Population
We prospectively enrolled 394 patients treated at the First Affiliated Hospital of Soochow University from November 2018 to December 2019. Fifty-five of them who were suspected of having coronary artery disease or HFpEF underwent left heart catherization. LV end-diastolic pressure (LVEDP) was invasively measured by left heart catheterization. Standard transthoracic echocardiography was performed during the 12 h before or after the procedure, and LA strain was obtained by speckle tracking echocardiography. Six of the patients were excluded because echocardiographic imaging was not good enough; thus, 49 patients served in the test group. The remaining patients (n = 339) were used to validate the results of the test group and evaluate the diagnostic performance of E/LASr in left ventricular diastolic dysfunction.
The inclusion criteria were as follows: (1) LVEF ≥ 50%; (2) no severe valvular heart disease; and (3) presence of sinus rhythm. The exclusion criteria were as follows: (1) hemodynamic instability; (2) LVEF < 50%; (3) atrial fibrillation, atrial flutter, supraventricular tachycardia, or irregular ventricular rhythm; (4) severe heart valve disease: any mitral or aortic stenosis, moderate or greater tricuspid regurgitation, moderate or greater mitral regurgitation, or experience with any valvular heart surgery or interventions; (5) insufficient echocardiographic imaging; and (6) acute ST-segment elevation myocardial infarction or acute non-ST-segment elevation myocardial infarction.
Conventional transthoracic echocardiography
Transthoracic echocardiographic measurements of all subjects were performed using a GE Vivid E9 or GE Vivid E95 (Norway) 2.5 MHz transducer in the left lateral decubitus position at rest. The biplane algorithm was used to measure the maximum volume of the left atrium in the standard apical four-chamber and two-chamber views before mitral valve opening for 1–2 frames. The LA maximal volume was divided by the body surface area to obtain the LA maximal volume index (LAVI). LVEF was measured in the standard apical four-chamber and two-chamber views by Simpson’s method biplane algorithm. Pulsed-wave Doppler (PW) was used to measure the peak early-diastolic (E) and peak end-diastolic (A) transmission velocity, E/A ratio, and E wave deceleration time at the level of the mitral leaflet tips from the apical four-chamber view. In the apical four-chamber view, the sampling points were placed at the levels of the basal portion of the septal and lateral mitral annulus. Tissue Doppler imaging (TDI) and PW were used to obtain the mitral annulus movement speed, and the peak value of the longitudinal movement in the early-diastolic period (i.e., septal e' and lateral e'). Then, the mean early-diastolic myocardial velocity (e′mean) and the ratio of E/e'mean were calculated. The maximum velocity of tricuspid regurgitation (TRmax) was measured by continuous-wave Doppler (CW) under the guidance of color Doppler from the parasternal long axis of the left ventricle or the apical four-chamber view. Researchers were blinded to the patient's LVEDP and clinical characteristics.
Two-dimensional speckle tracking echocardiography
The left atrial strain was measured using the two-dimensional strain analysis package provided by the Echo PAC workstation (GE Healthcare). The two base points of the mitral annulus and the top of the distal end of the LA were manually selected; the area of interest was adjusted to include the entire LA wall, each view was divided into six sections, and twelve sections from each patient were analyzed. The global longitudinal LA strain was measured as an average of 12 sections. The LA reservoir strain was measured as the average of the longitudinal positive peak of LA strain, which was from all LA segments (i.e., 12 segments) of the apical 4-chamber and 2-chamber views [8].
Invasive LV pressure measurements
The left ventricular filling pressure was measured using a 6F pigtail catheter. The invasive procedure was performed via the radial artery by an interventional cardiologist who was blinded to the echocardiography data. Before coronary angiography, transducers were balanced prior to the acquisition of hemodynamic data with zero level at the midaxillary line. After coronary angiography, left ventricular angiography was performed. The 6F pigtail catheter was reset routinely and placed in the left ventricle to obtain a stable pressure curve. Then, the ECG and left ventricular pressure curves were recorded simultaneously. Left ventricular end-diastolic pressure was measured at the QRS starting point for baseline stable left ventricular pressure curves. All parameters were averaged over three consecutive cardiac cycles. LVEDP > 16 mmHg was defined as elevated LVFP [1, 9].
Diagnosis of left ventricular diastolic dysfunction
According to the recommendations of the 2016 ASE/SCAI guideline[4], 339 patients (the validation group) were assessed for left ventricular diastolic dysfunction. The following are the abnormalvalues of conventional LV diastolic parameters: (1) e' of TDI mitral annulus (septal e' < 7 cm/s or lateral e' < 10 cm/s), (2) E/e 'mean > 14, (3) LAVI > 34 ml/m2, and (4) TRmax > 2.8 m/s. When more than 50% of the above criteria were positive, the patients were diagnosed with LVDD, and LV diastolic function was considered normal when less than 50% of the above criteria were positive. In addition, when only 50% of the criteria were positive, patients were diagnosed as having indeterminate LV diastolic function.
Definition of elevated left atrial pressure
According to the recommendations of the 2016 ASE/SCAI guideline [4], elevated left atrial pressure was defined as: mitral E/A ratio ≥ 2 or ≥ 2 positive criteria(LAVI > 34 mL/m2, E/e 'mean > 14, or TRmax > 2.8 m/s) when mitral E/A ratio ≤ 0.8 and E > 50 cm/s or mitral E/A ratio > 0.8 to < 2; and normal left atrial pressure was defined as: mitral E/A ratio ≤ 0.8 and E ≤ 50 cm/s or ≥ 2 negative criteria (LAVI > 34 mL/m2, E/e 'mean > 14, or TRmax > 2.8 m/s) when mitral E/A ratio ≤ 0.8 and E > 50 cm/s or mitral E/A ratio > 0.8 to < 2.
Left ventricular diastolic dysfunction grade
According to the 2016 ASE/SCAI algorithm [4], the severity of patients with left ventricular diastolic dysfunction in the validation group was graded: when mitral E/A ratio ≤ 0.8, E ≤ 50 cm/s, and more than two of the three criteria ( E/e 'mean > 14, LAVI > 34 ml/m2, TRmax > 2.8 m/s) were negative, it suggested that the corresponding grade of diastolic dysfunction was grade I; when mitral E/A ratio ≥ 0.8 and E > 50 cm/s, or if the mitral E/A ratio was > 0.8 but < 2, and two or three of the three criteria were positive at the same time, it indicated that the corresponding grade of diastolic dysfunction was grade II; when mitral E/A ratio ≥ 2, it was diagnosed as grade III diastolic dysfunction [4].
Diagnostic algorithm of HFpEF
According to the "HFA-PEFF diagnosis algorithm" offered by the 2019 ESC consensus recommendation [10], we performed clinical diagnosis of HFpEF on 339 patients in the validation group. The first was an initial workup, which included evaluating the symptoms and signs of heart failure and improving the clinical diagnosis of the primary disease (step 1). Then, the patients were assessed with echocardiography and natriuretic peptide. Diastolic function parameters of echocardiography and natriuretic peptide levels were used as the main basis for evaluating HFpEF. Then the patients were scored according to the scoring system (shown in Fig. 3 of the "HFA-PEFF Diagnosis Algorithm" [10]) (step 2). A score ≥ 5 points implied definite HFpEF. An intermediate score (2–4 points) implied diagnostic uncertainty and further hemodynamic testing was recommended, including echocardiography or invasive hemodynamic exercise stress testing (step 3). Symptoms compatible with HF could be confirmed to originate from the heart if hemodynamic abnormalities were detected either at rest or during exercise.
Definition of MACEs
Major adverse cardiac events (MACEs) included all-cause mortality, acute myocardial infarction, HF, stroke, and coronary revascularization.
Statistical analysis
Statistical analysis was performed using SPSS version 25.0 software. Continuous variables that were normally distributed are presented as the mean ± SD and were analyzed with an independent t-test. Variables that were not normally distributed are presented as medians with interquartile ranges (IQR = 25th–75th percentile) and were analyzed with the Mann–Whitney U test. Categorical data are expressed as absolute numbers or percentages and were analyzed with the chi-squared test. Univariate logistic regression was used to calculate odds ratios to predict LVEDP. All variables with p ≤ 0.100 (including LASr, E/LASr, peak A, E/e'mean, β-blockers) and age were included in the multiple logistic regression analysis to explore the relevance of LVEDP. P < 0.05 (two-tailed) was considered statistically significant. In our four models, LASr and E/LASr were analyzed separately due to their multicollinearity, with other control variables kept the same. The C-statistic was calculated in each model to allow comparison between them. In the test group, the area under the curve (AUC) of the receiver operating characteristic curve was used to compare the performance of multiple variables in determining elevated LVEDP. In the validation group, receiver operating characteristic curve analysis was used to evaluate the accuracy of the E/LASr ratio for diagnosing left ventricular diastolic dysfunction, grading the severity of LVDD and HFpEF. Univariate logistic regression was used to analyze the correlation between different variables and rehospitalization due to MACEs within one year.