1. Data sources
This study was a prospective study. We systematically yield the clinical data from a total of 295 inpatients with chest pain who underwent CAG, admitted to the Cardiovascular Department of Xiyuan Hospital of China Academy of Chinese Medical Sciences from October 2023 to June 2024, 128 patients were ultimately included (Fig. 1). And partial CMVD patients were from the AWARD (aromatic and warming-up management in coronary microvascular disease) study cohort. This study was approved by the Ethics Committee of Xiyuan Hospital (2024XLA031-2). The study adhered to the clinical practice standards set by the Declaration of Helsinki
2 Diagnostic criteria
The diagnostic criteria for CMVD in the Chinese Multidisciplinary Expert Consensus on the Diagnosis and Treatment of Microvascular Diseases issued by China in 202018 are as follows [16]: (ⅰ) symptoms of myocardial ischemia; (ⅱ) objective evidence of myocardial ischemia; (ⅲ) non-obstructive coronary artery disease, with coronary artery diameter stenosis of < 50% on coronary CT angiography (CTA) or invasive CAG; and (ⅳ) objective evidence of impaired coronary microcirculatory function, impaired coronary flow reserve - coronary flow reserve (CFR) < 2.5 or < 2.0, or the slow coronary flow phenomenon (TIMI frames > 25), or the index of microvascular resistance index (IMR) > 25.
3. Inclusion criteria:
(1) CMVD group: (ⅰ) no significant stenosis in 3 coronary arteries including the left anterior descending artery (LAD), left circumflex artery (LCX), and right coronary artery (RCA) (stenosis < 50%); (ⅱ) quantitative flow ratio (QFR) > 0.8 in 3 coronary arteries and angiography-derived index of microcirculation (AMR) > 250 mmHg*s/m in at least 1 coronary artery [17].
(2) Obstructive-CAD group: (ⅰ) stenosis (> 70%) or QFR < 0.8 in at least 1 coronary artery; (ⅱ) AMR < 250 mmHg*s/m both in non-culprit vessel and in the opened culprit vessels after operations.
4. Exclusion criteria
(1) Patients who underwent coronary revascularization within 1 month before CAG, including percutaneous coronary intervention (PCI), percutaneous transluminal coronary angioplasty (PTCA), and coronary artery bypass grafting (CABG);
(2) Organic heart disease, such as severe HF, malignant arrhythmia, or valvular disease;
(3) Acute and chronic inflammation, immune diseases, or malignant tumors;
(4) Coronary occlusive lesions on CAG, or severe tortuosity of the image of CAG;
(5) Severe renal insufficiency: glomerular filtration rate < 45 mL/min;
(6) Severe coagulation disorders or bleeding;
(7) Allergy to contrast media;
(8) Abnormal ventricular wall motion in the echocardiogram;
5. Cardiac catheterization examination
After CAG, we performed angiographic functional examination. Real-time CAG images were conducted to the QFR instrument (Pulse Medical, Shanghai) and analyzed under a single angle using the software AngioPlus Core software (version V2). The process was performed by a trained cardiologist. Firstly, selecting the optimal contrast view with minimal vessel overlap was selected, and then the lumen of the target vessel detected and outlined. The contrast flow rate is derived by dividing the vessel length by the contrast fill time, which is then converted to the filling flow rate. Subsequently, frames in which the lumen contour was fully exposed were selected as analysis frames to determine the main and side branches of the analyzed vessels. According to Murray's bifurcation law, the reference vessel diameter needs to be reconstructed [18–22]. Finally, according to fluid dynamics, the QFR and AMR were calculated via the following equations [23].
$$\:\mathbf{A}\mathbf{M}\mathbf{R}=\frac{\mathbf{P}\mathbf{d}}{{\mathbf{V}\mathbf{e}\mathbf{l}\mathbf{o}\mathbf{c}\mathbf{i}\mathbf{t}\mathbf{y}}_{\mathbf{h}\mathbf{y}\mathbf{p}}}=\frac{\varvec{P}\varvec{a}\times\:\varvec{u}\varvec{Q}\varvec{F}\varvec{R}}{{\mathbf{V}\mathbf{e}\mathbf{l}\mathbf{o}\mathbf{c}\mathbf{i}\mathbf{t}\mathbf{y}}_{\mathbf{h}\mathbf{y}\mathbf{p}}}$$
1
where Pa is the aortic pressure at baseline; Pd is the distal coronary artery pressure; and Velocityhyp indicates simulating hyperemic velocity.
6. Echocardiography examination
Well-trained echocardiographers performed routine transthoracic echocardiography and interpreted all echocardiograms: left ventricle internal dimension measured at end-diastole, left ventricle internal dimension measured at end-systole measured with M-mode, and early diastolic mitral inflow velocity/late diastolic mitral inflow velocity, early diastolic mitral inflow velocity/peak early diastolic mitral septal annular velocity (E/e'), and tricuspid regurgitation velocity (TR) measured [24].
The echocardiography-based diagnosis of LVDD was published in 2016 by the American Society of Echocardiography and the European Association of Cardiovascular Imaging [25]: (ⅰ) mitral annulus e' velocity (septal e'<7 cm/s or lateral e'<10 cm/s) [25]: (ⅱ) average E/e' >14; (iii) left atrial volume index > 34 ml/m2; and (iv) maximum tricuspid regurgitant velocity > 2.8 m/s.
Diagnosis of LVDD: A positive result for two or more of the above indicators suggests LVDD, and a negative result for two or more suggests non-LVDD.
7. Statistical analysis
We used R-4.2.2 (https://cran.r-project.org/) software for statistical analysis. First, the "Tableone" package was used to compare the statistical significance of categorical and continuous variables in the two groups via the chi-square, Kruskal‒Wallis test, and the Nemenyi test, and P-values < 0.05 were considered statistically significant. The Shapiro Test was used to determine whether the measurements were normally distributed; if the data were normally distributed, the data were expressed as the mean ± standard deviation (X ± S), and Student’s t test was used to explore differences in variables. If the data did not conform to a normal distribution, they were expressed as medians and interquartile ranges, and nonparametric tests were used for the statisticians.
The "ggpolt" package was used to explore the linear correlation of two variables via Spearman's correlation coefficient. Therefore, we analyzed the linear correlation between the indices of LVDD (including average E/e', septal e' and lateral e' velocity) and AMR (for the obstructive-CAD group, the AMR includ all vessels before and after operations). Furthermore, the relationships between LVDD and AMR indices were investigated via binary logistic regression with the "glm" function. And the results were represented by odds ratio (OR) and 95%confidence intervals (CI). In terms of stratified analyses, we clarified the distribution differences in AMR and simulating hyperemic velocity (SHV) between the LVDD and non-LVDD groups by plotting violin plots via the "ggplot" package. Finally, the "pROC" package was used to construct receiver operating characteristic (ROC) curves to investigate the ability of the AMR to identify LVDD between macrovascular and microvascular lesions.