Subject design and recruitment
A consecutive case series of T2DM patients hospitalized in the Department of Endocrinology at the Changzhou First People's Hospital (Changzhou, Jiangsu, China) were recruited from April 2018 to May 2022. The Inclusion criteria were: 1) diagnosed of T2DM according to the criteria of World Health Organization[12] and Chinese Diabetes Society[13] without cardiac symptoms; 2) aged from 18 to 70 years old independent of T2DM duration. The exclusion criteria were: 1) subjects with hypertension, CAD, atrial fibrillation, structural heart disease or history of any cardiovascular-related disease; 2) subjects with diabetic complications including macro and microvascular diseases such as neuropathy, retinopathy, kidney disease, stroke and peripheral vascular disease; 3) pregnancy; 4) other serious comorbidities, including thyroid disturbances, malignant tumors, liver and renal insufficiency, rheumatic diseases or major mental illness. Al the subjects signed written informed consent forms before the start of this study. The study was approved by the Institutional Review Committee and the Ethics Committee of the Third Affiliated Hospital of Soochow University.
Clinical and biochemical measurements
Baseline characteristics including age, sex, weight, height, body mass index (BMI), diabetic duration and other complete medical history were recorded in detail on the day of admission. After fasting for at least 8 hours, peripheral venous blood was collected before administration of hypoglycemic drugs on the morning after admission. Briefly, the concentration of HbA1c was evaluated through high performance liquid chromatography. Glutamic oxaloacetic transaminase (ALT), alanine aminotransferase (AST), creatinine (Cr), urea nitrogen (BUN), homocysteine, total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), fasting blood glucose (FBG), and C peptide (0min, 30min, 60min, 120min and 180min) were analysed by an automatic analyzer.
Echocardiographic measurements
The following parameters were measured and analyzed by echocardiography: the left atrial diameter (LAD), left ventricular end-systolic diameter (LVESD), left ventricular end-diastolic diameter (LVEDD), interventricular septal diameter (IVSD), left ventricular posterior wall thickness (LVPWT), left ventricular ejection fraction% (LVEF%), peak late diastolic trans-mitral flow velocity (MFV A), peak early diastolic trans-mitral flow velocity (MFV E), mitral valve septal velocity e, mitral valve lateral velocity e, and LA volume. e’ was defined as: (ventricular septal velocity e + mitral valve velocity e)/2. Left atrial volume index (LAVI) was defined as: LA volume/body surface area, where the body surface area was equal to 0.0128*weight (kg) + 0.006*height (cm) − 0.1529. Left ventricular mass index (LVMI) was defined as: 0.8*10.4 (IVSD + LVPWT + LVEDD). Relative ventricular wall thickness (RWT) was defined as: (LVPWT/LVEDD) *2.
TyG index and HFA-PEFF score definition
The TyG index was calculated as: In (fasting TG [mg/dl] x fasting glucose [mg/dl]/2). The Heart Failure Association (HFA)-PEFF score was originated from HFA-PEFF diagnostic algorithm, including functional, morphological, and biomarker domains[14]. Data of the peak tricuspid velocity and global longitudinal strain were not available in this study. In the HFA-PEFF diagnostic algorithm, a total score ≥ 5 points was identified to be diagnostic of heart failure with preserved ejection fraction (HFpEF), while a score ≤ 1 was considered to be very unlikely of HFpEF. Patients with an intermediate PEFF score (2–4 points) required further functional and aetiology assessment[14] .
Statistical analysis
All data in this study were analyzed using SPSS statistical software 26.0. P value < 0.05 was defined to be of statistical significance. The specific statistical analysis in this study were outlined as follows.
Baseline and echocardiographic data of subjects
The baseline and echocardiographic data of diabetic patients were stratified based on binary TyG index. The differences between two groups were evaluated, continuous normal distribution variables were expressed as mean ± standard deviation (SD) by independent sample t-test, nonnormal distribution variables were expressed as median P50 (P25, P75) by Mann-Whitney U test, and the categorical variables were presented as number (percentage) and analyzed by Х2 test.
Correlation analysis
Pearson correlation analysis was used to analysis independent variables with the TyG index. Partial correlation analysis was used to correct for suspicious confounding factors (make it/them a constant).
Logistic regression
A logistic multivariable regression analysis with cardiac diastolic dysfunction categorized as HFA-PEFF score (≤ 1, 2–4, and ≥ 5 points) was used to determine the associations between the TyG index and HFpEF. The goodness of fit of the regression model was assessed by Hosmer-Lemeshow test (P > 0.05). In the logistic regression analysis, three models were set up, Model 1: adjusted by age and sex; Model 2: adjusted by BMI and diabetic duration based on model 1; Mode 3: adjusted by TC and HbA1c based on model 2.
ROC curve analysis
Receiver operating characteristic curve (ROC) analysis was constructed to evaluate the predictive value of TyG index, HbA1c, FBG, and TG/HDL-C for the subclinical HFpEF presence according to the value of the area under the ROC curve (AUC).
Subgroup analysis
A stratified analysis was conducted based on sex, age, HbA1c and T2DM duration to eliminate the interference of confounding factors. Among them, means of age and HbA1c were defined as stratification criteria while the median of duration was used for the cut-off point since the latter does not conform to the normal distribution.