This study was a case-control study including 64 consecutive patients with T2DM with a left ventricular ejection fraction (LVEF) ≥50% and 30 matched non-diabetic controls, who were hospitalized in our hospital due to evidence of ischemia on non-invasive studies and had no apparent CAD on angiogram between 2021 and 2022. This study was conducted according to the Declaration of Helsinki (DoH) ethical principles, and was further approved by the Ethics Committee of Mazandaran University of Medical Sciences, Sari, Iran (ethics code no. IR.MAZUMS.REC.1400.628). All participants also signed an informed consent form. The patients with significant CAD, HF, severe valvular heart disease, valve replacement, an LVEF<50%, left bundle branch block, atrial flutter or fibrillation, cardiomyopathies, neoplastic illnesses, decreased glomerular filtration rate or serum creatinine ≥1.5 mg/dl, and different systemic diseases were excluded from the study. The serum levels of fasting blood sugar (FBS), cholesterol (Chol), triglyceride (TG), low-density lipoprotein (LDL), and hemoglobin A1c (HbA1c) were determined by the blood sample taken after a 10–12 h overnight fasting. Moreover, 2 h post-prandial serum BS was determined by the blood sample acquired 2 h after eating a meal. The ion exchange chromatography method (Bio Systems S.A, Barcelona, Spain) was used for determining HbA1c level. Body mass index (BMI) was explained as weight in kg divided by height in meters squared. DM was defined according to the guidelines of the American Diabetes Association (ADA), including FBS ≥126 mg/dl (7.0 mmol/L) or 2 h plasma BS during 75 g oral glucose tolerance test (OGTT) ≥200 mg/dl (11.1 mmol/L), or HbA1c ≥6.5% (48 mmol/mol), and included individuals who require to take insulin or oral hypoglycemic medications [6]. Hypertension (HTN) was explained as a systolic blood pressure (SBP) ≥140mmHg and/or a DBP ≥90mmHg, determined on three separate occasions or those receiving antihypertensive medications [7]. Hyperlipidemia (HLP) was characterized as the total Chol levels over 200 mg/dl, HDL-c levels less than 40 mg/dl in males, or less than 50 mg/dl in females [8]. Cigarette smoking was identified by a face-to-face survey. Coronary angiography was performed on all patients by a Siemens AG (Medical Solutions; Erlangen, Germany) within 24–48 h of admission. An experienced cardiologist blinded to the patient data then interpreted the angiograms. No apparent CAD was defined as a stenosis of less than 20% in all coronary artery territories or the presence of only luminal irregularities [9].

A transthoracic echocardiography was performed for all patients by the ACUSON SC2000 with a 4V1c transducer (Siemens Medical Solutions Inc., Mountain View, CA, USA) within 24–48 h after hospitalization. All movies and images were stored on a hard disk for further off-line analysis (using the eSie VVI software) by an expert echocardiographer blinded to the patients’data. Epicardial fat was identified as an echo-dense space between the outer wall of the myocardium and the visceral pericardium, anterior to the right ventricular wall in the parasternal long axis view. The point of measurement was vertical to the aortic annulus, and its thickness was measured at the end-systole in an average of three cardiac cycles [10] (Figure 1).

Transthoracic two-dimensional echocardiography in the parasternal long-axis view, showing epicardial fat thickness (arrow)

The two-dimensional (2D) grayscale movies were acquired in the apical four-, two-, and three-chamber views (three standard apical views) in three cardiac cycles. The peak longitudinal strain values from the basal, mid, and apical segments of the inferoseptal, anterolateral, inferior, anterior, inferolateral, and anteroseptal walls were measured by the tracking of the endocardial and epicardial walls. To identify the LV peak global longitudinal strain (LVGLS), the average value of all 18 myocardial segments was considered [11]. The LA diameter was identified as the vertical distance between the posterior wall of the aortic root and the posterior LA wall in the para-sternal long axis view at the end systole [12]. The eSie VVI software was used to trace the LA endocardial and epicardial borders in the apical four-chamber view. The average LARS in three segments containing left LA wall, roof, and right LA wall were used to determine LARS by the R-R gating method (Figure 2). After that, LARS was corrected for LVGLS (LARS/-LVGLS). Transmitral pulse-Doppler-derived early and late diastolic velocities (E and A waves) and deceleration time (DT) of the E-wave were determined by placing the cursor at the tip of mitral valve leaflets in the apical four-chamber view. To find the mean tissue-Doppler-derived mitral annulus septal and lateral early diastolic (e°) and peak systolic (s′) velocities, the cursor was inserted at the level of the mitral annulus. The M-mode echocardiography in the parasternal long-axis view was further employed to determine the end-systolic and -diastolic LV internal diameters and end-diastolic interventricular septal (IVS) and posterior wall thickness by inserting the cursor at the mitral valve leaflet tip. The LV mass index (LVMI) was calculated by the formula given below:

$$0.\left( .0\left[ } + } + }} \right) \, \left( }} \right)} \right]} \right) + 0.$$

Speckle-tracking echocardiography in apical four chamber view represents longitudinal LA reservoir strain curves in three segments containing left LA wall, roof, and right LA wall by considering the QRS complex (R-R gating) as the initiation of the strain calculation (LA Left atrium)

In which, LVIDd, PWT, and SWT indicate the end-diastolic LV internal dimension and posterior and IVS wall thickness, respectively. We also obtained peak velocity of tricuspid regurgitation jet by continuous-Doppler from multiple views [11]. The reproducibility of the EFT and speckle-tracking-derived LVGLS and LARS measurements were determined according to repeated measurements in 10 randomly selected patients by an echocardiography within 48 h, and the intra-observer correlation coefficients were found to be 0.91, 0.90, and 0.92, respectively.

The categorical variables were presented as frequency and percentage. The quantitative variables were expressed as the mean± standard deviation (for the normally distributed continuous variables) or median [lower-upper quartile] (for Non-normally distributed continuous variables). Normality was determined by the Shapiro-Wilk test. An independent t-test was used to compare the continuous variables that were normally distributed and the Mann-Whitney U test was used to compare the continuous variables that were non-normally distributed. The categorical variables were compared by Chi-square test and Fisher’s exact test. Moreover, we used a Pearson correlation to find the correlation between different echocardiographic variables in the patients with T2DM and the non-diabetic ones, separately. Multiple linear regression analysis was also conducted on all study population, including the patients with T2DM and the non-diabetic controls to determine the independent correlation between different demographic, laboratory-based, and echocardiographic variables and LARS/-LVGLS. The statistical analyses were performed by the SPSS/PASW (Predictive Analytics Software) Statistics version 18 (SPSS, Chicago, IL, USA). The P value of less than 0.05 signified that the variable was significant. The sample size was determined based on the findings of Evin et al study [2]. According to mean and SD of LA longitudinal strain in case and control groups of the aforementioned study, the effect size was estimated to be 0.69. Considering a confidence level of 95%, the test power of 80%,, a one-tailed test for the comparison of two means, and using the G*power software, the minimum sample size was determined to be 54 individuals (viz., 27 in each group). Given the available facilities and the aim to increase the power of the statistical tests, more samples (that is, 64 individuals in the case group and 30 people as the controls) were included in this study.