Abstract
Background
Heart failure (HF) is closely associated with abnormalities in glucose metabolism, which substantially influence clinical outcomes. The hemoglobin glycation index (HGI), reflecting the discrepancy between observed and predicted glycated hemoglobin, has emerged as a marker of interindividual variation in glycation beyond conventional HbA1c. However, the prognostic significance of HGI in hospitalized patients with HF remains unclear.
Methods
In this retrospective cohort study, we analyzed data from the MIMIC-IV database and included 3,470 adult patients hospitalized with HF. HGI was calculated as the difference between measured HbA1c and predicted HbA1c derived from fasting plasma glucose, and participants were categorized into quartiles. The primary outcome was 365-day all-cause mortality, and the secondary outcome was 30-day all-cause mortality. Kaplan–Meier survival analysis, restricted cubic spline modeling, multivariable Cox regression, and subgroup analyses were performed to evaluate the association between HGI and mortality.
Results
Among the 3,470 included patients, 837 (24.1%) died within 365 days after discharge. Restricted cubic spline analysis demonstrated a predominantly L-shaped association between HGI and both 30-day and 365-day all-cause mortality, with excess mortality risk mainly concentrated at low HGI values. In multivariable Cox models, compared with the lowest quartile, higher HGI quartiles were associated with significantly lower risks of both short-term and long-term mortality, supporting the particular vulnerability of patients with extremely low HGI. Subgroup analyses generally supported the robustness of these findings, although a significant interaction with diabetes status was observed for 365-day mortality.
Conclusions
HGI was independently associated with short-term and long-term all-cause mortality in hospitalized patients with HF, with a predominantly L-shaped pattern observed. Excess mortality risk was mainly concentrated at low HGI values. HGI may serve as a simple and accessible marker for risk stratification in this population.
Keywords: heart failure, hemoglobin glycation index, dysglycemia, all-cause mortality, risk stratification
Introduction
Heart failure (HF) represents a significant global health challenge, impacting more than 56 million individuals worldwide. While the incidence rate appears to be stabilizing, both prevalence and mortality rates remain elevated, imposing a considerable burden on healthcare systems. 1
Evidence indicates that a complex bi-directional relationship exists between HF and diabetes. 2 Diabetes is highly prevalent among patients with HF, affecting over 40% of this population. 3 Furthermore, diabetes is significantly associated with elevated hospitalization rates for HF and poorer prognoses in these patients. 4 As two highly interconnected conditions, the early identification and precise management of diabetes are essential for patients with HF. 5
The substantial benefits of newer anti-diabetic agents, such as sodium-glucose co-transporter-2 inhibitors (SGLT2i), in patients with HF, highlight the critical importance of managing metabolic risk factors, including blood glucose levels, to enhance HF outcomes 6 . Consequently, diabetocardiology may deserve increased attention. 7
Hemoglobin A1c (HbA1c), reflecting average blood glucose levels over the past 2 to 3 months, has been extensively utilized in clinical practice as one of the diagnostic criteria for diabetes. 8 However, an increasing number of studies have identified multiple limitations associated with HbA1c. 9 Research has shown that HbA1c can only reflect approximately 60% to 80% of average blood glucose levels. 10 This is associated with various genetic and environmental factors, including age, race, disease status, medication use, red blood cell lifespan, and differences in the transmembrane glucose gradient.11–15
To address this issue, Hempe et al. developed the Hemoglobin Glycation Index (HGI), which quantifies the relationship between glycosylated hemoglobin and changes in blood glucose. 16 A linear regression equation, utilizing fasting blood glucose and the actual measured glycosylated hemoglobin, is employed to predict glycosylated hemoglobin levels. The discrepancy between the predicted and actual values is referred to as the HGI. 17
Prior research has demonstrated a strong correlation between HGI and severe adverse cardiovascular events, cardiovascular and all-cause mortality risks, and comorbidities in diabetes individuals, as well as adverse prognoses in individuals with metabolic syndrome, hypertension, and coronary heart disease (CHD).18–23 Cheng et al. were the first to establish the relationship between HGI and clinical outcomes in a cohort of 427 patients with acute decompensated heart failure (ADHF). 24 However, the sample size of this study was relatively small and exclusively focused on patients with ADHF. HGI not only provides an accurate reflection of long-term blood glucose control but is also associated with immediate blood glucose levels. It offers a comprehensive assessment of blood glucose management across entire HF patients, rendering it a significant and notable indicator for this population. However, the relationship between HGI and prognosis requires further investigation in a broader spectrum of patients with HF, including those at different stages of disease progression.
This study utilizes the MIMIC database to explore the relationship between HGI and all-cause mortality in the entire hospitalized HF population. It aims to evaluate whether HGI can serve as a universal risk stratification tool applicable to hospitalized HF patients, with the expectation of identifying new methods for recognizing and intervening in high-risk HF groups in order to improve patient prognosis.
Methods
Study participants
This study utilizes data from the Medical Information Mart for Intensive Care (MIMIC-IV) computerized database, which is collaboratively developed and maintained by the Beth Israel Deaconess Medical Center (BIDMC) and the Massachusetts Institute of Technology (MIT). The latest version of the database, MIMIC-IV 3.1, contains relevant information for all patients admitted to BIDMC from 2008 to 2022. Given that the data have been de-identified and transformed, the Institutional Review Board at BIDMC granted a waiver of informed consent and approved the sharing of the research resource. 25 Jia-hao Dou, the author, was responsible for data extraction and processing and fulfilled all database requirements. This study includes adult patients diagnosed with HF during their first hospitalization, based on the International Classification of Diseases, 9th and 10th revisions while excluding patients who lacked blood glucose and glycated hemoglobin data on the day of admission, in addition to those hospital stays shorter than 24 hours. Figure 1 presents the patient screening flowchart.
Figure 1.
Flowchart of patient selection.
Data extraction
Data were extracted from the MIMIC-IV 3.1 database utilizing the pgAdmin PostgreSQL (version 6.1), specifically from five categories: (1) demographics, including age, gender, ethnicity, marriage status, and body mass index (BMI); (2) laboratory examinations, encompassing glucose, HbA1c, red blood cell (RBC), low-density lipoprotein (LDL), creatinine, aspartate aminotransferase (AST), potassium, cardiac troponins T (cTnT), creatine kinase (CK), and N-terminal pro-B-type natriuretic peptide (NT-proBNP); (3) comorbidities, such as CHD, diabetes, hypertension, atrial fibrillation (AF), and hyperlipidemia; (4) medication treatments, which included diuretics, antiplatelet agents, beta-blockers, and antidiabetic medications; and (5) length of hospital stay, follow-up survival situation, and follow-up time for survival (all discharged patients’ one-year follow-up data is included in the database).
All laboratory variables were derived exclusively from the initial 24 hours following patient admission. In instances where multiple results were available, the mean value was employed. To minimize potential bias, variables with missing values greater than 30% were eliminated, while variables with missing values less than 30% were imputed using the random forest imputation technique, which is used in the R software’s missForest package. 26
Definition of exposure variables and outcome events
The HGI is a linear regression residual calculated in two steps. First, a predicted HbA1c level is derived by inputting fasting plasma glucose (FPG) into a regression equation that describes the linear relationship between HbA1c and FPG in patients with HF. Using data from HF patients in this study, a regression equation was established, yielding the following prediction for HbA1c: predicted HbA1c = 0.012 × FPG + 4.781 (r = 0.289, p < 0.001). The relationship between fasting plasma glucose and measured HbA1c used to derive the predicted HbA1c equation is shown in Figure 2. The predicted HbA1c is then subtracted from the individual’s actual measured HbA1c, represented by the formula (HGI = actual measured HbA1c – predicted HbA1c), to obtain the HGI. 11 The main objective of this research was to evaluate 365-day all-cause mortality, with 30-day all-cause mortality serving as a secondary outcome.
Figure 2.
Scatter plot of fasting plasma glucose and measured HbA1c. Scatter plot showing the association between fasting plasma glucose (FPG) and measured HbA1c. The solid line represents the fitted linear regression line, and the shaded area indicates the 95% confidence interval. The regression equation used to calculate predicted HbA1c was: predicted HbA1c = 0.012 × FPG + 4.781 (r = 0.289, P < 0.001). HbA1c, glycated hemoglobin; FPG, fasting plasma glucose.
Statistical analysis
Statistical analyses were performed using t-tests or analysis of variance. Continuous variables were expressed as mean ± standard deviation or median (interquartile range, IQR) based on normality. Categorical variables were represented as proportions and analyzed using chi-square tests. The incidence of outcomes was assessed across various stratification groups based on HGI using Kaplan-Meier (KM) survival analysis. The log-rank test was employed to evaluate any observed differences. The association between HGI and mortality at 30 and 365 days was evaluated using univariate Cox regression analysis. Clinically significant variables were included in the multivariable Cox proportional hazards model with a significance threshold of P < 0.05. Model 1 only had HGI without any additional modifications. Age and gender adjustments were incorporated in Model 2. Model 3 further adjusted for the presence of comorbidities like hypertension, hyperlipidemia, diabetes, CHD, AF, chronic kidney disease (CKD) and chronic obstructive pulmonary disease (COPD), as well as the use of medications like beta-blockers, angiotensin-converting enzyme inhibitor or angiotensin receptor blocker (ACEI/ARB), antiplatelet agents, antidiabetic drugs, and diuretics. The lowest quartile of HGI was designated as the reference category for all models. We employed restricted cubic spline (RCS) analysis to explore the dose-effect relationship between HGI and the primary outcomes. Additionally, we conducted stratified analyses to ensure the robustness of the conclusions. The analysis considered whether variations in gender, age (≤69 years and > 70 years), marriage status and the existence of multiple comorbidities and medication use might influence the results. Data analysis was conducted using R software (version 4.2.2) and SPSS statistical software (version 26.0). A two-tailed P value of less than 0.05 was deemed statistically significant for all analyses.
Results
Baseline characteristics
As shown in Figure 2, measured HbA1c was positively associated with fasting plasma glucose, and the fitted regression equation used to calculate predicted HbA1c was: predicted HbA1c = 0.012 × FPG + 4.781 (r = 0.289, P < 0.001). Based on the quartile grouping of the HGI, Table 1 presents the baseline characteristics of 3,470 HF patients, among whom 837 (24.1%) died during the 365-day follow-up period post-discharge. The average age of the participants was 69 years, with 59.4% being male. Patients in the higher HGI group were generally younger and displayed elevated levels of BMI, RBC, platelets, triglycerides (TG), and hematocrit (HCT), while exhibiting lower levels of high-density lipoprotein (HDL), AST, and NT-proBNP. Compared to the lower HGI group, they also had higher prevalence rates of COPD, CHD, hyperlipidemia, and diabetes, alongside a lower prevalence of AF. In terms of pharmacological treatment, the higher HGI group demonstrated increased utilization of ACEI/ARB, statins, and antidiabetic drugs (all p < 0.05). Furthermore, the prevalence rates of white blood cells (WBC), glucose, CKD, and hypertension were higher in both the high and low HGI groups compared to the median HGI group. Across HGI quartiles, 365-day mortality decreased from Q1 to Q3 and was slightly higher in Q4 (28.2% vs. 22.5% vs. 21.5% vs. 24.2%, P = 0.006). A similar pattern was observed for 30-day mortality (15.0% vs. 10.9% vs. 7.9% vs. 9.7%, P < 0.001), although the excess risk remained most prominent in Q1.
Table 1.
Baseline characteristics of Heart Failure patients grouped according to HGI quartile.
| Categories | Overall (N=3470) | Q1 (N=868) | Q2 (N=871) | Q3 (N=864) | Q4 (N=867) | P-value |
|---|---|---|---|---|---|---|
| Demographic | ||||||
| Age, years, mean (SD) | 69.19 (13.77) | 68.19(14.77) | 70.33(13.19) | 71.06(12.84) | 67.51(12.84) | <0.001 |
| Male,n(%) | 2060(59.4) | 511(58.9) | 509(58.4) | 507(58.7) | 533(61.5) | 0.538 |
| Ethnicity,n(%) | <0.001 | |||||
| Asian | 72 (2.1) | 18 (2.1) | 17 (2.0) | 21 (2.4) | 16 (1.8) | |
| White | 2187 (63.0) | 539 (62.1) | 584 (67.0) | 546 (63.2) | 518 (59.7) | |
| Black | 369 (10.6) | 76 (8.8) | 79 (9.1) | 107 (12.4) | 107 (12.3) | |
| Hispanic/Latino | 120 (3.5) | 19 (2.2) | 21 (2.4) | 38 (4.4) | 42 (4.8) | |
| Other | 722 (20.8) | 216 (24.9) | 170 (19.5) | 152 (17.6) | 184 (21.2) | |
| BMI,kg/m2, median (IQR) | 29.20 (24.70-34.60) | 27.80 (23.85- 33.50) | 27.50 (24.20- 32.60) | 29.30 (25.10- 34.50) | 30.80 (26.45- 36.00) | <0.001 |
| Marital status(%) | 0.005 | |||||
| Married | 1482(42.7) | 366 (42.2) | 382 (43.9) | 375 (43.4) | 359 (41.4) | |
| Single | 808(23.3) | 195 (22.5) | 173 (19.9) | 210 (24.3) | 230 (26.5) | |
| Divorced | 284(8.2) | 62 (7.1) | 69 (7.9) | 67 (7.8) | 86 (9.9) | |
| other | 896(25.8) | 245 (28.2) | 247 (28.4) | 212 (24.5) | 192 (22.1) | |
| Laboratory tests | ||||||
| RBC,m/µL, median (IQR) | 3.95 (3.45- 4.46) | 3.76 (3.19- 4.24) | 3.96 (3.51- 4.47) | 4.03 (3.55- 4.51) | 4.08 (3.56- 4.54) | <0.001 |
| WBC, K/µL, median (IQR) | 8.80 (6.70-11.60) | 9.40 (7.00- 13.10) | 8.80 (6.67- 11.52) | 8.30 (6.50- 10.80) | 8.90 (6.80-11.40) | <0.001 |
| Hemoglobin,g/dL,median (IQR) | 11.7 (10.2- 13.2) | 11.3 (9.6- 13.0) | 11.9 (10.4- 13.5) | 11.7 (10.4- 13.1) | 11.7 (10.1- 13.2) | <0.001 |
| Platelet,K/µL, median (IQR) | 206 (164-259) | 203 (162- 256) | 201 (161, 255) | 202 (163- 255) | 218 (171- 272) | <0.001 |
| Glucose, mg/dL, median (IQR) | 123 (100- 164) | 137.5 (111- 179) | 111 (97- 135) | 108 (94- 137) | 149 (112- 201) | <0.001 |
| HbA1c,%, median (IQR) | 6.0 (5.5- 7.1) | 5.4 (5.1- 5.7) | 5.7 (5.5- 5.9) | 6.2 (5.9- 6.5) | 8.2 (7.2- 9.6) | <0.001 |
| TG, mg/dL, median (IQR) | 97 (73- 139) | 93 (69- 127) | 92.5 (70- 133) | 94 (74- 126) | 120 (85- 169) | <0.001 |
| LDL, mg/dL, median (IQR) | 76 (55- 105) | 79 (54, 108) | 80 (57- 108) | 75 (56- 100) | 75 (54- 101) | 0.265 |
| HDL,mg/dL, median (IQR) | 43 (34- 54) | 46 (35- 59) | 44 (35- 56) | 43 (34- 53) | 40 (33- 49) | <0.001 |
| TC, mg/dL, median (IQR) | 146 (118- 178) | 150 (118- 182) | 148 (119- 182) | 142 (117- 171) | 143 (117- 174) | 0.181 |
| TC/HDL, median (IQR) | 3.30 (2.60- 4.20) | 3.10 (2.50- 4.10) | 3.30 (2.60- 4.10) | 3.27 (2.60- 4.30) | 3.50 (2.90- 4.50) | <0.001 |
| NT-proBNP, pg/mL, median (IQR) | 3594 (1588- 9830) | 4360 (1977- 11865) | 3609 (1420- 10643) | 3054 (1581- 8220) | 2950 (1552- 8800) | 0.196 |
| cTnT, ng/L, median (IQR) | 0.20 (0.05- 0.99) | 0.40 (0.09- 1.65) | 0.20 (0.04- 0.97) | 0.12 (0.03- 0.68) | 0.12 (0.04- 0.71) | <0.001 |
| CK, IU/L, median (IQR) | 133 (66- 338) | 190 (78- 669) | 128 (63- 303) | 112 (55- 251) | 122 (67- 297) | <0.001 |
| Creatinine, mg/dL, median (IQR) | 1.10 (0.90- 1.50) | 1.15 (0.85- 1.70) | 1.05 (0.80- 1.40) | 1.10 (0.90- 1.44) | 1.15 (0.90- 1.65) | <0.001 |
| BUN,mg/dL, median (IQR) | 22 (16- 33) | 22 (16- 35) | 21 (16- 30) | 22 (16- 33) | 22 (17- 36) | <0.001 |
| AST,IU/L, median (IQR) | 27 (19- 46) | 30 (20- 63) | 27 (20- 44) | 25 (18- 38) | 26 (18- 42) | <0.001 |
| ALT,IU/L, median (IQR) | 22 (15- 38) | 23 (15- 43) | 23 (15- 37) | 21 (15- 34) | 22 (15- 38) | 0.371 |
| Albumin,g/dL, median (IQR) | 3.70 (3.30- 4.00) | 3.60 (3.20- 3.90) | 3.80 (3.40- 4.08) | 3.80 (3.40- 4.10) | 3.60 (3.20- 3.90) | <0.001 |
| Sodium,mEq/L, median (IQR) | 139 (137- 141) | 139 (136- 141) | 140 (137- 142) | 140 (137- 142) | 139 (136- 141) | <0.001 |
| Potassium,mEq/L, median (IQR) | 4.10 (3.82- 4.45) | 4.10 (3.80- 4.50) | 4.10 (3.80- 4.40) | 4.10 (3.80- 4.40) | 4.13 (3.90- 4.50) | 0.016 |
| Calcium, mg/dl, median (IQR) | 8.90 (8.45- 9.20) | 8.70 (8.30- 9.10) | 8.90 (8.43- 9.30) | 8.95 (8.60- 9.30) | 8.90 (8.50- 9.25) | <0.001 |
| Magnesium, mg/dl, median (IQR) | 2.00 (1.90- 2.20) | 2.00 (1.90- 2.20) | 2.05 (1.90- 2.20) | 2.00 (1.90- 2.20) | 2.00 (1.80- 2.15) | <0.001 |
| RDW | 14.3 (13.4- 15.6) | 14.2 (13.3- 15.6) | 14.2 (13.4- 15.4) | 14.5 (13.5- 15.6) | 14.3 (13.4- 15.6) | 0.032 |
| Hematocrit | 35.7 (31.5- 40.2) | 34.4 (29.7- 39.1) | 36.0 (32.0- 40.9) | 36.3 (32.0- 40.3) | 36.2 (31.7- 40.4) | <0.001 |
| HGI | -0.23 (-0.70, 0.43) | -1.05 (-1.35, -0.83) | -0.46 (-0.58, -0.35) | 0.03 (-0.12, 0.19) | 1.38 (0.74, 2.57) | <0.001 |
| Comorbidities | ||||||
| Hypertension,n,% | 743 (21.4) | 171 (19.7) | 182 (20.9) | 220 (25.5) | 170 (19.6) | 0.008 |
| Diabetes,n,% | 1,522 (43.9) | 228 (26.3) | 219 (25.1) | 336 (38.9) | 739 (85.2) | <0.001 |
| CHD, n (%) | 1,982 (57.1) | 496 (57.1) | 494 (56.7) | 455 (52.7) | 537 (61.9) | 0.002 |
| AF,n,% | 1,608 (46.3) | 412 (47.5) | 409 (47.0) | 456 (52.8) | 331 (38.2) | <0.001 |
| AMI,n,% | 816 (23.5) | 266 (30.6) | 204 (23.4) | 159 (18.4) | 187 (21.6) | <0.001 |
| COPD, n (%) | 344 (9.9) | 95 (10.9) | 72 (8.3) | 69 (8.0) | 108 (12.5) | 0.004 |
| Hyperlipidemia,n,% | 1,899 (54.7) | 433 (49.9) | 466 (53.5) | 486 (56.3) | 514 (59.3) | <0.001 |
| CKD,n,(%) | 1,088 (31.4) | 279 (32.1) | 240 (27.6) | 243 (28.1) | 326 (37.6) | <0.001 |
| Medicine | ||||||
| Diuretic | 2,770 (79.8) | 697 (80.3) | 679 (78.0) | 693 (80.2) | 701 (80.9) | 0.447 |
| ACEI/ARB | 1,710 (49.3) | 393 (45.3) | 428 (49.1) | 426 (49.3) | 463 (53.4) | 0.009 |
| Beta-blocker | 2,841 (81.9) | 709 (81.7) | 719 (82.5) | 723 (83.7) | 690 (79.6) | 0.154 |
| Antiplatelet | 2,561 (73.8) | 642 (74.0) | 660 (75.8) | 621 (71.9) | 638 (73.6) | 0.328 |
| Statin | 2,559 (73.7) | 599 (69.0) | 648 (74.4) | 619 (71.6) | 693 (79.9) | <0.001 |
| Antidiabetic drug | 2,492 (71.8) | 555 (63.9) | 533 (61.2) | 583 (67.5) | 821 (94.7) | <0.001 |
| Events | ||||||
| 30 days death,n,(%) | 377 (10.9) | 130 (15.0) | 95 (10.9) | 68 (7.9) | 84 (9.7) | <0.001 |
| 365 days death,n,(%) | 837 (24.1) | 245 (28.2) | 196 (22.5) | 186 (21.5) | 210 (24.2) | 0.006 |
HGI Q1 (<-0.70), Q2 (-0.70, -0.23), Q3 (-0.23, 0.43), Q4 (>0.43). HGI hemoglobin glycation index, BMI body mass index, RBC red blood cell, WBC white blood cell, HbA1c hemoglobin A1c, TC total cholesterol, TG triglyceride, LDL low-density lipoprotein, HDL high-density lipoprotein, TC/HDL total cholesterol/high-density lipoprotein ratio, NT-proBNP N-terminal pro-B-type natriuretic peptide, cTnT cardiac troponins T, CK creatine kinase, BUN blood urea nitrogen, ALT alanine aminotransferase, AST aspartate aminotransferase, RDW red blood cell distribution width, CHD coronary heart disease, AMI acute myocardial infarction, AF atrial fibrillation, COPD chronic obstructive pulmonary disease, CKD chronic kidney disease, ACEI/ARB angiotensin-converting enzyme inhibitor or angiotensin receptor blocker.
Survival analysis
Figure 3 illustrates Kaplan–Meier survival curves for 365-day and 30-day all-cause mortality according to HGI quartiles. Patients in the lowest HGI quartile showed the poorest survival, whereas those in the intermediate HGI quartiles generally had more favorable survival. Although the highest HGI quartile showed slightly worse unadjusted survival than the intermediate quartiles, the overall survival pattern indicated that excess mortality risk was most pronounced among patients with low HGI.
Figure 3.
Kaplan–Meier survival analysis curves for all-cause mortality. Panel (a) shows Kaplan–Meier curves for the primary outcome of 365-day all-cause mortality. Panel (b) shows Kaplan–Meier curves for the secondary outcome of 30-day all-cause mortality. HGI quartiles were defined as Q1 (<−0.70), Q2 (−0.70 to −0.23), Q3 (−0.23 to 0.43), and Q4 (>0.43).
Restricted cubic spline analysis
As illustrated in Figure 4, the RCS model revealed a predominantly L-shaped association between HGI and all-cause mortality. The estimated risk was highest at low HGI values and decreased sharply toward the intermediate HGI range. At higher HGI values, the curve showed no statistically robust rebound, as the confidence interval partly crossed 1.0. These findings indicate that the excess mortality risk was mainly concentrated at low HGI values rather than symmetrically distributed at both extremes of HGI.
Figure 4.
Restricted cubic spline Curve for the mortality rate of patients within 365-day (a), and 30-day (b). Restricted cubic spline curves showing the adjusted hazard ratios for 365-day all-cause mortality (a) and 30-day all-cause mortality (b) according to HGI as a continuous variable. The solid line represents the adjusted hazard ratio, and the shaded area represents the 95% confidence interval. The horizontal dashed line indicates a hazard ratio of 1.
As shown in Tables 2 and 3, when HGI was treated as a categorical variable, the multivariable Cox models showed that HGI was associated with both 365-day and 30-day all-cause mortality. Q1 was selected as the reference group to illustrate the trajectory of mortality risk from the lowest HGI range. Compared with Q1, the adjusted HRs were significantly lower in Q2, Q3, and Q4. Although the HR for Q4 was slightly higher than that for Q3 in the fully adjusted model for 365-day mortality, both Q3 and Q4 remained significantly lower than Q1. These findings support a predominantly L-shaped risk pattern, in which mortality risk was highest at extremely low HGI levels and then decreased substantially across higher HGI quartiles.
Table 2.
Cox proportional hazard ratios (HR) for 365-day all-cause mortality.
| Categories | Model 1 | Model 2 | Model 3 | ||||
|---|---|---|---|---|---|---|---|
| HGI | Events(%) | HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value |
| Q1(N=868) | 245 (28.23) | Ref | | Ref | | Ref | |
| Q2(N=871) | 196 (22.50) | 0.76(0.63-0.92) | 0.005 | 0.71(0.59-0.85) | <0.001 | 0.71(0.59-0.86) | <0.001 |
| Q3(N=864) | 186 (21.53) | 0.71(0.59-0.86) | <0.001 | 0.64(0.53-0.77) | <0.001 | 0.64(0.53-0.78) | <0.001 |
| Q4(N=867) | 210 (24.22) | 0.82(0.68-0.98) | 0.032 | 0.84(0.70-1.02) | 0.073 | 0.75(0.61-0.92) | 0.007 |
| P for trend | 0.021 | 0.032 | <0.001 | ||||
HGI hemoglobin glycation index, HR hazard ratio, CI confidence interval. Model 1: Unadjusted model. Model 2: adjusted for gender and age. Model 3: adjusted for gender, age, hypertension, diabetes, CHD, AF, COPD, CKD, hyperlipidemia, beta-blocker, ACEI/ARB, antiplatelet, antidiabetic drug, and diuretic.
Table 3.
Cox proportional hazard ratios (HR) for 30-day all-cause mortality.
| Categories | Model 1 | Model 2 | Model 3 | ||||
|---|---|---|---|---|---|---|---|
| HGI | Events(%) | HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value |
| Q1(N=868) | 130 (14.98) | Ref | | Ref | | Ref | |
| Q2(N=871) | 95 (10.91) | 0.71(0.54-0.92) | 0.010 | 0.65(0.50-0.85) | 0.002 | 0.66(0.50-0.86) | 0.002 |
| Q3(N=864) | 68 (7.87) | 0.50(0.37-0.67) | <0.001 | 0.45(0.34-0.60) | <0.001 | 0.46(0.34-0.61) | <0.001 |
| Q4(N=867) | 84 (9.69) | 0.62(0.47-0.82) | <0.001 | 0.65(0.49-0.85) | 0.002 | 0.58(0.42-0.78) | <0.001 |
| P for trend | <0.001 | <0.001 | <0.001 | ||||
HGI hemoglobin glycation index, HR hazard ratio, CI confidence interval. Model 1: Unadjusted model. Model 2: adjusted for gender and age. Model 3: adjusted for gender, age, hypertension, diabetes, CHD, AF, COPD, CKD, hyperlipidemia, beta-blocker, ACEI/ARB, antiplatelet, antidiabetic drug, and diuretic.
Subgroup analysis
Subgroup analyses showed generally consistent associations between HGI quartiles and mortality across most prespecified strata, including age, sex, marital status, comorbidities, and medication use. A significant interaction was observed between diabetes status and HGI for 365-day mortality (P for interaction = 0.002). No significant interaction was observed for 30-day mortality. Detailed subgroup results are provided in Supplementary Tables S1 and S2.
Discussion
In this retrospective cohort study of 3,470 hospitalized patients with HF from the MIMIC-IV database, we found that HGI was independently associated with both 30-day and 365-day all-cause mortality. The association was predominantly L-shaped, with the excess risk mainly concentrated at low HGI values and the most favorable prognosis observed in the intermediate HGI range. At higher HGI values, mortality risk did not show a statistically robust rebound, suggesting that the observed association should not be interpreted as a classic symmetric U-shaped relationship. These findings extend previous evidence on the prognostic value of HGI in metabolic and cardiovascular disorders and suggest that HGI may provide additional value for risk stratification in hospitalized patients with HF.
HF and diabetes are closely interconnected. The consensus report from the American Diabetes Association notes that HF is an underestimated and increasingly prevalent complication of diabetes. 27 The Framingham Heart Study, the Reykjavik Study, and other studies have shown that patients with diabetes or prediabetes have a substantially higher risk of developing HF than individuals without diabetes.28,29 In addition, diabetes is associated with worse clinical status and poorer outcomes in patients with HF across different ejection fraction phenotypes.30,31 Conversely, HF may also promote the onset and progression of diabetes, further contributing to adverse outcomes. 32 These observations highlight the importance of glycometabolic assessment in patients with HF.
However, HbA1c does not fully reflect ambient glucose exposure in all individuals because it can be influenced by non-glycemic factors, including erythrocyte lifespan, age, ethnicity, renal function, inflammation, medication use, and other disease-related conditions.11,33,34 To address this discrepancy, HGI was introduced as an indicator reflecting the difference between measured HbA1c and glucose-predicted HbA1c. 16 Previous studies have linked HGI to diabetic microvascular complications, cardiovascular disease, cardiovascular mortality, and all-cause mortality in populations with diabetes, hypertension, metabolic syndrome, and coronary heart disease.18–23 Cheng et al. first reported the association between HGI and clinical outcomes in a cohort of patients with acute decompensated HF, but the sample size was relatively limited and the population was restricted to ADHF. 24 Evidence regarding HGI in broader hospitalized HF populations therefore remains limited. The present study utilized data from 3,470 hospitalized HF patients in the MIMIC-IV database to derive an HF-specific HGI equation and to evaluate the nonlinear association between HGI and short-term and long-term mortality in this population.
The RCS analysis further supported a predominantly L-shaped association between HGI and mortality. The estimated risk was highest at low HGI values and declined steeply as HGI increased toward the intermediate range. At higher HGI values, the curve tended to flatten, and the confidence interval partly crossed 1.0, indicating that the estimated risk at higher HGI values did not show a statistically robust increase. Therefore, the excess mortality risk in this cohort appears to be mainly driven by low HGI rather than by equivalent risk elevations at both ends of the HGI distribution.
Several mechanisms may explain the pronounced risk observed in the lowest HGI quartile. HGI is calculated as the difference between measured HbA1c and glucose-predicted HbA1c; therefore, low HGI indicates that HbA1c is lower than expected for a given glucose level. This discrepancy may reflect non-glycemic influences on HbA1c, including shortened erythrocyte survival, anemia, altered erythrocyte turnover, inflammation, renal dysfunction, and acute systemic illness.11,33,34 In the present cohort, patients in Q1 had lower BMI, RBC count, hemoglobin, and hematocrit, together with higher WBC levels and higher mortality. These findings suggest that low HGI may partly capture a high-risk phenotype characterized by altered erythrocyte biology, inflammatory burden, and greater systemic illness burden, rather than simply reflecting favorable long-term glycemic status.35,36
In contrast, the absence of a statistically robust risk rebound at higher HGI values may reflect the heterogeneous clinical characteristics of patients in the highest HGI quartile. Patients in Q4 had higher HbA1c, glucose, BMI, triglycerides, diabetes prevalence, and use of ACEI/ARB, statins, and antidiabetic drugs. These characteristics suggest a phenotype dominated by chronic dysglycemia and cardiometabolic disease. Although hyperglycemia and related metabolic abnormalities can contribute to adverse cardiovascular outcomes through oxidative stress, endothelial dysfunction, inflammation, prothrombotic activation, and myocardial remodeling,37–45 the expected adverse prognostic effect of high HGI may have been attenuated by higher BMI, greater recognition of diabetes or metabolic disease, and more intensive cardiometabolic treatment. However, this explanation remains speculative because confounding by indication cannot be excluded, and because treatment indication, disease severity, diabetes duration, glycemic variability, and post-discharge management could not be fully captured in this retrospective analysis.
The significant interaction between HGI and diabetes status for 365-day mortality should also be interpreted cautiously. Among patients with diabetes, higher HGI may identify individuals with established chronic dysglycemia who were more likely to receive antidiabetic therapy and comprehensive cardiometabolic management. In contrast, among patients without diagnosed diabetes, higher HGI may represent unrecognized dysglycemia or stress-related metabolic disturbance. However, because diabetes duration, treatment intensity, glycemic variability, and post-discharge management were not fully captured in this retrospective analysis, this subgroup finding should be considered exploratory and requires external validation.
Several limitations should be acknowledged. First, this study focused only on baseline HGI at hospital admission. Dynamic changes in glucose, HbA1c, and HGI during hospitalization and after discharge were not available, and their prognostic value warrants further investigation. Second, because this was a retrospective observational study, causality cannot be inferred. Although multivariable adjustment and subgroup analyses were performed, residual confounding remains possible. Third, HGI may be influenced by non-glycemic factors that affect HbA1c, including anemia, erythrocyte lifespan, renal dysfunction, inflammation, and acute illness. Differences in hematological and biochemical characteristics across HGI groups may therefore have contributed to both HGI estimation and mortality risk. Fourth, although the RCS analysis supported a nonlinear association, the risk pattern was predominantly L-shaped, with excess risk mainly concentrated at low HGI values. The lack of a statistically robust risk increase at higher HGI values indicates that the prognostic implication of high HGI should be interpreted cautiously. Finally, this study was conducted using data from a single tertiary medical center in the United States, and external validation in multicenter and prospective cohorts is needed.
Conclusion
In conclusion, HGI was independently associated with both short-term and long-term all-cause mortality in hospitalized patients with HF. The association was predominantly L-shaped, with excess risk mainly concentrated at low HGI values. These findings suggest that HGI may be a simple and clinically accessible marker for identifying high-risk patients at hospital admission. Further prospective multicenter studies are warranted to validate these findings and to determine whether dynamic monitoring of HGI could improve risk assessment and management in patients with HF.
Supplemental material
Supplemental material for Predominantly L-shaped association between hemoglobin glycation index and all-cause mortality in hospitalized patients with heart failure by Rui-yun Wu, Jia-hao Dou, Chen Guo and Jin Wei1 in Diabetes & Vascular Disease Research.
Acknowledgements
We sincerely appreciate the dedicated efforts of the personnel responsible for the design and maintenance of the MIMIV-IV database. We would also like to express our gratitude for the funding support provided by the National Natural Science Foundation of China, the Shaanxi Provincial Science and Technology Department, and the Second Affiliated Hospital of Xi’an Jiaotong University.
Appendix.
List of abbreviations
- ACCORD
Action to Control Cardiovascular Risk in Diabetes
- ACEI/ARB
Angiotensin-converting enzyme inhibitor or angiotensin receptor blocker
- ADHF
Acute decompensated heart failure
- AF
Atrial fibrillation
- AST
Aspartate aminotransferase
- BIDMC
Beth Israel Deaconess Medical Center
- BMI
Body mass index
- CHD
Coronary heart disease
- CK
Creatine kinase
- CKD
Chronic kidney disease
- COPD
Chronic obstructive pulmonary disease
- DCCT
Data from the Diabetes Control and Complications Trial
- FPG
Fasting plasma glucose
- HbA1c
Hemoglobin A1c
- HCT
Hematocrit
- HDL
High-density lipoprotein
- HF
Heart failure
- HR
Hazard ratio
- KM
Kaplan-Meier
- LDL
Low-density lipoprotein
- MACE
Major adverse cardiovascular events
- MIMIC-IV
Medical Information Mart for Intensive Care
- MIT
Massachusetts Institute of Technology
- NT-proBNP
N-terminal pro-B-type natriuretic peptide
- RBC
Red blood cell
- RCS
Restricted cubic spline
- SGLT2i
sodium-glucose co-transporter-2 inhibitors
- TG
Triglyceride
- WBC
White blood cell.
Author contributions: RYW, JHD, CG, and JW were responsible for the study concept, RYW and JHD for the study design. Data extraction was undertaken by JHD. RYW were responsible for data analysis. Drafting of the manuscript: CG. Critical revision of the manuscript for important intellectual content: JW. All authors read and approved the final manuscript.
Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by grants from the National Natural Science Foundation of China (No. 32400952), Shaanxi Provincial Science and Technology Department (No. 2024JC-YBQN-0791, No.2024SF-LCZX-20, and No. 2020ZDLSF02-09), and IIT Clinical Research Fund of The Second Affiliated Hospital of Xi’an Jiaotong University (No. M023).
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Supplemental material: Supplemental material for this article is available online.
ORCID iDs
Rui-yun Wu https://orcid.org/0000-0001-8173-8936
Data Availability Statement
The corresponding author can be contacted to receive the datasets generated and utilized in this work upon reasonable request and with MIMIC’s permission.*
References
- 1.Khan MS, Shahid I, Bennis A, et al. Global epidemiology of heart failure. Nat Rev Cardiol 2024; 21: 717–734. 10.1038/s41569-024-01046-6 [DOI] [PubMed] [Google Scholar]
- 2.Valensi P. Evidence of a bi-directional relationship between heart failure and diabetes: a strategy for the detection of glucose abnormalities and diabetes prevention in patients with heart failure. Cardiovasc Diabetol 2024; 23: 354. 10.1186/s12933-024-02436-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mchugh K, Devore AD, Wu J, et al. Heart Failure With Preserved Ejection Fraction and Diabetes: JACC State-of-the-Art Review. J Am Coll Cardiol 2019; 73: 602–611. 10.1016/j.jacc.2018.11.033 [DOI] [PubMed] [Google Scholar]
- 4.Seferovic PM, Petrie MC, Filippatos GS, et al. Type 2 diabetes mellitus and heart failure: a position statement from the Heart Failure Association of the European Society of Cardiology. Eur J Heart Fail 2018; 20: 853–872. 10.1002/ejhf.1170 [DOI] [PubMed] [Google Scholar]
- 5.Hoek AG, Dal Canto E, Wenker E, et al. Epidemiology of heart failure in diabetes: a disease in disguise. Diabetologia 2024; 67: 574–601. 10.1007/s00125-023-06068-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Usman MS, Bhatt DL, Hameed I, et al. Effect of SGLT2 inhibitors on heart failure outcomes and cardiovascular death across the cardiometabolic disease spectrum: a systematic review and meta-analysis. Lancet Diabetes Endocrinol 2024; 12: 447–461. 10.1016/S2213-8587(24)00102-5 [DOI] [PubMed] [Google Scholar]
- 7.Braunwald E. Diabetocardiology: a new subspecialty? Eur Heart J 2023; 44: 4214–4216. 10.1093/eurheartj/ehad541 [DOI] [PubMed] [Google Scholar]
- 8.2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes-2024. Diabetes Care 2024; 47: S20–S42. 10.2337/dc24-S002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bovee LB, Hirsch IB. Should We Bury HbA1c? Diabetes Technol Ther 2024; 26: 509–513. 10.1089/dia.2024.0028 [DOI] [PubMed] [Google Scholar]
- 10.Nathan DM, Kuenen J, Borg R, et al. Translating the A1C assay into estimated average glucose values. Diabetes Care 2008; 31: 1473–1478. 10.2337/dc08-0545 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hempe JM, Hsia DS. Variation in the hemoglobin glycation index. J Diabetes Complications 2022; 36: 108223. 10.1016/j.jdiacomp.2022.108223 [DOI] [PubMed] [Google Scholar]
- 12.Herman WH. Are There Clinical Implications of Racial Differences in HbA1c? Yes, to Not Consider Can Do Great Harm. Diabetes Care 2016; 39: 1458–1461. 10.2337/dc15-2686 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Pani LN, Korenda L, Meigs JB, et al. Effect of aging on A1C levels in individuals without diabetes: evidence from the Framingham Offspring Study and the National Health and Nutrition Examination Survey 2001-2004. Diabetes Care 2008; 31: 1991–1996. 10.2337/dc08-0577 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.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: 359ra130. 10.1126/scitranslmed.aaf9304 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Khera PK, Joiner CH, Carruthers A, 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. 10.2337/db07-1820 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hempe JM, Gomez R, Mccarter RJ, et al. High and low hemoglobin glycation phenotypes in type 1 diabetes: a challenge for interpretation of glycemic control. J Diabetes Complications 2002; 16: 313–320. 10.1016/s1056-8727(01)00227-6 [DOI] [PubMed] [Google Scholar]
- 17.Klein KR, Franek E, Marso S, 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 Res Care 2021; 9: e002339. 10.1136/bmjdrc-2021-002339 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wang Y, Liu H, Hu X, et al. Association between hemoglobin glycation index and 5-year major adverse cardiovascular events: the REACTION cohort study. Chin Med J (Engl) 2023; 136: 2468–2475. 10.1097/CM9.0000000000002717 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Yang J, Shangguan Q, Xie G, et al. Sex-specific associations between haemoglobin glycation index and the risk of cardiovascular and all-cause mortality in individuals with pre-diabetes and diabetes: A large prospective cohort study. Diabetes Obes Metab 2024; 26: 2275–2283. 10.1111/dom.15541 [DOI] [PubMed] [Google Scholar]
- 20.Carette C, Czernichow S. Harms and benefits of the haemoglobin glycation index (HGI). Eur J Prev Cardiol 2017; 24: 1402–1404. 10.1177/2047487317717821 [DOI] [PubMed] [Google Scholar]
- 21.Wei X, Chen X, Zhang Z, 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: 113. 10.1186/s12933-024-02206-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Shangguan Q, Yang J, Li B, et al. Association of the hemoglobin glycation index with cardiovascular and all-cause mortality in individuals with hypertension: findings from NHANES 1999-2018. Front Endocrinol (Lausanne) 2024; 15: 1401317. 10.3389/fendo.2024.1401317 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhao L, Li C, Lv H, et al. Association of hemoglobin glycation index with all-cause and cardio-cerebrovascular mortality among people with metabolic syndrome. Front Endocrinol (Lausanne) 2024; 15: 1447184. 10.3389/fendo.2024.1447184 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cheng W, Huang R, Pu Y, et al. Association between the haemoglobin glycation index (HGI) and clinical outcomes in patients with acute decompensated heart failure. Ann Med 2024; 56: 2330615. 10.1080/07853890.2024.2330615 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Johnson A, Bulgarelli L, Shen L, et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data 2023; 10: 1. 10.1038/s41597-022-01899-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Stekhoven DJ, Buhlmann P. MissForest--non-parametric missing value imputation for mixed-type data. Bioinformatics 2012; 28: 112–118. 10.1093/bioinformatics/btr597 [DOI] [PubMed] [Google Scholar]
- 27.Pop-Busui R, Januzzi JL, Bruemmer D, et al. Heart Failure: An Underappreciated Complication of Diabetes. A Consensus Report of the American Diabetes Association. Diabetes Care 2022; 45: 1670–1690. 10.2337/dci22-0014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Kannel WB, Hjortland M, Castelli WP. Role of diabetes in congestive heart failure: the Framingham study. Am J Cardiol 1974; 34: 29–34. 10.1016/0002-9149(74)90089-7 [DOI] [PubMed] [Google Scholar]
- 29.Thrainsdottir IS, Aspelund T, Thorgeirsson G, et al. The association between glucose abnormalities and heart failure in the population-based Reykjavik study. Diabetes Care 2005; 28: 612–616. 10.2337/diacare.28.3.612 [DOI] [PubMed] [Google Scholar]
- 30.Jackson AM, Rorth R, Liu J, et al. Diabetes and pre-diabetes in patients with heart failure and preserved ejection fraction. Eur J Heart Fail 2022; 24: 497–509. 10.1002/ejhf.2403 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kristensen SL, Jhund PS, Lee M, et al. Prevalence of Prediabetes and Undiagnosed Diabetes in Patients with HFpEF and HFrEF and Associated Clinical Outcomes. Cardiovasc Drugs Ther 2017; 31: 545–549. 10.1007/s10557-017-6754-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Rossing P, Inzucchi SE, Vart P, et al. Dapagliflozin and new-onset type 2 diabetes in patients with chronic kidney disease or heart failure: pooled analysis of the DAPA-CKD and DAPA-HF trials. Lancet Diabetes Endocrinol 2022; 10: 24–34. 10.1016/S2213-8587(21)00295-3 [DOI] [PubMed] [Google Scholar]
- 33.Campbell L, Pepper T, Shipman K. HbA1c: a review of non-glycaemic variables. J Clin Pathol 2019; 72: 12–19. 10.1136/jclinpath-2017-204755 [DOI] [PubMed] [Google Scholar]
- 34.Nayak AU, Singh BM, Dunmore SJ. Potential Clinical Error Arising From Use of HbA1c in Diabetes: Effects of the Glycation Gap. Endocr Rev 2019; 40: 988–999. 10.1210/er.2018-00284 [DOI] [PubMed] [Google Scholar]
- 35.Hanna A, Frangogiannis NG. Inflammatory Cytokines and Chemokines as Therapeutic Targets in Heart Failure. Cardiovasc Drugs Ther 2020; 34: 849–863. 10.1007/s10557-020-07071-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Murphy SP, Kakkar R, Mccarthy CP, et al. Inflammation in Heart Failure: JACC State-of-the-Art Review. J Am Coll Cardiol 2020; 75: 1324–1340. 10.1016/j.jacc.2020.01.014 [DOI] [PubMed] [Google Scholar]
- 37.Aune D, Schlesinger S, Neuenschwander M, et al. Diabetes mellitus, blood glucose and the risk of heart failure: A systematic review and meta-analysis of prospective studies. Nutr Metab Cardiovasc Dis 2018; 28: 1081–1091. 10.1016/j.numecd.2018.07.005 [DOI] [PubMed] [Google Scholar]
- 38.Krinock MJ, Singhal NS. Diabetes, stroke, and neuroresilience: looking beyond hyperglycemia. Ann N Y Acad Sci 2021; 1495: 78–98. 10.1111/nyas.14583 [DOI] [PubMed] [Google Scholar]
- 39.Luc K, Schramm-Luc A, Guzik TJ, et al. Oxidative stress and inflammatory markers in prediabetes and diabetes. J Physiol Pharmacol 2019; 70: 809–824. 10.26402/jpp.2019.6.01 [DOI] [PubMed] [Google Scholar]
- 40.Paolisso P, Foa A, Bergamaschi L, et al. Hyperglycemia, inflammatory response and infarct size in obstructive acute myocardial infarction and MINOCA. Cardiovasc Diabetol 2021; 20: 33. 10.1186/s12933-021-01222-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Lemkes BA, Hermanides J, Devries JH, et al. Hyperglycemia: a prothrombotic factor? J Thromb Haemost 2010; 8: 1663–1669. 10.1111/j.1538-7836.2010.03910.x [DOI] [PubMed] [Google Scholar]
- 42.Beverly JK, Budoff MJ. Atherosclerosis: Pathophysiology of insulin resistance, hyperglycemia, hyperlipidemia, and inflammation. J Diabetes 2020; 12: 102–104. 10.1111/1753-0407.12970 [DOI] [PubMed] [Google Scholar]
- 43.Zhan J, Chen C, Wang DW, et al. Hyperglycemic memory in diabetic cardiomyopathy. Front Med 2022; 16: 25–38. 10.1007/s11684-021-0881-2 [DOI] [PubMed] [Google Scholar]
- 44.Murtaza G, Virk H, Khalid M, et al. Diabetic cardiomyopathy - A comprehensive updated review. Prog Cardiovasc Dis 2019; 62: 315–326. 10.1016/j.pcad.2019.03.003 [DOI] [PubMed] [Google Scholar]
- 45.Radzioch E, Dabek B, Balcerczyk-Lis M, et al. Diabetic Cardiomyopathy-From Basics through Diagnosis to Treatment. Biomedicines 2024; 12: 765. 10.3390/biomedicines12040765 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental material for Predominantly L-shaped association between hemoglobin glycation index and all-cause mortality in hospitalized patients with heart failure by Rui-yun Wu, Jia-hao Dou, Chen Guo and Jin Wei1 in Diabetes & Vascular Disease Research.
Data Availability Statement
The corresponding author can be contacted to receive the datasets generated and utilized in this work upon reasonable request and with MIMIC’s permission.*