Abstract
Background:
Glycemic abnormalities are highly prevalent in chronic heart failure (CHF) and exacerbate adverse cardiac remodeling. Epicardial fat volume (EFV) and the triglyceride-glucose (TyG) index may both reflect metabolic-cardiovascular interactions, but their combined prognostic utility in CHF patients with dysglycemia remains unexplored.
Methods:
We analyzed 516 CHF patients stratified by glycemic status: normal glucose (n = 230), prediabetes (n = 174), and diabetes (n = 112). Baseline characteristics, echocardiography (TTE), and cardiac magnetic resonance (CMR) data were assessed. The primary endpoint was defined as a composite of heart-failure hospitalization or all-cause mortality and was evaluated during a follow-up period of up to 3 years. Associations between EFV, TyG index, and primary endpoint were evaluated using stratified, joint, and incremental prognostic models.
Results:
Diabetic CHF patients were generally older, had higher smoking rates, worse renal function, and poorer nutritional status. CMR and TTE revealed more severe cardiac structural and functional impairments in dysglycemic groups, including reduced left ventricular ejection fraction (LVEF) and elevated EFV. EFV predicted primary endpoint with increasing accuracy across glycemic strata (AUC: 0.631–0.711 at 1–3 years), with the strongest predictive performance in diabetes. A significant correlation between TyG index and EFV was observed in diabetic patients (r = 0.24, p = 0.00021). Joint EFV (cutoff: 74.68 mL/m²) and TyG (median: 8.48) analysis showed that high TyG alone or combined high EFV/TyG conferred the highest the risk of primary endpoint (p for trend < 0.001), while isolated EFV elevation did not. Incorporating EFV into TyG-based models significantly improved prediction (AUC increased from 0.698 to 0.729 at 1 year). Net reclassification and integrated discrimination improvements confirmed the incremental prognostic value of EFV, particularly in diabetes and over longer follow-up.
Conclusion:
In CHF patients with glycemic dysregulation, EFV and TyG index synergistically enhance clinical endpoint prediction, with EFV providing significant incremental prognostic value beyond TyG. These metrics may improve risk stratification and personalized management in comorbid heart failure and dysglycemia.
Keywords: chronic heart failure, epicardial fat volume, triglyceride-glucose index, glycemic dysregulation, prognostic modeling
1. Introduction
Chronic heart failure (CHF) represents a major and growing public health burden worldwide, characterized by high rates of morbidity, mortality, and recurrent hospitalization [1,2]. The pathophysiology of heart failure (HF) is complex and intertwined with numerous comorbid conditions, among which metabolic dysregulation plays a pivotal role. Abnormal energy substrate utilization, insulin resistance, and chronic inflammatory states contribute significantly to myocardial remodeling, systolic and diastolic dysfunction, and the overall progression of the HF syndrome [3,4].
Glycemic abnormalities, encompassing the spectrum from insulin resistance and prediabetes to overt type 2 diabetes mellitus (T2DM), are exceptionally prevalent in the HF population and are recognized as potent drivers of adverse outcomes [5,6]. Dysglycemia exacerbates HF through multiple mechanisms, including promotion of endothelial dysfunction, accumulation of advanced glycation end-products, and induction of lipotoxicity, which collectively accelerate adverse cardiac remodeling and increase the risk of clinical endpoint [7,8]. Despite the clear association, the optimal tools for risk stratification within this high-risk comorbid population remain limited, necessitating the exploration of novel biomarkers that reflect the underlying metabolic-cardiac interface.
Two promising markers have emerged in this context: epicardial fat volume (EFV) and the triglyceride-glucose (TyG) index. Epicardial adipose tissue, quantified by EFV via cardiac magnetic resonance (CMR) or computed tomography, is a metabolically active organ that directly surrounds the coronary arteries and myocardium. Due to its anatomic proximity and lack of a fascial barrier, it can locally secrete pro-inflammatory cytokines and free fatty acids, directly contributing to coronary atherosclerosis, myocardial fibrosis, and ventricular dysfunction [9,10]. The TyG index is a simple and reliable surrogate marker of insulin resistance [11]. It has been demonstrated to predict new-onset HF and adverse prognosis in cardiovascular diseases [12,13]. Both EFV and the TyG index are elevated in settings of glycemic dysregulation and are independently associated with worse cardiovascular outcomes.
However, their interplay and combined prognostic utility specifically in patients with established HF and comorbid glycemic dysregulation are not well understood. It is plausible that these markers capture complementary pathophysiological pathways of systemic insulin resistance (TyG index) and local paracrine cardiac toxicity (EFV) could synergistically offer superior risk stratification.
Therefore, the primary objective of this study was to investigate the individual and synergistic prognostic value of EFV and the TyG index for predicting endpoint in the cohort of chronic HF patients, stratified by glycemic status. We hypothesized that the combination of high EFV and a high TyG index would identify a subgroup of HF patients with glycemic dysregulation at the highest risk for endpoint and that EFV would provide incremental prognostic value beyond traditional risk factors and the TyG index alone.
2. Methods
2.1 Study Design
This retrospective single-center cohort study enrolled 516 patients with HF who underwent CMR evaluation between January 2018 and December 2020 in the First Affiliated Hospital of Nanjing Medical University (Nanjing, China). Participants with age over 18 years old and less than 85 years were included in our design. The diagnosis of HF were conducted according to the Chinese guideline for diagnosis and treatment of heart failure [14]. The evaluation and treatment of HF were performed in accordance with clinical history, physical assessment, laboratory investigations, cardiac imaging, and functional testing results.
The study protocol was approved by the independent institutional ethics committee of the First Affiliated Hospital of Nanjing Medical University in accordance with the Declaration of Helsinki (Approval No. 2022-SR-052). All participants provided written informed consent prior to enrollment.
2.2 Definition of Covariates
Demographic information, physical examination reports, New York Heart Association (NYHA) functional class, electrocardiographic (ECG) results, laboratory tests, clinical characteristics, comorbidities, and underlying etiology of CHF were extracted from the hospital’s electronic medical record system. All laboratory analyses were performed in the central laboratory of our institution.
We conducted transthoracic echocardiography (TTE) with a Vivid E9 ultrasound system (GE Medical System, USA), and left ventricular ejection fraction (LVEF) was measured following the modified Simpson’s method.
Diabetes was defined as fasting blood glucose (FBG) >126 mg/dL, HbA1c ≥6.5%, medical diagnosis, or use of glucose-lowering medication. While Prediabetes was clarified as FBG between 100 and 125 mg/dL or HbA1c between 5.7% and 6.4%, without previous diagnosed as diabetes or consuming any hypoglycemics agents [15]. Meanwhile, hypertension was characterized as systolic blood pressure (SBP) ≥140 mmHg, diastolic blood pressure (DBP) ≥90 mmHg, antihypertensive medication use, or clinical diagnoses [16].
Especially, the triglyceride-glucose index (TyG), a marker of insulin resistance, was calculated with the formula: TyG = ln[fasting triglycerides (mg/dL) × fasting glucose (mg/dL)/2].
2.3 Measurement of Outcomes
The primary endpoint was a composite outcome of heart-failure hospitalization or all-cause mortality during follow-up. Heart-failure hospitalization was defined as admission requiring treatment for worsening heart failure symptoms. Information on endpoints and patients’ status were obtained via telephone follow-up and/or outpatient visits, with confirmation from the patient’s family or treating physician.
2.4 CMR and Epicardial Fat Quantification (EFV)
Participants underwent CMR imaging in the supine position using a 3T scanner (MAGNETOM Skyra, Siemens Healthcare, Germany) with an 18-channel phased-array body coil anteriorly and a spine coil posteriorly. All acquisitions were collected under ECG-gated breath-hold conditions. The imaging protocol included balanced steady-state free-precession (b-SSFP) cine sequences in three long-axis views (two-, three-, and four-chamber) and a contiguous short-axis stack covering the left ventricle from base to apex (TR/TE 3.4/1.4 ms, flip angle 47°, field of view 360 × 360 mm2, matrix 208 × 188, voxel size 1.6 × 1.6 × 8.0 mm3, slice thickness 8 mm with 2-mm gap, temporal resolution 34 ms). Late gadolinium enhancement (LGE) imaging was performed 10–15 minutes after intravenous administration of gadolinium-DTPA (0.2 mmol/kg, Magnevist, Schering, Berlin, Germany) utilizing a phase-sensitive inversion recovery (PSIR) sequence.
All images were analyzed offline using CVI 42 software (Circle Cardiovascular Imaging Inc., Calgary, Canada) by two independent clinicians, with a third cardiovascular radiologist adjudicating in cases of disagreement. End-diastolic volumes, stroke volumes, LGE, and T1 mapping values were recorded according to guidelines. Extracellular volume (ECV) was calculated from native and post-contrast T1 measurements at the septal mid-ventricular short-axis level.
EFV was measured based on established methods [17,18]. Endocardial and epicardial borders of both ventricles were manually traced on end-diastolic short-axis slices from base to apex. The mitral valve annulus was applied to distinguish atrial from ventricular epicardial adipose tissue. Ventricular epicardial adipose tissue was identified as fat surrounding the ventricles from the mitral valve to the apex. Coronary arteries were manually excluded. EFV was then indexed to body surface area (BSA). Subsequently, in order to confirm that the primary representative value as adipose tissue of measured EFV, T1 mapping was operated at the mid-ventricular level to assess T1 relaxation times of myocardium, epicardial fat, subcutaneous fat, and blood pool. Comparable T1 values between epicardial and subcutaneous fat ensured that the volume measurements reflected adipose tissue rather than fluids.
2.5 Statistical Analyses
Continuous variables were presented as mean ± standard deviation (SD) for normally distributed data or median (interquartile range, IQR) for non-normally distributed data. Categorical variables were expressed as counts (percentages). Group comparisons were performed using Student’s t-test or Mann–Whitney U test for continuous variables and the chi-square (χ2) test for categorical variables.
Internal validation of the epicardial fat volume (EFV) cut-off derived from receiver operating characteristic curves (ROC) was performed using bootstrap resampling (1000 iterations) to assess the stability of the estimated threshold. Kaplan–Meier (KM) curves were applied to evaluate differences in the incidence of the primary endpoint across subgroups. Cox proportional hazards regression models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI). Covariates included in the multivariable models were selected based on clinical relevance and prior evidence regarding prognostic factors in HF. The fully adjusted model included age, gender, SBP, DBP, red blood cell distribution width (RDW), total bilirubin (TBIL), blood urea nitrogen (BUN), estimated glomerular filtration rate (eGFR), uric acid (UA), N-terminal pro-brain natriuretic peptide (NT-proBNP), left atrial diameter (LAD), LVEF, left ventricular end-systolic diameter (LVDs), left ventricular end-diastolic diameter (LVDd), atrial fibrillation (AF), NYHA class.
Collinearity among covariates was evaluated using variance inflation factors (VIF). All variables included in the analyses were complete, and no missing data were observed. Model performance and incremental predictive ability were assessed using time-dependent area under the curve (AUC), C-index, net reclassification improvement (NRI), and integrated discrimination improvement (IDI).
3. Results
3.1 Baseline Characteristics
According to baseline information, participants were categorized into three groups: normal glucose (n = 230), prediabetes (n = 174), and diabetes (n = 112). Baseline demographic and clinical characteristics across different subgroups were summarized in Table 1. Participants diagnosed as diabetes were more likely to be older, smoker and suffered impaired renal function and poorer nutritional status. However, no significant differences were observed in control of blood pressure and lipids. Both TTE and CMR data consistently indicated that patients with glycemic dysregulation exhibited compromised cardiac function and structural abnormalities, particularly reduced LVEF and elevated EFV. Meanwhile, significantly decreased incidence ratio of primary endpoint was observed in normal individuals (p = 0.024).
Table 1.
Baseline characteristics.
| Variables | Normal (n = 230) | Prediabetes (n = 174) | Diabetes (n = 112) | p | |
|---|---|---|---|---|---|
| Age, (years) | 46.45 ± 16.75 | 52.34 ± 14.54 | 56.28 ± 15.36 | <0.001 | |
| BMI, (kg/m²) | 24.01 ± 4.72 | 24.86 ± 4.03 | 25.00 ± 3.55 | 0.054 | |
| SBP, (mmHg) | 121.15 ± 16.08 | 118.91 ± 18.45 | 123.62 ± 17.12 | 0.075 | |
| DBP, (mmHg) | 75.02 ± 12.23 | 75.97 ± 12.69 | 76.12 ± 12.86 | 0.655 | |
| RDW, (%) | 12.99 ± 1.49 | 13.61 ± 1.60 | 13.58 ± 1.79 | <0.001 | |
| ALB, (g/L) | 38.64 ± 3.86 | 37.79 ± 4.05 | 37.24 ± 4.33 | 0.007 | |
| TC, (mmol/L) | 4.04 ± 0.98 | 4.22 ± 1.07 | 4.09 ± 1.20 | 0.220 | |
| TG, (mmol/L) | 1.40 ± 0.94 | 1.40 ± 0.75 | 1.63 ± 1.27 | 0.089 | |
| HDL-c, (mmol/L) | 1.11 ± 0.92 | 1.03 ± 0.27 | 0.98 ± 0.24 | 0.143 | |
| LDL-c, (mmol/L) | 2.45 ± 0.72 | 2.63 ± 0.80 | 2.51 ± 0.87 | 0.069 | |
| TBIL, (μmol/L) | 13.66 ± 6.89 | 14.92 ± 9.02 | 16.92 ± 14.60 | 0.015 | |
| BUN, (mmol/L) | 6.27 ± 3.58 | 6.68 ± 4.68 | 7.10 ± 2.70 | 0.159 | |
| eGFR, (mL/min/1.73 m2) | 103.61 ± 40.96 | 98.34 ± 26.89 | 89.17 ± 30.47 | 0.001 | |
| Glucose, (mmol/L) | 4.50 ± 0.55 | 4.73 ± 0.70 | 6.64 ± 2.82 | <0.001 | |
| HbA1c, (%) | 5.31 ± 0.27 | 5.93 ± 0.22 | 7.36 ± 1.43 | <0.001 | |
| TyG | 8.34 ± 0.60 | 8.45 ± 0.50 | 8.81 ± 0.71 | <0.001 | |
| UA, (μmol/L) | 373.59 ± 118.51 | 429.91 ± 139.89 | 425.42 ± 127.66 | <0.001 | |
| LAD, (mm) | 39.05 ± 7.82 | 43.94 ± 8.12 | 44.19 ± 7.67 | <0.001 | |
| LVDd, (mm) | 54.36 ± 10.30 | 60.70 ± 11.89 | 60.87 ± 11.64 | <0.001 | |
| LVDs, (mm) | 40.28 ± 12.42 | 48.11 ± 13.61 | 48.65 ± 14.32 | <0.001 | |
| LVEF, (%) | 51.72 ± 14.41 | 41.56 ± 14.33 | 40.94 ± 14.53 | <0.001 | |
| SV, (mL) | 78.40 ± 25.82 | 70.10 ± 24.22 | 67.58 ± 23.60 | <0.001 | |
| CO, (L/min) | 5.18 ± 1.70 | 4.88 ± 1.50 | 4.72 ± 1.49 | 0.026 | |
| LVM, (g) | 157.79 ± 60.56 | 181.48 ± 71.27 | 183.16 ± 55.90 | <0.001 | |
| LVMED, (g) | 156.47 ± 60.08 | 180.81 ± 72.07 | 183.89 ± 58.04 | <0.001 | |
| EDV, (mL) | 201.31 ± 99.54 | 253.11 ± 131.59 | 250.77 ± 117.42 | <0.001 | |
| ESV, (mL) | 120.74 ± 95.56 | 183.31 ± 124.97 | 183.12 ± 114.41 | <0.001 | |
| T1, (ms) | 1367.62 ± 105.69 | 1384.04 ± 118.65 | 1364.55 ± 122.02 | 0.258 | |
| ECV, (%) | 34.02 ± 13.38 | 34.46 ± 11.36 | 34.66 ± 11.99 | 0.889 | |
| EFV, (mL/m²) | 52.87 ± 23.52 | 63.27 ± 25.97 | 66.64 ± 26.18 | <0.001 | |
| NT-proBNP, (pg/mL) | 555.40 (173.75, 1387.50) | 1105.00 (369.40, 2738.20) | 1393.00 (355.78, 3287.50) | <0.001 | |
| ALT, (U/L) | 22.05 (15.03, 31.90) | 25.35 (17.33, 45.85) | 26.85 (18.20, 45.20) | <0.001 | |
| AST, (U/L) | 21.85 (18.00, 29.62) | 25.50 (19.60, 35.85) | 26.40 (19.82, 36.78) | <0.001 | |
| D-dimer, (mg/L FEU) | 0.21 (0.12, 0.52) | 0.29 (0.16, 0.69) | 0.41 (0.17, 0.92) | <0.001 | |
| Gender | 0.189 | ||||
| Male | 79 (34.35) | 47 (27.01) | 30 (26.79) | ||
| Female | 151 (65.65) | 127 (72.99) | 82 (73.21) | ||
| Primary endpoint, n (%) | 0.024 | ||||
| No | 183 (79.57) | 120 (68.97) | 77 (68.75) | ||
| Yes | 47 (20.43) | 54 (31.03) | 35 (31.25) | ||
| NYHA, n (%) | <0.001 | ||||
| 1 | 69 (30.00) | 32 (18.39) | 11 (9.82) | ||
| 2 | 93 (40.43) | 55 (31.61) | 44 (39.29) | ||
| 3 | 56 (24.35) | 73 (41.95) | 39 (34.82) | ||
| 4 | 12 (5.22) | 14 (8.05) | 18 (16.07) | ||
| HBP, n (%) | <0.001 | ||||
| No | 170 (73.91) | 114 (65.52) | 57 (50.89) | ||
| Yes | 60 (26.09) | 60 (34.48) | 55 (49.11) | ||
| CAD, n (%) | 0.002 | ||||
| No | 198 (86.09) | 143 (82.18) | 79 (70.54) | ||
| Yes | 32 (13.91) | 31 (17.82) | 33 (29.46) | ||
| AF, n (%) | 0.001 | ||||
| No | 202 (87.83) | 128 (73.56) | 92 (82.14) | ||
| Yes | 28 (12.17) | 46 (26.44) | 20 (17.86) | ||
| Smoke, n (%) | 0.011 | ||||
| No | 176 (76.52) | 112 (64.37) | 72 (64.29) | ||
| Yes | 54 (23.48) | 62 (35.63) | 40 (35.71) | ||
| Drink, n (%) | 0.479 | ||||
| No | 176 (76.52) | 124 (71.26) | 82 (73.21) | ||
| Yes | 54 (23.48) | 50 (28.74) | 30 (26.79) | ||
Abbreviations: BMI, body mass index; NT-proBNP, N-terminal pro–B-type natriuretic peptide; SBP, systolic blood pressure; DBP, diastolic blood pressure; RDW, red cell distribution width; ALB, albumin; TC, total cholesterol; TG, triglycerides; HDL-c, high-density lipoprotein cholesterol; LDL-c, low-density lipoprotein cholesterol; TBIL, total bilirubin; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate; Glucose, fasting blood glucose; HbA1c, glycated hemoglobin; TyG, triglyceride–glucose index; UA, uric acid; D-dimer, D-dimer; LAD, left atrial diameter; LVDd, left ventricular end-diastolic diameter; LVDs, left ventricular end-systolic diameter; LVEF, left ventricular ejection fraction; SV, stroke volume; CO, cardiac output; LVM, left ventricular mass; LVMED, left ventricular mass indexed to end-diastolic volume; EDV, end-diastolic volume; ESV, end-systolic volume; T1, native T1 relaxation time; ECV, extracellular volume fraction; EFV, epicardial fat volume; NYHA, New York Heart Association functional class; HBP, hypertension; CAD, coronary artery disease; AF, atrial fibrillation.
Values are presented as mean ± standard deviation (SD) for normally distributed data, median (interquartile range, IQR) for skewed variables, and number (percentage) for categorical variables.
3.2 Stratified Analyses
Stratified analyses by glycemic status were presented in Fig. 1. Consistent with our previous conclusions among the overall cohort, EFV indicated modest association with the occurrence of primary endpoint in patients with normal glucose metabolism. However, the improved prognostic performance of EFV varied according to glycemic status with AUCs achieving 0.631, 0.648, and 0.683 for patients with prediabetes, AUCs of 0.668, 0.673, and 0.711 for patients with diabetes at 1-, 2-, and 3-year respectively. Similarly, Cox regression of Table 2 suggested only correlation between EFV and primary endpoint among patients with normal glycemia faded into insignificance at different timepoints (HR = 1.00, 95% CI: 0.99–1.02, p = 0.793 for 1-year timepoint; HR = 1.00, 95% CI: 0.99–1.02 for 2-years timepoint, p = 0.661; HR = 0.99, 95% CI: 0.98–1.00, p = 0.453 for 3-years timepoint). Meanwhile, we also observed the strongest positive relation between the TyG index and EFV in patients with diabetes (r = 0.24, p = 0.00021, Fig. 2).
Fig. 1.
Time-dependent ROC analyses for 1-, 2-, and 3-year primary endpoint of epicardial fat volume in different subgroups of heart failure patients. (A) Normal; (B) Prediabetes; (C) Diabetes. Abbreviations: ROC, receiver operating characteristic curve; AUC, area under curve.
Table 2.
Cox regression across subgroups stratified by glycemic status.
| Subgroups | |||
|---|---|---|---|
| Normal (n = 230) | Prediabetes (n = 174) | Diabetes (n = 112) | |
| HR for 1-year (95% CI) | 1.00 (0.99, 1.02) | 1.02 (1.01, 1.03) | 1.03 (1.01, 1.05) |
| p | 0.793 | 0.042 | 0.002 |
| HR for 2-year (95% CI) | 1.00 (0.99, 1.02) | 1.02 (1.01, 1.03) | 1.02 (1.01, 1.03) |
| p | 0.661 | 0.032 | 0.024 |
| HR for 3-year (95% CI) | 0.99 (0.98, 1.00) | 1.01 (1.00, 1.02) | 1.01 (1.00, 1.02) |
| p | 0.453 | 0.036 | 0.012 |
Adjusted by: age, gender, systolic blood pressure, diastolic blood pressure, red blood cell distribution width, total bilirubin, blood urea nitrogen, estimated glomerular filtration rate, uric acid, N-terminal pro-brain natriuretic peptide, left atrial diameter, left ventricular end-diastolic diameter, left ventricular end-systolic diameter, left ventricular ejection fraction, atrial fibrillation, New York Heart Association class, extracellular volume.
Abbreviations: HR, hazard ratio; CI, confidence interval.
Fig. 2.
The association between TyG and epicardial fat volume.
3.3 Joint Analyses of EFV and TyG With Primary Endpoint
Furthermore, we applied ROC to identify the cutoff value of EFV as 74.68 mL/m² with a 95% confidence interval of 46.19–111.54 for predicting primary endpoint among patients with glycemic disorder (n = 286). Combined with calculated median value of TyG as 8.48, we divided the entire cohort as four groups (High/Low EFV with High/Low TyG).
As summarized in Fig. 3 and Table 3, when patients were stratified according to combined EFV and TyG categories, those with elevated TyG alone (Group 1) and those with concomitantly elevated EFV and TyG (Group 3) exhibited a significantly higher risk of primary endpoint compared to the reference group in both unadjusted and multivariable-adjusted models. In contrast, elevated EFV in the absence of high TyG (Group 2) was not independently associated with adverse endpoints (HR = 1.18, 95% CI: 0.69–2.01 for model 1, p = 0.326; HR = 1.47, 95% CI: 0.84–2.58 for model 2, p = 0.239). The overall trend across the four groups was statistically significant in both models (p for trend <0.001), indicating that EFV offered incremental prognostic information beyond TyG among patients with glycemic dysregulation.
Fig. 3.
Kaplan-Meier survival curve of different groups in predicting primary endpoint of heart failure patients with glycemic disorder. G0: patients with EFV < cutoff and TyG < median; G1: patients with EFV < cutoff and TyG ≥ median; G2: patients with EFV ≥ cutoff and TyG < median; G3: patients with EFV ≥ cutoff and TyG ≥ median.
Table 3.
Cox regression of different groups in predicting primary endpoint of heart failure patients with glycemic disorder.
| Models | Groups | p for trend | |||
|---|---|---|---|---|---|
| Group 0 | Group 1 | Group 2 | Group 3 | ||
| Model 1 | Reference | 2.14 (1.14, 4.06) | 1.18 (0.69, 2.01) | 2.10 (1.17, 3.76) | <0.001 |
| Model 2 | Reference | 2.04 (1.06, 3.95) | 1.47 (0.84, 2.58) | 2.09 (1.13, 3.85) | <0.001 |
Model 1: Unadjusted;
Model 2: Adjusted by age, gender, systolic blood pressure, diastolic blood pressure, red blood cell distribution width, total bilirubin, blood urea nitrogen, estimated glomerular filtration rate, uric acid, N-terminal pro-brain natriuretic peptide, left atrial diameter, left ventricular end-diastolic diameter, left ventricular end-systolic diameter, left ventricular ejection fraction, atrial fibrillation, New York Heart Association class.
G0: patients with EFV < cutoff and TyG < median;
G1: patients with EFV < cutoff and TyG ≥ median;
G2: patients with EFV ≥ cutoff and TyG < median;
G3: patients with EFV ≥ cutoff and TyG ≥ median.
3.4 Prognostic Models and Incremental Efficacy
VIF analyses of including variables confirmed the absence of severe collinearity with all VIF values <5 (Table 4), ensuring the stability of constructed models.
Table 4.
Variance inflation factor analyses of selected covariates.
| Variables | VIF |
|---|---|
| Age | 1.669285 |
| Gender | 1.225420 |
| BMI | 1.596397 |
| HBP | 1.305942 |
| eGFR | 1.500354 |
| UA | 1.314859 |
| NT-proBNP | 1.448500 |
| NYHA | 1.640213 |
| AF | 1.172935 |
| CAD | 1.140519 |
| TyG | 1.153706 |
| LVEF | 1.518661 |
| EFV | 1.225672 |
| Smoke | 1.804316 |
| Drink | 1.773341 |
Among dysglycemic individuals, stepwise modifications of the base model by incorporation of TyG and EFV caused incremental improvements in model performance for predicting primary endpoint. Time-dependent ROC analyses demonstrated modest enhancement in the AUC after the addition of TyG, whereas further inclusion of EFV resulted in significant elevated efficiency, particularly at longer follow-up durations. For example, in the overall population, the AUC increased from 0.698 in the base model to 0.703 after inclusion of TyG and to 0.729 after additional adjustment for EFV at 1 year. Similar patterns were observed in individuals with prediabetes and diabetes, with the highest AUC values were constantly achieved by the TyG + EFV model (Fig. 4, Table 5). Time-specific calibration slope analyses demonstrated slopes approaching unity across all time points, particularly at 3 years, indicating stable long-term performance and robust calibration of the combined TyG + EFV model across glycemic subgroups (Table 6). Consistent with the AUC analyses, discrimination assessed by the C-index verified incremental variations following sequential model adjustment (Table 7).
Fig. 4.
Time-dependent ROC analyses for 1-, 2-, and 3-year primary endpoint of epicardial fat volume in different subgroups of heart failure patients.
Table 5.
AUC of different models adjusted by triglyceride–glucose index (TyG) and epicardial fat volume (EFV) for the incidence of primary endpoint at 1-, 2-, 3-year.
| Group | Time | Model | AUC |
|---|---|---|---|
| All | 1 year | Base | 0.698 (0.622–0.774) |
| All | 1 year | Base+TyG | 0.703 (0.628–0.779) |
| All | 1 year | Base+TyG+EFV | 0.729 (0.655–0.803) |
| All | 2 years | Base | 0.689 (0.618–0.760) |
| All | 2 years | Base+TyG | 0.691 (0.620–0.762) |
| All | 2 years | Base+TyG+EFV | 0.726 (0.657–0.795) |
| All | 3 years | Base | 0.685 (0.618–0.751) |
| All | 3 years | Base+TyG | 0.686 (0.620–0.753) |
| All | 3 years | Base+TyG+EFV | 0.746 (0.685–0.808) |
| Prediabetes | 1 year | Base | 0.716 (0.618–0.814) |
| Prediabetes | 1 year | Base+TyG | 0.723 (0.626–0.821) |
| Prediabetes | 1 year | Base+TyG+EFV | 0.734 (0.640–0.829) |
| Prediabetes | 2 years | Base | 0.681 (0.587–0.775) |
| Prediabetes | 2 years | Base+TyG | 0.685 (0.592–0.778) |
| Prediabetes | 2 years | Base+TyG+EFV | 0.702 (0.610–0.794) |
| Prediabetes | 3 years | Base | 0.676 (0.588–0.763) |
| Prediabetes | 3 years | Base+TyG | 0.683 (0.596–0.769) |
| Prediabetes | 3 years | Base+TyG+EFV | 0.724 (0.642–0.806) |
| Diabetes | 1 year | Base | 0.671 (0.546–0.795) |
| Diabetes | 1 year | Base+TyG | 0.672 (0.548–0.797) |
| Diabetes | 1 year | Base+TyG+EFV | 0.719 (0.598–0.840) |
| Diabetes | 2 years | Base | 0.699 (0.587–0.811) |
| Diabetes | 2 years | Base+TyG | 0.699 (0.587–0.810) |
| Diabetes | 2 years | Base+TyG+EFV | 0.762 (0.658–0.866) |
| Diabetes | 3 years | Base | 0.695 (0.589–0.801) |
| Diabetes | 3 years | Base+TyG | 0.689 (0.583–0.796) |
| Diabetes | 3 years | Base+TyG+EFV | 0.782 (0.690–0.875) |
Table 6.
Time-specific calibration slope of models adjusted by TyG and EFV for the incidence of primary endpoint at 1-, 2-, 3-year.
| Group | Time | Slope | Intercept |
|---|---|---|---|
| All | 1 year | 1.321 | –2.127 |
| All | 2 years | 1.132 | –1.602 |
| All | 3 years | 1.107 | –1.347 |
| Prediabetes | 1 year | 1.384 | –2.102 |
| Prediabetes | 2 years | 1.027 | –1.509 |
| Prediabetes | 3 years | 1.01 | –1.234 |
| Diabetes | 1 year | 1.247 | –2.178 |
| Diabetes | 2 years | 1.313 | –1.778 |
| Diabetes | 3 years | 1.291 | –1.572 |
Table 7.
C-index of different models adjusted by TyG and EFV for the incidence of primary endpoint.
| Group | Model | Cindex |
|---|---|---|
| All | Base | 0.673 (0.613–0.733) |
| All | Base+TyG | 0.676 (0.616–0.737) |
| All | Base+TyG+EFV | 0.684 (0.624–0.745) |
| Prediabetes | Base | 0.676 (0.598–0.754) |
| Prediabetes | Base+TyG | 0.681 (0.604–0.758) |
| Prediabetes | Base+TyG+EFV | 0.694 (0.621–0.768) |
| Diabetes | Base | 0.684 (0.585–0.784) |
| Diabetes | Base+TyG | 0.717 (0.626–0.809) |
| Diabetes | Base+TyG+EFV | 0.757 (0.676–0.838) |
NRI and IDI analyses provided complementary evidence for the incremental prognostic value of EFV. Integration of EFV into the TyG based model yielded further and substantial reclassification gains, particularly at longer follow-up periods and among participants with diabetes (Fig. 5, Table 8). Moreover, IDI analyses delivered that inclusion of EFV prompted the model discrimination especially in prediabetes and diabetes subgroups, while the addition of TyG alone was associated with relatively smaller changes in discrimination (Fig. 5, Table 9).
Fig. 5.
Incremental reclassification and discrimination improvement after adding EFV to TyG-based prognostic models. (A) NRI of different models; (B) IDI of different models. Abbreviations: NRI, net reclassification improvement index; IDI, integrated discrimination improvement.
Table 8.
NRI of different models adjusted by TyG and EFV for the incidence of primary endpoint at 1-, 2-, 3-year.
| Group | Time | Comparison | NRI_total | NRI_Lower | NRI_Upper | p |
|---|---|---|---|---|---|---|
| All | 1 year | Base → TyG | 0.0348 | –0.0196 | 0.1013 | 0.252 |
| All | 1 year | TyG → TyG+EFV | 0.0848 | –0.0595 | 0.2495 | 0.274 |
| All | 2 years | Base → TyG | 0.0462 | –0.0046 | 0.1052 | 0.084 |
| All | 2 years | TyG → TyG+EFV | 0.1660 | 0.0018 | 0.3366 | 0.039 |
| All | 3 years | Base → TyG | 0.0684 | –0.0803 | 0.2148 | 0.402 |
| All | 3 years | TyG → TyG+EFV | 0.1133 | 0.0192 | 0.3139 | 0.033 |
| Prediabetes | 1 year | Base → TyG | 0.0877 | –0.0420 | 0.1597 | 0.518 |
| Prediabetes | 1 year | TyG → TyG+EFV | 0.1279 | –0.0666 | 0.3096 | 0.214 |
| Prediabetes | 2 years | Base → TyG | 0.0527 | –0.0084 | 0.1419 | 0.637 |
| Prediabetes | 2 years | TyG → TyG+EFV | 0.1626 | 0.0641 | 0.2671 | 0.020 |
| Prediabetes | 3 years | Base → TyG | 0.1055 | –0.0577 | 0.2682 | 0.825 |
| Prediabetes | 3 years | TyG → TyG+EFV | 0.1915 | 0.0556 | 0.2828 | 0.011 |
| Diabetes | 1 year | Base → TyG | 0.0011 | –0.0429 | 0.0345 | 0.768 |
| Diabetes | 1 year | TyG → TyG+EFV | 0.0993 | –0.0882 | 0.2889 | 0.191 |
| Diabetes | 2 years | Base → TyG | 0.0342 | –0.0092 | 0.0631 | 0.567 |
| Diabetes | 2 years | TyG → TyG+EFV | 0.2499 | 0.1079 | 0.4076 | 0.005 |
| Diabetes | 3 years | Base → TyG | 0.1571 | 0.0394 | 0.2914 | 0.040 |
| Diabetes | 3 years | TyG → TyG+EFV | 0.3643 | 0.0811 | 0.6863 | 0.014 |
Base model: N-terminal pro–B-type natriuretic peptide+systolic blood pressure+diastolic blood pressure+estimated glomerular filtration rate+uric acid+left ventricular ejection fraction+New York Heart Association functional class+hypertension+coronary artery disease+atrial fibrillation.
Table 9.
IDI of different models adjusted by TyG and EFV for the incidence of primary endpoint at 1-, 2-, 3-year.
| Group | Time | Comparison | IDI_total | IDI_Lower | IDI_Upper | p |
|---|---|---|---|---|---|---|
| All | 1 year | Base → TyG | 0.0024 | –0.0015 | 0.0063 | 0.229 |
| All | 1 year | TyG → TyG+EFV | 0.0382 | 0.0130 | 0.0633 | 0.003 |
| All | 2 years | Base → TyG | 0.0002 | –0.0035 | 0.0053 | 0.692 |
| All | 2 years | TyG → TyG+EFV | 0.0759 | 0.0381 | 0.1138 | 0.024 |
| All | 3 years | Base → TyG | 0.0027 | –0.0041 | 0.0095 | 0.441 |
| All | 3 years | TyG → TyG+EFV | 0.1344 | 0.0688 | 0.2000 | <0.001 |
| Prediabetes | 1 year | Base → TyG | 0.0081 | –0.0052 | 0.0214 | 0.233 |
| Prediabetes | 1 year | TyG → TyG+EFV | 0.0320 | –0.0004 | 0.0643 | 0.053 |
| Prediabetes | 2 years | Base → TyG | 0.0005 | –0.0141 | 0.0151 | 0.947 |
| Prediabetes | 2 years | TyG → TyG+EFV | 0.0535 | 0.0092 | 0.0978 | 0.018 |
| Prediabetes | 3 years | Base → TyG | 0.0112 | –0.0105 | 0.0330 | 0.312 |
| Prediabetes | 3 years | TyG → TyG+EFV | 0.0555 | 0.0146 | 0.1256 | 0.021 |
| Diabetes | 1 year | Base → TyG | –0.0021 | –0.0077 | 0.0036 | 0.478 |
| Diabetes | 1 year | TyG → TyG+EFV | 0.0434 | –0.0018 | 0.0886 | 0.060 |
| Diabetes | 2 years | Base → TyG | –0.0011 | –0.0065 | 0.0043 | 0.754 |
| Diabetes | 2 years | TyG → TyG+EFV | 0.0897 | 0.0163 | 0.1631 | 0.017 |
| Diabetes | 3 years | Base → TyG | 0.0013 | –0.0049 | 0.0074 | 0.688 |
| Diabetes | 3 years | TyG → TyG+EFV | 0.1911 | 0.0940 | 0.2883 | <0.001 |
Base model: N-terminal pro–B-type natriuretic peptide+systolic blood pressure+diastolic blood pressure+estimated glomerular filtration rate+uric acid+left ventricular ejection fraction+New York Heart Association functional class+hypertension+coronary artery disease+atrial fibrillation.
4. Discussion
The principal finding of this study is that epicardial fat volume (EFV) and the triglyceride-glucose (TyG) index act as synergistic and incremental prognostic markers for primary endpoint in patients with CHF, particularly in the presence of glycemic dysregulation. To be noted, the prognostic power of EFV is gradient-dependent across the spectrum from normoglycemia to diabetes, being strongest in diabetes and significant in prediabetes. While a high TyG index alone significantly elevates the risk of primary endpoint, the highest risk is observed when both EFV and TyG are elevated. Also, EFV provides substantial incremental prognostic value beyond established risk factors and the TyG index, significantly improving model discrimination and reclassification.
Our results are consistent with and extend the current literature. Previous investigations have independently linked both EFV and the TyG index to adverse outcomes in various cardiovascular diseases [19,20]. However, our study is among the first to integrate them specifically in a HF cohort stratified by glycemic status. A crucial and novel aspect of our work is the highlighting of prediabetes as a critical risk stratum. Prediabetes is often under-managed in clinical practice [21,22]. Our data showed that patients with HF and prediabetes exhibit a clear intermediate phenotype, both in terms of baseline characteristics and prognostic marker performance, lying between normoglycemia and diabetes. The predictive accuracy of EFV in prediabetes was markedly stronger than in normoglycemic patients and approached the level seen in diabetes, indicating that the pathophysiological damage leading to adverse remodeling begins early in the dysglycemic continuum. This underscores the importance of early identification and intensified management in HF patients with even mild glycemic abnormalities, a population that may otherwise be overlooked [23,24,25].
Such synergistic relationship between EFV and the TyG index can be explained by their roles in the vicious cycle of metabolic-cardiac dysfunction. The TyG index, as a marker of systemic insulin resistance, promotes lipolysis and increases the flux of free fatty acids to ectopic fat depots, including the epicardium, thereby driving EFV expansion [26,27]. In the context of HF, this is particularly deleterious. A high EFV signifies a local source of pro-inflammatory cytokines (e.g., TNF-α, IL-6), adipokines, and free fatty acids, which can directly infiltrate the adjacent myocardium through paracrine and vasocrine mechanisms [28]. Notably, the adipokine profile from epicardial fat is often dysregulated in metabolic dysfunction—with increased secretion of pro-inflammatory adipokines (e.g., resistin, leptin) and reduced production of anti-inflammatory adipokines (e.g., adiponectin). This imbalance further perpetuates insulin resistance, promotes oxidative stress, and directly impairs myocardial contractility and energy metabolism, amplifying the local cardiac insult. Together with pro-inflammatory cytokines and free fatty acids, these adipokines exacerbate myocardial fibrosis, impair calcium handling, and promote endothelial dysfunction, leading to progressive ventricular remodeling and arrhythmogenesis [29,30,31]. Consequently, in diabetic HF patients, we observe the strongest correlation between TyG and EFV and the highest predictive power for primary endpoint, as both systemic and local metabolic insults are maximally activated.
The combined assessment of EFV and the TyG index can significantly improve risk prediction in heart failure patients, especially those with dysglycemia. This can help identify high-risk individuals who may benefit from more aggressive monitoring and therapy. Furthermore, EFV may represent a novel therapeutic target. Interventions aimed at reducing ectopic fat deposition, such as SGLT2 inhibitors or GLP-1 receptor agonists may exert part of their benefit by modulating epicardial fat [32]. Finally, clinicians should be vigilant about metabolic health in heart failure patients with prediabetes, as this group is already at significantly elevated risk.
This study demonstrated a powerful synergy between a marker of local cardiac metabolic burden (EFV) and a marker of systemic insulin resistance (TyG index) in predicting adverse outcomes in chronic HF. The risk is gradient-dependent on glycemic status, with a pronounced effect already evident in the often-neglected prediabetes population. The pathophysiological interplay between these markers creates a perfect storm that accelerates HF progression. The incorporation of EFV significantly refines risk prediction models beyond current standards.
5. Limitations
Several limitations of this study should be acknowledged. First, as a retrospective, single-center observational study, this study could not establish a definitive causal link between elevated EFV and endpoint. As a surrogate marker, EFV might reflect advanced metabolic exhaustion rather than a primary driver of adverse outcomes. Second, the diabetes subgroup (n = 112) was relatively moderate in size while the overall cohort was substantial. Meanwhile, the exclusive inclusion of patients who underwent CMR imaging might introduce selection bias. And although internal validation of the EFV cut-off was performed based on bootstrap resampling, the lack of external validation remained an important limitation of the present study. Baseline characteristics of the study population and inclusion criteria might introduce bias in the derived EFV cutoff value. Based on the research directions proposed in this study, findings regarding the EFV cutoff value require validation in larger, multicenter prospective cohorts to ensure broader generalizability, including subgroup analyses according to key clinical factors such as age and LVEF. Future studies will also aim to externally validate the EFV-TyG prognostic framework using independent datasets across different institutions, imaging protocols, and patient populations. Moreover, although we adjusted for numerous confounders, residual confounding might persist, particularly regarding clinical therapies. Finally, we recognized that EFV quantification based on CMR is relatively resource-intensive. The requirement for advanced imaging and specialized analyses will limit the immediate, widespread adoption of this framework in clinical settings with constrained resources.
6. Conclusion
Among patients with chronic heart failure, the combination of elevated epicardial fat volume and a high triglyceride-glucose index powerfully identifies those with glycemic dysregulation—including prediabetes and diabetes—who are at the greatest risk for clinical events. These findings provide a conceptual framework for integrated metabolic-cardiovascular risk stratification. Future research should focus on validating these markers in larger populations and exploring more cost-effective imaging alternatives to facilitate targeted metabolic interventions in this vulnerable population.
Acknowledgment
We sincerely appreciate all patients who participated in this study and contributed their clinical information for research purposes.
Funding Statement
This work was supported by General Program of National Natural Science Foundation of China (82370389, 81970339 to XLZ), the project "Treatment of Coronary Heart Disease and Heart Failure with Representative Formulas of the QLQX Theory" (303100036AA19 to XL), Project from Gusu School (GSRCKY20210204 to HZ), and Gusu Health Personnel Training Project (GSWS2021042 to HZ).
Footnotes
Publisher’s Note: IMR Press stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Iokfai Cheang, Email: steven0510@sina.com.
Xinli Li, Email: xinli3267@njmu.edu.cn.
Availability of Data and Materials
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Prof. Xinli Li (xinli3267@njmu.edu.cn).
Author Contributions
ZC: Formal analysis, Software, Writing original draft, Review & Editing. YY: Conceptualization, Methodology, Writing original draft. HC: Validation, Investigation. QL: Resources, Data Curation. HZ: Conception, Funding, Review & Editing. IC: Formal analysis, Methodology, Review & Editing. XL: Conceptualization, Methodology, Supervision, Funding acquisition. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.
Ethics Approval and Consent to Participate
Study protocol was approved by the independent institutional ethical committee of the First Affiliated Hospital of Nanjing Medical University and conducted in accordance with the Declaration of Helsinki (Approval No. 2022-SR-052). All participants provided written informed consent.
Funding
This work was supported by General Program of National Natural Science Foundation of China (82370389, 81970339 to XLZ), the project "Treatment of Coronary Heart Disease and Heart Failure with Representative Formulas of the QLQX Theory" (303100036AA19 to XL), Project from Gusu School (GSRCKY20210204 to HZ), and Gusu Health Personnel Training Project (GSWS2021042 to HZ).
Conflicts of Interest
The authors declare no conflicts interests.
References
- [1].Yang X, Zhao ZP, Shi Y, Han GY, Xu Y, Li YC, et al. The evolving burden of heart failure in China: a 34-year subnational analysis of trends and causes from the Global Burden of Disease Study 2023. Military Medical Research. 2025;12:65. doi: 10.1186/s40779-025-00650-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Savarese G, Becher PM, Lund LH, Seferovic P, Rosano GMC, Coats AJS. Global burden of heart failure: a comprehensive and updated review of epidemiology. Cardiovascular Research. 2023;118:3272–3287. doi: 10.1093/cvr/cvac013. [DOI] [PubMed] [Google Scholar]
- [3].Fang Z, Raza U, Song J, Lu J, Yao S, Liu X, et al. Systemic aging fuels heart failure: Molecular mechanisms and therapeutic avenues. ESC Heart Failure. 2025;12:1059–1080. doi: 10.1002/ehf2.14947. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Normand C, Kaye DM, Povsic TJ, Dickstein K. Beyond pharmacological treatment: an insight into therapies that target specific aspects of heart failure pathophysiology. Lancet (London, England) 2019;393:1045–1055. doi: 10.1016/S0140-6736(18)32216-5. [DOI] [PubMed] [Google Scholar]
- [5].Wong ND, Sattar N. Cardiovascular risk in diabetes mellitus: epidemiology, assessment and prevention. Nature Reviews. Cardiology. 2023;20:685–695. doi: 10.1038/s41569-023-00877-z. [DOI] [PubMed] [Google Scholar]
- [6].Gonzalez-Manzanares R, Anguita-Gámez M, Muñiz J, Barrios V, Gimeno-Orna JA, Pérez A, et al. Prevalence and incidence of heart failure in type 2 diabetes patients: results from a nationwide prospective cohort-the DIABET-IC study. Cardiovascular Diabetology. 2024;23:253. doi: 10.1186/s12933-024-02358-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Kim JA, Montagnani M, Chandrasekran S, Quon MJ. Role of lipotoxicity in endothelial dysfunction. Heart Failure Clinics. 2012;8:589–607. doi: 10.1016/j.hfc.2012.06.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Giraldo-Gonzalez GC, Roman-Gonzalez A, Cañas F, Garcia A. Molecular Mechanisms of Type 2 Diabetes-Related Heart Disease and Therapeutic Insights. International Journal of Molecular Sciences. 2025;26:4548. doi: 10.3390/ijms26104548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Packer M. Epicardial Adipose Tissue May Mediate Deleterious Effects of Obesity and Inflammation on the Myocardium. Journal of the American College of Cardiology. 2018;71:2360–2372. doi: 10.1016/j.jacc.2018.03.509. [DOI] [PubMed] [Google Scholar]
- [10].Wang W, Tian Y, Wang W, Yin H, Yin D, Tian Y. Role of epicardial adipose tissue in triggering and maintaining atrial fibrillation. Cardiovascular Innovations and Applications. 2022;7:997. doi: 10.15212/CVIA.2022.0012. [DOI] [Google Scholar]
- [11].Sun Y, Ji H, Sun W, An X, Lian F. Triglyceride glucose (TyG) index: A promising biomarker for diagnosis and treatment of different diseases. European Journal of Internal Medicine. 2025;131:3–14. doi: 10.1016/j.ejim.2024.08.026. [DOI] [PubMed] [Google Scholar]
- [12].Dou J, Guo C, Wang Y, Peng Z, Wu R, Li Q, et al. Association between triglyceride glucose-body mass and one-year all-cause mortality of patients with heart failure: a retrospective study utilizing the MIMIC-IV database. Cardiovascular Diabetology. 2023;22:309. doi: 10.1186/s12933-023-02047-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Li X, Chan JSK, Guan B, Peng S, Wu X, Lu X, et al. Triglyceride-glucose index and the risk of heart failure: Evidence from two large cohorts and a mendelian randomization analysis. Cardiovascular Diabetology. 2022;21:229. doi: 10.1186/s12933-022-01658-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Heart Failure Group of Chinese Society of Cardiology of Chinese Medical Association. Chinese Heart Failure Association of Chinese Medical Doctor Association. Editorial Board of Chinese Journal of Cardiology [Chinese guidelines for the diagnosis and treatment of heart failure 2018] Zhonghua Xin Xue Guan Bing Za Zhi. 2018;46:760–789. doi: 10.3760/cma.j.issn.0253-3758.2018.10.004. (In Chinese) [DOI] [PubMed] [Google Scholar]
- [15].ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, et al. 2. Classification and Diagnosis of Diabetes: Standards of Care in Diabetes-2023. Diabetes Care. 2023;46:S19–S40. doi: 10.2337/dc23-S002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Unger T, Borghi C, Charchar F, Khan NA, Poulter NR, Prabhakaran D, et al. 2020 International Society of Hypertension global hypertension practice guidelines. Journal of Hypertension. 2020;38:982–1004. doi: 10.1097/HJH.0000000000002453. [DOI] [PubMed] [Google Scholar]
- [17].Cheang I, Zhu X, Yue X, Tang Y, Gao Y, Lu X, et al. Prognostic value of ventricle epicardial fat volume by cardiovascular magnetic resonance in chronic heart failure. iscience. 2023;26:106755. doi: 10.1016/j.isci.2023.106755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Khidr SS, Bakeer BM, Helmy HAR, El-Naggar HM. Cardiac magnetic resonance quantified epicardial fat volume is associated with complex coronary artery disease among diabetics. Cardiovascular Diabetology. 2025;24:64. doi: 10.1186/s12933-025-02606-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Yang L, Yu W, Wan P, Wang J, Shao X, Zhang F, et al. Epicardial fat volume, an independent risk factor for major adverse cardiovascular events, had an incremental prognostic value to myocardial perfusion imaging in Chinese populations with suspected or known coronary artery disease with a normal left ventricular ejection fraction. Frontiers in Cardiovascular Medicine. 2023;10:1261215. doi: 10.3389/fcvm.2023.1261215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Li C, Zhao S, Li Y, He S, Jian W, Liu Y, et al. Triglyceride-glucose index predicts adverse cardiovascular events in patients with H-type hypertension combined with coronary heart disease: a retrospective cohort study. Cardiovascular Diabetology. 2025;24:45. doi: 10.1186/s12933-024-02555-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Tabák AG, Herder C, Rathmann W, Brunner EJ, Kivimäki M. Prediabetes: a high-risk state for diabetes development. Lancet (London, England) 2012;379:2279–2290. doi: 10.1016/S0140-6736(12)60283-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Ping WX, Hu S, Su JQ, Ouyang SY. Metabolic disorders in prediabetes: From mechanisms to therapeutic management. World Journal of Diabetes. 2024;15:361–377. doi: 10.4239/wjd.v15.i3.361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Cheang I, Zhu X, Huang JY, Tse YK, Li HL, Ren QW, et al. Prediabetes is associated with increased cardiac events in patients with cancer who are prescribed anthracyclines. Cancer. 2024;130:2795–2806. doi: 10.1002/cncr.35322. [DOI] [PubMed] [Google Scholar]
- [24].Davoodian N, Lotfaliany M, Huxley RR, Lee CMY, Pasco JA, Adams RJ, et al. Regression from prediabetes to normoglycaemia and the role of cardiometabolic risk factors on the subsequent risk of developing type 2 diabetes. Diabetologia. 2026;69:69–80. doi: 10.1007/s00125-025-06555-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Dinh W, Lankisch M, Nickl W, Scheyer D, Scheffold T, Kramer F, et al. Insulin resistance and glycemic abnormalities are associated with deterioration of left ventricular diastolic function: a cross-sectional study. Cardiovascular Diabetology. 2010;9:63. doi: 10.1186/1475-2840-9-63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Wu S, Tan J, Zhang H, Hou DX, He J. Tissue-specific mechanisms of fat metabolism that focus on insulin actions. Journal of Advanced Research. 2023;53:187–198. doi: 10.1016/j.jare.2022.12.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Iacobellis G. Local and systemic effects of the multifaceted epicardial adipose tissue depot. Nature Reviews. Endocrinology. 2015;11:363–371. doi: 10.1038/nrendo.2015.58. [DOI] [PubMed] [Google Scholar]
- [28].Verma S, McMurray JJV. SGLT2 inhibitors and mechanisms of cardiovascular benefit: a state-of-the-art review. Diabetologia. 2018;61:2108–2117. doi: 10.1007/s00125-018-4670-7. [DOI] [PubMed] [Google Scholar]
- [29].Shah A, Mehta N, Reilly MP. Adipose inflammation, insulin resistance, and cardiovascular disease. JPEN. Journal of Parenteral and Enteral Nutrition. 2008;32:638–644. doi: 10.1177/0148607108325251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Baraka M, Leon R, Guindy R, Abdelrazek Ali Y. Relationship of epicardial adipose tissue measured by transthoracic echocardiography with coronary artery disease. Cardiology Plus. 2024;9:243–250. doi: 10.1097/CP9.0000000000000100. [DOI] [Google Scholar]
- [31].Gaborit B, Sengenes C, Ancel P, Jacquier A, Dutour A. Role of Epicardial Adipose Tissue in Health and Disease: A Matter of Fat? Comprehensive Physiology. 2017;7:1051–1082. doi: 10.1002/cphy.c160034. [DOI] [PubMed] [Google Scholar]
- [32].Ran CQ, He WT. Epicardial Adipose Tissue: A Potential Target to Improve Left Ventricular Diastolic Dysfunction. Reviews in Cardiovascular Medicine. 2025;26:39224. doi: 10.31083/RCM39224. [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.
Data Availability Statement
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Prof. Xinli Li (xinli3267@njmu.edu.cn).