95% confidence interval (CI)) = 1.232 (1.065, 1.426); p = 0.005), cardiac mortality (HR (95% CI) = 1.516 (1.100, 2.088); p = 0.011) and cardiovascular mortality (HR (95% CI) = 1.334 (1.013, 1.756); p = 0.039). Low HGI was associated only with all-cause mortality (Q1: HR (95% CI) = 1.269 (1.082, 1.488); p = 0.003). RCS analysis confirmed a U-shaped relationship between the HGI and all three outcome events. Subgroup analyses and interaction tests supported the robustness of the conclusions.

Conclusion:

This study demonstrates a U-shaped association between the HGI and overall mortality, cardiac mortality, and cardiometabolic mortality in the general population. Specifically, the high HGI value represented a risk factor for all-cause, cardiac, and cardiovascular mortality. In contrast, low HGI values were associated only with all-cause mortality in the general population.

Graphical abstract

Keywords

glycated hemoglobin / NHANES / mortality / cardiovascular diseases

1. Introduction

Cardiovascular diseases (CVDs), characterized by their high morbidity and mortality rates, pose a significant global health challenge and a substantial burden on healthcare systems worldwide. As populations age, CVDs have emerged as the leading cause of death globally, responsible for nearly one-third of all fatalities, with a marked increase of 12.5% observed over the past decade [1, 2]. CVDs now rank as the foremost contributor to reduced life expectancy among older people [2, 3]. To effectively reduce CVD risk, early identification of individuals at high risk during the progression of cardiovascular disease is crucial.

The hemoglobin glycation index (HGI) is derived from the difference between observed and predicted glycosylated hemoglobin (HbA1c) values. Moreover, the HGI serves as a reliable metric to quantify glucose metabolism and individual variability, having been validated as an effective measure of deviations in HbA1c levels [4]. Although HbA1c remains the gold standard for assessing glycemic control in diabetes, its levels can be influenced by factors such as red blood cell turnover and glucose gradients across the red blood cell membrane, reflecting only 60–80% of the glucose levels found in the body [5]. To address these limitations and provide a more accurate assessment of glycemic variability, Hempe and colleagues [6] introduced the HGI, which has subsequently been demonstrated to improve blood glucose variability capturing in diverse populations [7].

Given that the HGI represents blood glucose variability, prior research has primarily focused on its prognostic implications in diabetic populations. However, studies investigating the correlation between the HGI and overall mortality rates in the general population are scarce [8, 9]. Indeed, recent research indicated a U-shaped relationship between the HGI and all-cause mortality in the general population [10]. However, the association between the HGI and cardiac and cardiovascular mortality in the general population remains unclear.

To our knowledge, no studies have yet investigated the relationship between HGI and cardiometabolic mortality. Therefore, this study aimed to utilize data from the National Health and Nutrition Examination Survey (NHANES) database to conduct a retrospective cohort analysis. Moreover, this study sought to explore the associations and disparities of the HGI with all-cause mortality, cardiac mortality, and cardiovascular mortality in the general population.

2. Methods and Materials

2.1 Study Population

The NHANES database is a periodic survey conducted by the National Center for Health Statistics (NCHS) of the United States. The NHANES database systematically examines a random sample of American citizens through comprehensive physical examinations and questionnaires. Indeed, the NHANES collects data on physiological measurements, nutritional status, health surveys, and environmental factors to assess the health and nutritional status of the U.S. population (https://wwwn.cdc.gov/nchs/nhanes/Default.aspx). The NHANES has received approval from the NCHS Research Ethics Review Board to ensure ethical standards in research.

The exclusion criteria included individuals with severe liver or kidney disease, those lacking follow-up data or inadequate follow-ups, individuals missing baseline data such as glycated hemoglobin, fasting glucose, or lipid levels, and those with less than 8 hours of fasting before blood specimen collection. A flowchart outlining the patient selection process is presented in Fig. 1. Following the application of these criteria, our study included 18,171 samples spanning from 1999 to 2018, with participants undergoing up to 10 follow-up assessments over a median follow-up period of 112 months.

2.2 Data Collection

Our study incorporated four primary categories of covariates: (1) demographic characteristics, including sex, age, race, smoking status, alcohol consumption, marital status, and education level; (2) general physical measures, such as body mass index (BMI); (3) laboratory parameters encompassing alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine (CR), blood urea nitrogen (BUN), triglyceride (TG) level, total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), uric acid (UA), fasting plasma glucose (FPG), HbA1c, white blood cell count (WBC), neutrophil count (NC), lymphocyte count (LC), platelet count (PLT), and red blood cell count (RBC); (4) medical history including hypertension, coronary artery disease (CAD), angina pectoris, chronic heart failure (CHF), diabetes mellitus (DM), chronic obstructive pulmonary disease (COPD), cancer, chronic kidney disease (CKD), and hyperlipidemia. The HGI was calculated using the formula HGI = observed HbA1c – predicted values (predicted HbA1c values derived from a linear regression equation according to the FPG and observed HbA1c) [11]. All blood specimens were collected following a minimum 8-hour fasting period.

2.3 Definition of Exposure Variables and Outcome Events

The primary endpoint of this study was defined as all-cause mortality (cause of death: Alzheimer’s disease (G30), diseases of the heart (I00–I09, I11, I13, I20–I51), chronic lower respiratory diseases (J40–J47), malignant neoplasms (C00–C97), all other causes (residual), cerebrovascular diseases (I60–I69), accidents (unintentional injuries) (V01–X59, Y85–Y86), diabetes mellitus (E10–E14), influenza and pneumonia (J09–J18), nephritis, nephrotic syndrome, and nephrosis (N00–N07, N17–N19, N25–N27)), while the secondary endpoint was defined as cardiac mortality (diseases of the heart (I00–I09, I11, I13, I20–I51)) or cardiovascular mortality (diseases of the heart (I00–I09, I11, I13, I20–I51), cerebrovascular diseases (I60–I69)). Follow-up continued until the time of death.

2.4 Statistical Analyses

Statistical analyses were performed using R (version 4.2.3; The R Foundation for Statistical Computing, Vienna, Austria), EmpowerStats (version 4.2; X&Y Solutions, Hangzhou, Zhejiang, China), and GraphPad Prism (version 9.0; GraphPad Software, San Diego, CA, USA). Participants were stratified into four quartiles based on the HGI values: Q1 (HGI –0.287, n = 4535), Q2 (–0.287 \le HGI –0.028, n = 4540), Q3 (–0.028 \le HGI 0.230, n = 4552), and Q4 (HGI \ge 0.230, n = 4544). Continuous variables were expressed as medians (interquartile range: P25, P75), and categorical variables were reported as counts (percentages). The normality test was conducted using the Kolmogorov-Smirnov test. Group differences for continuous variables were assessed using analysis of variance (ANOVA) and ANOVA post hoc test or the Kruskal–Wallis test, while chi-square tests were used for categorical variables. A two-sided p-value 0.05 was considered statistically significant.

The association between the HGI and overall mortality, cardiac mortality, and cardiovascular mortality in the general population was evaluated using multivariable Cox proportional hazards models. Given that the Q2 group exhibited the lowest relative risk, this group was selected as the reference group for the Cox regression analysis. Kaplan–Meier survival analysis was performed to compare survival outcomes across the HGI quartiles, with differences assessed using the log-rank test. Restricted cubic spline (RCS) curves were employed to explore potential nonlinear relationships between the HGI and the primary and secondary endpoints, including identifying inflection points. RCS is a statistical method that partitions the independent variable into multiple intervals and applies cubic polynomials within each segment to capture complex nonlinear trends. Finally, subgroup analyses were conducted using forest plots to evaluate the influence of confounding variables on the study outcomes, ensuring the robustness of the findings based on clinical expertise.

3. Results

3.1 Characteristics of the Study Population Based on HGI Quartiles

This study analyzed data from 18,171 participants in the NHANES database, including 9044 males (48.99%) and 9127 females (51.01%). Participants were stratified into quartiles based on their HGI levels (Table 1).

Significant differences were observed across the HGI quartiles. Gender distribution revealed a trend whereby the proportion of females included in the higher HGI levels increased whereas the male participants decreased (females: Q1 = 1878 (41.35%), Q2 = 2292 (52.17%), Q3 = 2458 (56.07%), Q4 = 2499 (57.01%); males: Q1 = 2657 (58.65%), Q2 = 2248 (47.83%), Q3 = 2094 (43.93%), Q4 = 2045(42.99%)) (p 0.05) (Table 1). The racial analysis showed that the proportion of non-Hispanic white individuals decreased, while non-Hispanic black individuals increased with higher HGI levels (Non-Hispanic white: Q1 = 2423 (75.66%), Q2 = 2299 (72.58%), Q3 = 2115 (69.59%), Q4 = 1541 (57.36%), Non-Hispanic black: Q1 = 594 (6.50%), Q2 = 565 (6.67%), Q3 = 907 (11.57%), Q4 = 1445 (21.44%)); other racial groups exhibited variable trends (p 0.05; Table 1). Marital history analysis indicated that the proportions of divorced and widowed individuals increased with higher HGI levels, while other marital statuses fluctuated irregularly (p 0.05; Table 1). Regarding alcohol consumption, the prevalence of individuals who never drank or were former drinkers was significantly higher in Q4 compared to Q1, whereas heavy drinkers were more prevalent in Q1 than in Q4; other categories showed inconsistent trends (p 0.05; Table 1). The normality assessment demonstrated that the distributions of age (years), BMI, fasting blood–glucose (FBG), HbA1c, ALT, AST, CR, TG, TC, HDL-C, LDL-C, systemic immune-inflammatory index (SII), WBC, LC, NC, PLT, and HGI were not consistent with a normal distribution (Supplementary Table 1). Additionally, AST, CR, TG, and DM prevalence were significantly higher in Q1 and Q4 compared to Q2 and Q3 (p 0.05; Table 1; Supplementary Table 2). Conversely, HDL-C, LDL-C, and SII were significantly higher in Q2 and Q3 compared to Q1 and Q4. Age, BMI, glycated hemoglobin, WBC, NC, and LC were significantly higher in Q4 than in Q1, Q2, and Q3. In contrast, ALT and FBG levels were highest in Q1, while PLT and TC peaked in Q3 (Table 1; Supplementary Table 2).

The prevalence of smoking, hypertension, coronary heart disease (CHD), CKD, CHF, COPD, hyperlipidemia, DM, cancer, and stroke was significantly greater in Q4 compared to other quartiles (p 0.05; Table 1). Regarding study outcomes, both all-cause mortality (primary endpoint) and cardiovascular-related mortality (secondary endpoints including cardiac mortality and cardiovascular mortality) exhibited a U-shaped trend across the HGI quartiles, with the highest rates observed in Q4 (p 0.05; Table 1; Fig. 2; Supplementary Table 2).

3.2 Survival Analysis Based on HGI Quartiles

This study had a median follow-up duration of 112 months. Our analysis revealed significantly higher incidence rates of all-cause mortality in the highest quartile (Q4) compared to Q1, Q2, and Q3 (p 0.05). Similarly, we observed elevated cardiac and cardiovascular mortality rates in the Q4 group compared to lower quartiles (p 0.05; Fig. 3).

3.3 Correlation of the HGI With Outcome Events

Using Cox proportional hazards models with Q2 as the reference category, we examined the association between the HGI and outcome events.

For all-cause mortality, both low HGI (Q1) (hazard ratio (HR) (95% confidence interval (CI)) = 1.269 (1.082, 1.488); p = 0.003) and high HGI (Q4) values (HR (95% CI) = 1.232 (1.065, 1.426); p = 0.005) were associated with increased risks, independent of adjusted risk factors (Table 2). The trend tests did not indicate a significant linear trend (p > 0.05), likely due to both low and high HGI values being risk factors for all-cause mortality.

For cardiac mortality, high HGI (Q4) was a significant risk factor (HR (95% CI) = 1.516 (1.100, 2.088); p = 0.011), while low HGI (Q1) showed no significant association (p > 0.05). The trend tests confirmed that a high HGI value was a risk factor for cardiac mortality (HR (95% CI) = 1.131 (1.029, 1.243); p = 0.010; Table 2).

Regarding cardiovascular mortality, a high HGI (Q4) was significantly associated with increased risk (HR (95% CI) = 1.334 (1.013, 1.756); p = 0.039), whereas a low HGI (Q1) did not demonstrate clinical significance (p > 0.05; Table 2).

3.4 Threshold Effect Analysis of Three Different Follow-up Endpoints

The RCS analyses revealed a U-shaped nonlinear relationship between the HGI and all-cause mortality (p for nonlinearity 0.001), with an inflection point at HGI = –0.25. Similarly, the U-shaped relationships were observed for cardiac mortality (p for nonlinearity 0.001, inflection point = –0.34) and cardiovascular mortality (p for nonlinearity 0.001, inflection point = –0.31) (Table 3; Fig. 4).

3.5 Subgroup Analysis

Subgroup analyses were performed using forest plots to evaluate the robustness of study outcomes and assess whether the HGI differs as a common risk factor in all-cause mortality, cardiac mortality, and cardiovascular mortality. Subgroup analyses for outcome events were conducted based on sex, age, BMI, race, smoking status, history of stroke, hypertension, CHD, and CHF. Our findings revealed that among groups with all-cause mortality as the primary endpoint, except for sex, elevated HGI values were closely associated with higher rates of all-cause mortality in all other subgroups (Fig. 5). In the subsequent Cox regression analysis stratified by gender, both Q1 and Q4 were found to be associated with all-cause mortality in males. In contrast, in females, only Q1 was related to all-cause mortality (Supplementary Table 3). Moreover, in groups where cardiac mortality and cardiovascular mortality were endpoints, except for age, an elevated HGI was consistently associated with adverse outcomes in the general population (Figs. 6,7). In the subsequent Cox analysis stratified by age subgroups, the following associations were observed: For cardiac mortality, Q3 and Q4 were identified as significant risk factors in the 20–39 age group, whereas no significant associations were found across Q1–Q4 in the 40–59 age group. In the 65–85 age group, Q1 and Q4 demonstrated significant associations with cardiac mortality (Supplementary Table 4). Regarding cardiovascular mortality, Q3 and Q4 showed significant associations in the 20–39 age group, while no significant relationships were observed across Q1–Q4 in the 40–59 age group; meanwhile, in the 65–85 age group, Q1 and Q4 were significantly associated with cardiovascular mortality (Supplementary Table 4).

4. Discussion

This study evaluated the associations between the HGI and all-cause mortality, cardiac mortality, and cardiovascular mortality in the general population. The results revealed a U-shaped relationship between the HGI and all-cause mortality, with high and low HGI levels associated with increased mortality rates. However, only high HGI values were significantly associated with increased cardiac and cardiovascular mortalities, while low HGI values showed no significant statistical relevance. This suggests that while both high and low HGI values may independently contribute as risk factors for all-cause mortality in the general population, only high HGI values emerge as an independent risk factor for cardiac mortality and cardiovascular mortality.

The HGI, introduced by Hempe et al. [6] in 2002, offers a novel approach to quantifying discrepancies between individual A1c levels and average blood glucose levels. Unlike FPG and HbA1c, elevated HGI values correlate with increased diabetes risk independent of current blood glucose levels. Previous research has demonstrated the congruence of continuous glucose monitoring (CGM) with the HGI, underscoring its reliability in assessing blood glucose status [12]. A study has also indicated that as populations transition from low to high HGI subgroups, average HbA1c levels rise despite decreasing average FPG levels, suggesting that postprandial glucose fluctuations do not significantly impact variations in the HGI [13]. Thus, high HGI values may signify long-term blood glucose stability. Clinically, discrepancies between HbA1c levels and actual blood–glucose control are common due to glycemic gaps. Consequently, patients with similar mean blood–glucose levels can exhibit substantial differences in HbA1c levels, potentially leading clinicians to overlook these gaps and misjudge therapeutic strategies, ultimately compromising patient treatment [14]. In contrast to HbA1c, utilizing the HGI for patient blood–glucose assessments may offer superior benefits in clinical management, mitigating the risk of therapeutic misjudgments [15]. Considering that HbA1c is currently the widely recognized indicator for blood–glucose control analysis and the challenges of implementing the HGI in clinical practice, future research may need to explore further the similarities and differences between the HGI and HbA1c. Randomized controlled trials guided by HGI thresholds (e.g., implementing differentiated glycemic targets for patients with high HGI values) are also required to develop decision support systems that dynamically integrate HGI, CGM data, and machine learning algorithms to balance blood–glucose control with hypoglycemia risk.

In the two decades since the HGI was proposed, this index has continued to be a subject of active research. The HGI has been established as a predictor of diabetes onset and complications such as diabetic nephropathy and CVD in diabetic patients [16, 17, 18, 19, 20]. Moreover, studies have demonstrated a direct association between elevated HGI values and metabolic syndrome risk in older populations [9, 21]. Higher HGI levels have also proven effective in identifying susceptibility to non-alcoholic fatty liver disease and hypertension among individuals with diabetes [22, 23, 24, 25]. Regarding CVDs, the HGI reliably predicts cardiovascular risk in diabetic populations [26, 27, 28, 29, 30, 31, 32]. Large-scale clinical trials, including the DEVOTE trial, have notably linked elevated HGI levels in type 2 diabetes mellitus (T2DM) patients to increased risks of major adverse cardiovascular events over extended follow-up periods [33]. Similarly, findings from the Ale-Cardio trial highlighted a 16% heightened risk of cardiovascular death per percentage point increase in the HGI [34]. Meta-analyses have further confirmed elevated HGI levels as significantly associated with heightened risks of cardiovascular complications and overall mortality in people with T2DM [20]. Despite these insights, most studies have been cross-sectional and confined to specific diabetic populations, often lacking long-term follow-ups. This limits definitive conclusions about broader prognostic implications of the HGI across the general population. Additionally, associations between the HGI and metabolic and cardiovascular outcomes appear applicable beyond diabetes [35]. A recent study demonstrated a correlation between increased HGI values and telomere attrition. As the HGI levels rose, so did telomere wear, indicating potential impacts on lifespan [28]. These findings suggest that the HGI may be an independent risk factor for all-cause mortality in the general population. Our study findings support this conclusion. In our investigation, individuals with elevated HGI values showed significantly higher inflammatory and metabolic markers, including BMI, WBC, and HbA1c. Moreover, compared to those with lower HGI values, these participants exhibited increased risks of age-related conditions such as hypertension and CVDs. Importantly, the high HGI group also demonstrated markedly elevated risks of all-cause mortality, cardiac mortality, and cardiovascular mortality compared to the low HGI group, aligning with prior research hypotheses.

Recent studies have highlighted that high and low HGI values can influence patient outcomes differently. For instance, a cohort study by Østergaard et al. [36] involving 1910 patients with T2DM suggested that low HGI values might increase the risk of myocardial infarction in patients with CHD. Pan et al. [32] also observed a U-shaped relationship between the HGI value and one-year stroke risk, indicating that low and high HGI levels are associated with adverse cerebrovascular outcomes. A prospective study of 5260 critically ill patients with CHD admitted to the ICU found that both high and low HGI values were significantly linked to negative outcomes at 30 days and 365 days [37]. However, the association between a low HGI and cardiovascular prognosis in the general population remains contentious. This study suggests that the correlation analysis between a low HGI and death from cardiac disease and cardiovascular mortality did not reach statistical significance; however, the p-values for these associations were close to 0.05. This indicates a potential risk that cannot be discounted outright. However, potential biases cannot be neglected, such as selection bias due to missing data, which may not fully represent the entire population. Moreover, the limited types of variables studied might have omitted important risk factors affecting outcomes and risk assessment stability.

The impact of HGI on outcome variables in the subgroup analysis conducted in this study remained consistent across predefined subgroups, except for differences related to sex in all-cause mortality and age in cardiac and cardiovascular mortalities. The observed sex disparity in all-cause mortality may reflect inherent differences in life expectancy between genders. Regarding cardiac mortality and cardiometabolic mortality, individuals aged 20–39 exhibited the highest mortality risk. The better self-care capabilities and perceived notions of being healthy may contribute to younger adults often underestimating their susceptibility to metabolic and CVDs such as hypertension, DM, and CHD. Clinical observations indicate that early-onset cardiovascular and metabolic diseases among younger adults can lead to poorer prognoses compared to older adults. Additionally, the forest plot suggests that patients without stroke, hypertension, coronary heart disease, or CHF but with high HGI values may have a higher mortality rate. This phenomenon is surprising, and to some extent, it is possible that the HGI is more closely associated with mortality in this group of patients. However, due to the retrospective cohort design of this study, study bias may exist. Therefore, further prospective studies may be needed in the future to explore this aspect.

This 20-year cohort study examines the association between the HGI and the risk of all-cause mortality, cardiac mortality, and cardiovascular mortality in the general population. Notably, this study marks the first documentation of the association between HGI values and cardiovascular mortality in the general population. In clinical settings, healthcare providers can assess the risk of cardiovascular mortality based on HGI levels, enabling timely intervention for high-risk patients.

5. Limitations

This study has several limitations. Firstly, the study draws upon public data from the NHANES database, primarily derived from sampled questionnaire surveys of the general population, inherently introducing population data biases. Secondly, inherent relative errors and biases persist. Therefore, future prospective studies involving larger and more diverse populations should explore the association between HGI values and outcomes such as all-cause mortality, cardiac mortality, and cardiometabolic mortality. Lastly, given the many factors influencing mortality in the general population, not all relevant risk factors could be addressed in this study, necessitating further refinement and enhancement of its findings.

6. Conclusion

This study demonstrates that the HGI is associated with all-cause mortality, cardiac mortality, and cardiovascular mortality in the general population. Elevated HGI values emerged as a risk factor for all-cause mortality and cardiovascular mortality in this population, whereas a lower HGI level was linked solely to the general population.

References

[1] Tsao CW, Aday AW, Almarzooq ZI, Anderson CAM, Arora P, Avery CL, et al. Heart Disease and Stroke Statistics-2023 Update: A Report From the American Heart Association. Circulation. 2023; 147: e93–e621. https://doi.org/10.1161/CIR.0000000000001123.

[2] GBD 2021 Forecasting Collaborators. Burden of disease scenarios for 204 countries and territories, 2022-2050: a forecasting analysis for the Global Burden of Disease Study 2021. Lancet (London, England). 2024; 403: 2204–2256. https://doi.org/10.1016/S0140-6736(24)00685-8.

[3] Danaei G, Finucane MM, Lu Y, Singh GM, Cowan MJ, Paciorek CJ, et al. National, regional, and global trends in fasting plasma glucose and diabetes prevalence since 1980: systematic analysis of health examination surveys and epidemiological studies with 370 country-years and 2·7 million participants. Lancet (London, England). 2011; 378: 31–40. https://doi.org/10.1016/S0140-6736(11)60679-X.

[4] 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: 1473–1478. https://doi.org/10.2337/dc08-0545.

[5] 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: 2445–2452. https://doi.org/10.2337/db07-1820.

[6] 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. Journal of Diabetes and its Complications. 2002; 16: 313–320. https://doi.org/10.1016/s1056-8727(01)00227-6.

[7] Soros AA, Chalew SA, McCarter RJ, Shepard R, Hempe JM. Hemoglobin glycation index: a robust measure of hemoglobin A1c bias in pediatric type 1 diabetes patients. Pediatric Diabetes. 2010; 11: 455–461. https://doi.org/10.1111/j.1399-5448.2009.00630.x.

[8] 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. Scientific Reports. 2024; 14: 19629. https://doi.org/10.1038/s41598-024-70666-2.

[9] Zhao L, Li C, Lv H, Zeng C, Peng Y. Association of hemoglobin glycation index with all-cause and cardio-cerebrovascular mortality among people with metabolic syndrome. Frontiers in Endocrinology (Lausanne). 2024; 15: 1447184. https://doi.org/10.3389/fendo.2024.1447184.

[10] Vainshelboim B, Kokkinos P, Myers J. Prognostic Value and Clinical Usefulness of the Hemodynamic Gain Index in Men. The American Journal of Cardiology. 2019; 124: 644–649. https://doi.org/10.1016/j.amjcard.2019.05.015.

[11] Hempe JM, Yang S, Liu S, Hsia DS. Standardizing the haemoglobin glycation index. Endocrinology, Diabetes & Metabolism. 2021; 4: e00299. https://doi.org/10.1002/edm2.299.

[12] Joung HN, Kwon HS, Baek KH, Song KH, Kim MK. Consistency of the Glycation Gap with the Hemoglobin Glycation Index Derived from a Continuous Glucose Monitoring System. Endocrinology and Metabolism (Seoul, Korea). 2020; 35: 377–383. https://doi.org/10.3803/EnM.2020.35.2.377.

[13] Lin L, Wang A, Jia X, Wang H, He Y, Mu Y, et al. High hemoglobin glycation index is associated with increased risk of diabetes: A population-based cohort study in China. Frontiers in Endocrinology. 2023; 14: 1081520. https://doi.org/10.3389/fendo.2023.1081520.

[14] Hsia DS, Rasouli N, Pittas AG, Lary CW, Peters A, Lewis MR, et al. Implications of the Hemoglobin Glycation Index on the Diagnosis of Prediabetes and Diabetes. The Journal of Clinical Endocrinology and Metabolism. 2020; 105: e130–e138. https://doi.org/10.1210/clinem/dgaa029.

[15] Nayak AU, Singh BM, Dunmore SJ. Potential Clinical Error Arising From Use of HbA1c in Diabetes: Effects of the Glycation Gap. Endocrine Reviews. 2019; 40: 988–999. https://doi.org/10.1210/er.2018-00284.

[16] Ibarra-Salce R, Pozos-Varela FJ, Martinez-Zavala N, Lam-Chung CE, Mena-Ureta TS, Janka-Zires M, et al. Correlation Between Hemoglobin Glycation Index Measured by Continuous Glucose Monitoring With Complications in Type 1 Diabetes. Endocrine Practice: Official Journal of the American College of Endocrinology and the American Association of Clinical Endocrinologists. 2023; 29: 162–167. https://doi.org/10.1016/j.eprac.2023.01.001.

[17] Lin CH, Lai YC, Chang TJ, Jiang YD, Chang YC, Chuang LM. Hemoglobin glycation index predicts renal function deterioration in patients with type 2 diabetes and a low risk of chronic kidney disease. Diabetes Research and Clinical Practice. 2022; 186: 109834. https://doi.org/10.1016/j.diabres.2022.109834.

[18] Kim W, Go T, Kang DR, Lee EJ, Huh JH. Hemoglobin glycation index is associated with incident chronic kidney disease in subjects with impaired glucose metabolism: A 10-year longitudinal cohort study. Journal of Diabetes and its Complications. 2021; 35: 107760. https://doi.org/10.1016/j.jdiacomp.2020.107760.

[19] Nakasone Y, Miyakoshi T, Sakuma T, Toda S, Yamada Y, Oguchi T, et al. Hemoglobin Glycation Index: A Novel Risk Factor for Incident Chronic Kidney Disease in an Apparently Healthy Population. The Journal of Clinical Endocrinology and Metabolism. 2024; 109: e1055–e1060. https://doi.org/10.1210/clinem/dgad638.

[20] Zhang L, Wang M, Zhang R, Zhong Y, Fan H, Wang M, et al. Hemoglobin glycation index in relationship to the risk of cardiovascular complication in patients with type 2 diabetes: A systematic review and meta-analysis. Journal of Diabetes and its Complications. 2020; 34: 107673. https://doi.org/10.1016/j.jdiacomp.2020.107673.

[21] Xie SS, Luo XT, Dong MH, Wang Q, Li J, Wu QF. Association Between Hemoglobin Glycation Index and Metabolic Syndrome in Middle-Aged and Older People. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy. 2023; 16: 1471–1479. https://doi.org/10.2147/DMSO.S406660.

[22] Fiorentino TV, Marini MA, Succurro E, Andreozzi F, Sciacqua A, Hribal ML, et al. Association between hemoglobin glycation index and hepatic steatosis in non-diabetic individuals. Diabetes Research and Clinical Practice. 2017; 134: 53–61. https://doi.org/10.1016/j.diabres.2017.09.017.

[23] Wang M, Li S, Zhang X, Li X, Cui J. Association between hemoglobin glycation index and non-alcoholic fatty liver disease in the patients with type 2 diabetes mellitus. Journal of Diabetes Investigation. 2023; 14: 1303–1311. https://doi.org/10.1111/jdi.14066.

[24] Xing Y, Zhen Y, Yang L, Huo L, Ma H. Association between hemoglobin glycation index and non-alcoholic fatty liver disease. Frontiers in Endocrinology. 2023; 14: 1094101. https://doi.org/10.3389/fendo.2023.1094101.

[25] Mi J, Song J, Zhao Y, Wu X. Association of hemoglobin glycation index and its interaction with obesity/family history of hypertension on hypertension risk: a community-based cross-sectional survey. BMC Cardiovascular Disorders. 2020; 20: 477. https://doi.org/10.1186/s12872-020-01762-0.

[26] Guo R, Wang X, Liu Y, Huang M, Ma M, He Y, et al. The Association Between Hemoglobin Glycation Index and Carotid Artery Plaque in Patients With Coronary Heart Disease. Angiology. 2025; 76: 183–192. https://doi.org/10.1177/00033197231198688.

[27] Rhee EJ, Cho JH, Kwon H, Park SE, Park CY, Oh KW, et al. Association Between Coronary Artery Calcification and the Hemoglobin Glycation Index: The Kangbuk Samsung Health Study. The Journal of Clinical Endocrinology and Metabolism. 2017; 102: 4634–4641. https://doi.org/10.1210/jc.2017-01723.

[28] Ahn CH, Min SH, Lee DH, Oh TJ, Kim KM, Moon JH, et al. Hemoglobin Glycation Index Is Associated With Cardiovascular Diseases in People With Impaired Glucose Metabolism. The Journal of Clinical Endocrinology and Metabolism. 2017; 102: 2905–2913. https://doi.org/10.1210/jc.2017-00191.

[29] 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 & Metabolic Syndrome. 2022; 16: 102592. https://doi.org/10.1016/j.dsx.2022.102592.

[30] Shangguan Q, Yang J, Li B, Chen H, Yang L. Association of the hemoglobin glycation index with cardiovascular and all-cause mortality in individuals with hypertension: findings from NHANES 1999-2018. Frontiers in Endocrinology. 2024; 15: 1401317. https://doi.org/10.3389/fendo.2024.1401317.

[31] Wang R, Chen C, Xu G, Jin Z. Association of triglyceride glucose-body mass index and hemoglobin glycation index with heart failure prevalence in hypertensive populations: a study across different glucose metabolism status. Lipids in Health and Disease. 2024; 23: 53. https://doi.org/10.1186/s12944-024-02045-9.

[32] Pan Y, Jing J, Wang Y, Liu L, Wang Y, He Y. Association of hemoglobin glycation index with outcomes of acute ischemic stroke in type 2 diabetic patients. Neurological Research. 2018; 40: 573–580. https://doi.org/10.1080/01616412.2018.1453991.

[33] Klein KR, Franek E, Marso S, Pieber TR, Pratley RE, Gowda A, et al. Hemoglobin glycation index, calculated from a single fasting glucose value, as a prediction tool for severe hypoglycemia and major adverse cardiovascular events in DEVOTE. BMJ Open Diabetes Research & Care. 2021; 9: e002339. https://doi.org/10.1136/bmjdrc-2021-002339.

[34] van Steen SC, Schrieks IC, Hoekstra JB, Lincoff AM, Tardif JC, Mellbin LG, et al. The haemoglobin glycation index as predictor of diabetes-related complications in the AleCardio trial. European Journal of Preventive Cardiology. 2017; 24: 858–866. https://doi.org/10.1177/2047487317692664.

[35] Marini MA, Fiorentino TV, Succurro E, Pedace E, Andreozzi F, Sciacqua A, et al. Association between hemoglobin glycation index with insulin resistance and carotid atherosclerosis in non-diabetic individuals. PloS One. 2017; 12: e0175547. https://doi.org/10.1371/journal.pone.0175547.

[36] Østergaard HB, Mandrup-Poulsen T, Berkelmans GFN, van der Graaf Y, Visseren FLJ, Westerink J, et al. Limited benefit of haemoglobin glycation index as risk factor for cardiovascular disease in type 2 diabetes patients. Diabetes & Metabolism. 2019; 45: 254–260. https://doi.org/10.1016/j.diabet.2018.04.006.

[37] 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. Cardiovascular Diabetology. 2024; 23: 113. https://doi.org/10.1186/s12933-024-02206-1.