Background
Regional strain by speckle-tracking echocardiography can be used to detect occult right ventricular (RV) myocardial dysfunction. However, in patients with coronary artery disease, the impact of RV strain is unknown. The aim of this study was to validate the prognostic value of RV strain in patients with non–acute coronary syndrome angina.
Methods
In total, 208 patients with coronary artery disease proved by coronary angiography were retrospectively identified (mean age, 63.81 ± 10.12 years; 161 men). In addition to clinical and traditional echocardiographic parameters, RV free wall longitudinal strain (RVLS_FW) was measured using speckle-tracking echocardiography from the apical four-chamber view. Cardiac mortality and major cardiovascular events were recorded.
Results
During follow-up (23.1 ± 6.6 months), 27 patients (12.98%) died of cardiovascular causes. These patients were found to have lower left ventricular ejection fractions and greater reductions in the amplitudes of left ventricular peak systolic global longitudinal strain and RVLS_FW. Using −18% as a cutoff point for RV strain, patients with strain ≤ −18% had superior outcomes (log-rank χ 2 = 9.04 and 6.94 for cardiovascular mortality and arrhythmia, respectively, P = .003). In Cox multivariate regression analysis, RVLS_FW was an independent prognostic factor for both cardiovascular mortality (hazard ratio, 1.27; 95% CI, 1.03–1.72; P = .01) and hemodynamically unstable ventricular arrhythmia (hazard ratio, 1.72; 95% CI, 1.08–2.41; P = .01).
Conclusions
RVLS_FW is significantly related to cardiovascular outcomes and hemodynamically unstable ventricular arrhythmia in patients with non–acute coronary syndrome angina.
Right ventricular (RV) myocardial dysfunction has been reported to increase in-hospital mortality, at least in part because of a significantly higher risk for hemodynamically compromising arrhythmias. It has also been shown that the right ventricle plays a pivotal role in maintaining hemodynamic stability and arrhythmia and that it is associated with the outcomes of various diseases. Despite the emergence of several echocardiographic tools, such as Doppler tissue imaging and tricuspid annular plane systolic excursion, the complex geometry of the right ventricle still makes comprehensive evaluations difficult. Therefore, an optimal parameter to accurately investigate occult RV dysfunction is required. Speckle-tracking echocardiography (STE) is an emerging technique that is used to evaluate early myocardial dysfunction with angle-independent characteristics. In recent years, an increasing number of studies have shown that left ventricular (LV) longitudinal strain is significantly related to long-term outcomes in patients with chronic ischemic cardiomyopathy and those who have had myocardial infarctions and that it is superior to LV ejection fraction (LVEF). Conversely, whereas most studies regarding the right ventricle have focused on congenital heart defects, pulmonary hypertension, pulmonary embolism, and right cardiomyopathy, only a few studies have investigated the relationship between RV speckle-tracking and coronary artery disease (CAD). Among these investigations, although acute myocardial infarction has been discussed more than chronic ischemic heart disease, the prognostic impact and referenced normal range of RV strain are still lacking. In a previous study, we successfully demonstrated the usefulness of RV strain in the diagnosis of RV involvement in patients with chronic CAD. The aim of this study was to evaluate the association between RV strain and the outcomes of non–acute coronary syndrome (ACS) angina. We hypothesized that occult systolic dysfunction of the right ventricle may influence the prognosis in patients with non-ACS angina.
Methods
Subjects
Two hundred fifty-six patients with suspected CAD who underwent elective coronary angiography were identified retrospectively. All patients presented with angina and underwent stress tests, including thallium scans and treadmill or stress echocardiography before angiography. CAD was diagnosed according to coronary angiographic findings and was defined as significant coronary stenosis (>50% stenosis in diameter). Echocardiography was also performed before the procedure. The exclusion criteria were acute or prior myocardial infarction, arrhythmia during the examination, moderate or severe valvular heart disease, moderate or severe pulmonary hypertension (>50 mm Hg), and hypotension. All patients underwent extensive evaluations of baseline clinical history, laboratory test results, and cardiovascular risk factors. This study adhered to the principles of the Declaration of Helsinki and was approved by the Human Research and Ethics Committee of National Cheng Kung University Hospital. All enrolled patients or their families provided written informed consent. In addition, 50 age- and gender-matched control subjects with negative stress test results were also recruited as the comparison cohort.
Echocardiography
Standard echocardiography was performed (Vivid 7; GE Vingmed Ultrasound AS, Horten, Norway) with a 3.5-MHz multiphase-array probe in accordance with the recommendations of the American Society of Echocardiography. Chamber dimensions and LVEF were measured using the two-dimensionally guided M-mode method and the biplane Simpson method, respectively. RV fractional area change was measured in an apical four-chamber view. Left atrial volume index was calculated as 0.85 × (area in the four-chamber view × area in the two-chamber view)/average of vertical axes in the four-chamber and two-chamber views. LV diastolic function–associated parameters, including isovolumic relaxation time, deceleration time, transmitral early filling velocity (E)/atrial velocity (A) ratio, and mitral E/early diastolic mitral annular velocity (e′) ratio, were also measured. Peak systolic pulsed Doppler tissue imaging was performed at the tricuspid annulus (S′). RV dimensions were defined as basal, midcavity, and longitudinal diameter in an apical four-chamber view. RV outflow tract dimensions at the proximal or subvalvular level and at the distal or pulmonic valve and tricuspid annular plane systolic excursion were also measured. In addition, pulmonary artery systolic pressure was obtained by summation of estimated trans–tricuspid valve pressure and estimated right atrial pressure. The severity of tricuspid regurgitation was defined by the vena contracta. The echocardiographic readers who analyzed the data were blinded to the coronary angiographic results and the patients’ clinical information.
Speckle-Tracking Echocardiographic Analysis of Deformation
Standard apical four-, two-, and three-chamber views were recorded in digital loops for deformation analysis of the left ventricle, and an apical four-chamber view focusing on the right ventricle was used for RV deformation analysis. The images were acquired with frame rates of 70 to 90 frames/sec and stored for three cycles. The images were analyzed offline using computer software (EchoPAC; GE Vingmed Ultrasound AS). As described elsewhere, we used automated function imaging software to measure LV peak systolic global longitudinal strain (GLS). In brief, LV GLS was calculated automatically by the software after defining the timing of aortic valve closure. RV deformation was measured by two-dimensional STE in the apical four-chamber view. RV free wall longitudinal strain (RVLS_FW) and strain rate were derived from the average of three regional strain values comprising the lateral wall ( Figures 1A and 1B ).
Reproducibility
Using Bland-Altman limits of agreement and interclass correlation coefficients, 20 patients were randomly selected to assess intra- and interobserver variability.
Follow-Up and End Points
After the diagnosis of CAD by catheterization, patients who received treatment including interventions were followed at our outpatient clinics after discharge. In general, the interval between each visit was 3 months, but additional visits were scheduled on the basis of patients’ clinical requirements. Twelve-lead electrocardiography was performed on arrival and repeated on the basis of clinical requirements. New-onset arrhythmia was defined by documented electrocardiography, 24-hour Holter recording, or bedside electrocardiographic monitoring during the stay in the postcatheterization care unit and every follow-up visit at the clinic. The diagnosis was confirmed by expert cardiologists not involved in this study. Among the diagnoses, the patients with atrial fibrillation, ventricular tachycardia, and ventricular fibrillation >30 sec were enrolled in this study. Some patients developed more than one kind of arrhythmia, and they were counted according to the various types of arrhythmias. Regarding management, the primary strategy was to restore sinus rhythm with overdrive pacing, electrical cardioversion, and/or antiarrhythmic drug therapy, or whatever seemed appropriate in the given clinical context. Hemodynamic status was also recorded when arrhythmia occurred, and a hemodynamically unstable ventricular arrhythmia (HUVA) was defined as a ventricular arrhythmia that required cardioversion or defibrillation. Data on the occurrence of mortality and adverse events at follow-up were collected by careful medical record reviews during the follow-up period (December 2010 to June 2013).
Cardiovascular mortality due to myocardial infarction, heart failure, or arrhythmia was regarded as the primary outcome. The secondary outcomes were major adverse cardiovascular events such as cardiovascular death, ischemic events including admission for angina and coronary revascularization, new-onset arrhythmia, and especially hospitalizations for HUVA and heart failure ( Supplemental Figure 1 ). Patients without specific data during the prior 6-month period were considered lost to clinical follow-up. The data from these patients were included as of the date of last follow-up.
Statistical Analysis
Differences between patients with or without end points were compared using Student’s t tests for normally distributed continuous variables, nonparametric tests for continuous variables not normally distributed, and χ 2 tests for categorical variables. Blood pressure is presented as median (interquartile range). Group differences were analyzed using analysis of variance and post hoc Scheffé tests. Variables with P values < .10 in the univariate analysis were included in the multivariate Cox regression analysis. To analyze RVLS_FW as a continuous or categorical variable, two Cox regression models were used (model 1 and model 2, respectively). Factors including RVLS_FW that were significant in the univariate analysis were entered into model 1. In model 2, a dichotomized RVLS_FW value by −18% was used instead of RVLS_FW. Event rates from the time of inclusion were calculated using the Kaplan-Meier method. The log-rank test for time-to-event data with respect to the outcome was used for statistical comparisons between the two patient groups. P values < .05 were considered to indicate statistical significance. All analyses were performed with SPSS version 18 for Windows (SPSS, Chicago, IL).
Results
Clinical and Echocardiographic Characteristics of Survivors versus Patients with Cardiovascular Death, Arrhythmia, and Heart Failure
After excluding patients whose images were of suboptimal quality ( n = 17 [7.4%]), those who were lost to follow-up ( n = 3 [1.3%]), and those who died of sepsis and bleeding-related complications, a total of 208 patients (mean age, 63.8 ± 10.2 years; 161 men) were entered into the analysis ( Supplemental Figure 1 ). Among them, the incidence rates of left main, left coronary artery, and right coronary artery (RCA) lesions were 20.2%, 76.4%, and 62.5%, respectively. The median follow-up duration for the entire patient population was 23.1 ± 6.6 months. During follow-up, 27 patients (12.98%) died of cardiovascular causes, including myocardial infarction ( n = 12 [44.4%]), heart failure ( n = 7 [25.9%]), and arrhythmia ( n = 8 [29.62%]). These patients had significantly higher percentages of left main or RCA lesions, hospitalization for exacerbated heart failure, and new-onset arrhythmia compared with survivors ( Table 1 ). In echocardiographic analysis, cardiovascular mortality was significantly associated with a lower LVEF and a greater reduction in the amplitude of LV GLS or RVLS_FW ( Table 2 ). In addition, compared with the gender- and age-matched control subjects ( n = 50), patients with non-ACS angina (both survivors and nonsurvivors) had reduced amplitudes of both LV GLS and RVLS_FW.
| Characteristic | Control ( n = 50) | Survivors ( n = 181) | CV death ( n = 27) |
|---|---|---|---|
| Demographic characteristics | |||
| Age (y) | 63 ± 10.7 | 63.11 ± 11.9 | 67.5 ± 9.7 |
| Male sex | 40 (80.0%) | 141 (77.9%) | 20 (74.0%) |
| Medical history | |||
| Diabetes | 12 (24.0%) | 76 (41.9%) | 13 (48.1%) ∗ |
| Hypertension | 14 (28.0%) | 137 (75.7%) | 20 (74.1%) ∗ |
| Dyslipidemia || | 12 (24.0%) | 92 (50.8%) | 13 (48.1%) ∗ |
| CKD § | 2 (4.0%) | 28 (15.4%) | 6 (22.2%) ∗ |
| Smoking | 15 (30.0%) | 52 (28.7%) | 8 (29.6%) |
| Heart failure, NYHA functional class ≥ III | 0 (0.0%) | 12 (5.7%) | 2 (7.4%) |
| Systolic blood pressure (mm Hg) | 135 (115–142) | 138 (115–162) | 127 (108–148) |
| Diastolic blood pressure (mm Hg) | 82 (70–93) | 83 (71–96) | 78 (65–92) |
| Therapeutic history | |||
| β-blocker | 6 (12.0%) | 75 (41.4%) | 12 (44.4%) |
| Antiarrhythmia medication | 0 | 62 (34.2%) | 10 (37.0%) ‡ |
| Cardioversion | 0 | 8 (4.4%) | 8 (29.6%) † |
| ICD implantation | 0 | 2 (1.1%) | 2 (7.4%) † |
| Rehospitalization for exacerbated heart failure | — | 25 (13.8%) | 8 (29.6%) † |
| Coronary angiographic or interventional data | |||
| LM | — | 32 (17.7%) | 10 (37.0%) † |
| LAD | — | 139 (76.8%) | 20 (74.1%) |
| RCA | — | 106 (58.5%) | 24 (88.9%) † |
| LCX | — | 110 (60.7%) | 11 (66.6%) |
| 1VD | — | 28 (15.4%) | 3 (11.1%) |
| 2VD | — | 36 (19.8%) | 11 (40.7%) |
| 3VD | — | 66 (36.4%) | 16 (59.2%) † |
| Percutaneous coronary intervention | — | 110 (60.7%) | 19 (70.4%) |
| Coronary artery bypass grafting | — | 25 (13.8%) | 6 (22.2%) |
| New-onset arrhythmia | — | 38 (20.9%) | 21 (77.8%) † |
| AF or SVT | — | 28 (20.9%) | 12 (44.4%) † |
| Stable VT | — | 12 (12.1%) | 5 (14.8%) |
| Hemodynamically unstable VT or VF | — | 6 (3.3%) | 10 (37.0%) ‡ |
† P < .05 versus survivors (one-way analysis of variance, post hoc Scheffé test).
§ Chronic kidney disease was defined as glomerular filtration rate < 60 mL/min/1.73 m 2 , including dialysis.
|| Hyperlipidemia was defined as according to the Adult Treatment Panel III as total cholesterol > 240 mg/dL or triglyceride > 200 mg/dL.
| Characteristic | Control ( n = 50) | Survivors ( n = 181) | CV death ( n = 27) |
|---|---|---|---|
| Left heart–associated parameters | |||
| LVEDV index (mL/m 2 ) | 75.98 ± 8.3 | 76.3 ± 7.1 | 78.9 ± 9.2 |
| LVMI (g/m 2 ) | 92.83 ± 35.4 | 93.32 ± 28.2 | 99.66 ± 48.8 |
| LVEF (%) | 63.96 ± 13.8 | 61.13 ± 14.5 | 53.96 ± 13.8 ∗ ‡ |
| LV GLS (%) | −19.07 ± 6.9 | −14.74 ± 4.2 | −12.18 ± 2.3 † § |
| IVRT (msec) | 93.79 ± 23 | 91.7 ± 13.8 | 89.37 ± 21.2 |
| DT (msec) | 209.35 ± 68.5 | 209.35 ± 62.5 | 196.3 ± 51.1 |
| E/A ratio | 1.09 ± 0.4 | 1.19 ± 0.4 | 1.22 ± 0.8 |
| E/E′ ratio (mean) | 7.61 ± 2.5 | 9.12 ± 2.5 | 10.4 ± 4.5 ∗ ‡ |
| LAV index (mL/m 2 ) | 27.41 ± 5.8 | 25.78 ± 6.7 | 29.81 ± 8.4 |
| Right heart–associated parameters | |||
| RV dimension 1 (cm) | 5.16 ± 0.5 | 5.08 ± 0.5 | 5.87 ± 0.6 |
| RV dimension 2 (cm) | 2.05 ± 0.5 | 2.44 ± 0.3 | 2.51 ± 0.3 |
| RV dimension 3 (cm) | 2.33 ± 0.8 | 2.27 ± 0.7 | 2.15 ± 0.6 |
| RVOTp (cm) | 2.66 ± 0.9 | 2.4 ± 0.4 | 2.1 ± 0.6 |
| RVOTd (cm) | 2.00 ± 0.5 | 2.18 ± 0.3 | 2.19 ± 0.7 |
| Estimated pulmonary artery systolic pressure (mm Hg) | 19.69 ± 7.5 | 19.10 ± 3.9 | 23.2 ± 5.1 |
| Tricuspid valve vena contracta (cm) | 0.18 ± 0.1 | 0.23 ± 0.2 | 0.21 ± 0.1 |
| RV FAC (%) | 70.13 ± 10.9 | 69.73 ± 13.6 | 67.3 ± 14.0 |
| TAPSE (cm/sec) | 1.87 ± 2.7 | 1.96 ± 0.5 | 1.93 ± 0.5 |
| S′ (cm/sec) | 11.89 ± 2.3 | 11.46 ± 2.3 | 12.13 ± 2.6 |
| RVLS_FW (%) | −19.01 ± 3.5 | −17.12 ± 5.7 | −12.47 ± 3.7 † § |
| RVLSR_FW (%) | −1.29 ± 0.5 | −1.35 ± 0.4 | −1.14 ± 0.3 |
‡ P < .05 versus survivors (one-way analysis of variance, post hoc Scheffé test).
With respect to other cardiovascular events during postcatheterization care, 59 of the 208 patients (28.36%) developed new arrhythmias, including 40 (19.23%) supraventricular arrhythmias (including atrial fibrillation, atrial flutter, and paroxysmal supraventricular tachycardia), 17 (8.17%) hemodynamically stable ventricular arrhythmias, and 16 (7.69%) HUVA. Compared with the nonarrhythmia group, the patients who experienced any type of arrhythmia had a higher rate of RCA lesions, attenuated LVEFs, and reduced amplitudes of LV GLS and RVLS_FW ( Supplemental Tables 1 and 2 ).
Similarly, heart failure progression was closely related to multivessel CAD or left main involvement. Furthermore, chronic kidney disease was significantly related to hospitalization for worsening heart failure, in contrast to the group with stable heart function (34.4% vs 8.9%, P = .001).
The Impact of RV Strain on Cardiovascular Death and New-Onset HUVA
According to the reference value of our previous study, using −18% as a cutoff point for RVLS_FW yielded sensitivity and specificity of 73.4% and 74.9%, respectively, for a diagnosis of proximal RCA lesions. Therefore, we dichotomized RV strain at the value of −18%. The results of Cox regression analysis showed that patients with strain ≤ −18% had superior outcomes compared with those with RVLS_FW > −18% (log-rank χ 2 = 4, P = .003) ( Figure 2 A). During the follow-up period, a cumulative 3.9%, 10.5%, and 22.3% of patients with RV strain > −18% (reduced amplitude of RV strain) died at 1, 2, and 3 years, respectively. In contrast, patients with RV strain ≤ −18% (increased amplitude of RV strain) had a significantly lower incidence of cardiovascular death and were even free from mortality in the first 2 years. Interestingly, by dichotomizing the population at this cutoff point, we illustrated a similar effect of impaired RV strain to newly developed arrhythmia (log-rank χ 2 = 6.94, P = .003) ( Figure 2 B). During the follow-up period, the cumulative rates of newly developed arrhythmia in the group of patients with RV strain > −18% were 13%, 26%, and 34.2% at 1, 2, and 3 years, respectively. In contrast, the group of patients with RV strain ≤ −18% had fewer new events with cumulative rates of 6.5%, 9.2%, and 11.8% at 1, 2, and 3 years, respectively. When focusing on HUVA, the correlation between preserved RV strain and lower event rate was significantly persistent (log-rank χ 2 = 7.45, P = .006) ( Figure 2 C).
The Correlation between RV Function and HUVA
According to our findings, a higher rate of cardiovascular death was correlated with a reduced amplitude of RV strain and any type of newly developed arrhythmia, especially HUVA. Given the specific relationship between fatal arrhythmia and RV dysfunction, we divided the patients with reduced amplitude of RV strain (>−18%) by HUVA events. The results showed that in patients with reduced amplitude of RV strain, the incidence of HUVA and the risk for cardiovascular death were significantly higher. During the follow-up period, the cumulative mortality rates in the patients with preserved RVLS_FW, impaired RVLS_FW, and impaired RVLS_FW plus the development of HUVA increased to 1.3%, 5.9%, and 50%, respectively, at 30 months ( Figure 2 D).
Prognostic Significance of RV Strain
In the Cox multivariate regression analysis, compared with other clinical and echocardiographic parameters, RVLS_FW was significantly correlated with cardiovascular mortality ( Table 3 ). By dichotomizing RVLS_FW at −18%, we found that it was significantly linked to cardiovascular mortality ( Table 3 ) and HUVA ( Supplemental Table 3 ). To validate the significance of RVLS_FW, we sequentially included various parameters into the multivariate Cox regression analysis, and RVLS_FW remained significant in the different models ( Supplemental Table 4 ). Furthermore, old age and chronic kidney disease also had significant effects in multivariate regression analysis for hospitalization for exacerbation of heart failure ( Supplemental Table 5 ).
| Predictor | Univariate analysis | Multivariate analysis | ||||
|---|---|---|---|---|---|---|
| Model 1 | Model 2 | |||||
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | |
| Cardiovascular death | ||||||
| Age | 2.3 (0.9–5.7) | .58 | ||||
| LM involvement | 3.1 (0.6–7.6) | .52 | ||||
| RCA involvement | 5.2 (1.0–58.5) | .04 ∗ | 3.1 (0.8–73.4) | .25 | 2.1 (0.5–33.2) | .16 |
| CAD/3VD | 1.7 (0.8–2.2) | .03 ∗ | 2.4 (0.7–5.6) | .09 ∗ | 1.3 (0.8–1.9) | .05 ∗ |
| Decreased LVEF (<50%) | 0.9 (0.7–1.8) | .14 | ||||
| Impaired GLS (>−15%) | 1.1 (1–1.3) | .02 ∗ | 1.1 (0.9–4.3) | .08 | 1.1 (0.9–1.3) | .06 |
| RVLS_FW | 1.8 | .002 ∗ | 1.1 (1.0–1.6) | .04 ∗ | ||
| Impaired RVLS_FW (>−18%) | 2.2 (1.1–4.3) | .002 ∗ | 1.2 (1.0–1.7) | .01 ∗ | ||
The Value of RV Strain on Outcomes in Patients with Preserved LV GLS
To avoid the influence of LV dysfunction on cardiovascular outcome, we performed subgroup analysis by excluding patients with reduced LV GLS amplitude (>−15%) to investigate the pure effect of RVLS_FW on RV dysfunction and subsequent outcomes. Among the recruited 107 patients, the 11 (10.28%) who died of cardiovascular events had higher rates of RCA lesions, new onset of any arrhythmia, and reduced RVLS_FW amplitude ( Supplemental Table 6 ). Using −18% as a cutoff value, the patients with preserved RVLS_FW had superior cardiovascular survival (log-rank χ 2 = 7.6, P = .001) ( Supplemental Figure 2 ). In the Cox multivariate regression analysis, only RV strain was significantly correlated with cardiovascular mortality. According to these findings, RVLS_FW was associated with cardiovascular death independently of LV dysfunction.
Using an LVEF cutoff value of 35%, we divided patients with non-ACS angina into two groups. Interestingly, we found that LVEF and LV GLS were associated with mortality in both groups, but the amplitude of RVLS_FW was significantly reduced in the nonsurvivors who had LVEFs ≥ 35% ( Supplemental Table 7 ). This indicated that the value of RVLS_FW was more significant in patients with relatively preserved LVEFs.
Reproducibility of LV and RV Strain
The echocardiographic images of 20 randomly selected patients were analyzed by two readers a total of three times each. Each measurement was taken at 15-min intervals. Readers could select the best cardiac cycle by themselves and were blinded to previous measurements. The intra- and interobserver interclass correlation coefficients for RVLS_FW were 0.94 (0.7–0.95) and 0.97 (0.94–0.99), respectively. For LV GLS, the intra- and interobserver interclass correlation coefficients were 0.88 (0.64–0.92) and 0.94 (0.88–0.98), respectively. The mean intra- and interobserver differences in RVLS_FW were −0.87 ± 0.29 (limits of agreement, −1.93% to 3.69%) and −0.88 ± 0.37 (limits of agreement, −2.26% to 4.3%), respectively. For LV GLS, the mean intra- and interobserver differences were −0.71 ± 0.28 (limits of agreement, −2.55% to 2.35%) and −0.73 ± 0.28 (limits of agreement, −1.96% to 3.03%), respectively.
Results
Clinical and Echocardiographic Characteristics of Survivors versus Patients with Cardiovascular Death, Arrhythmia, and Heart Failure
After excluding patients whose images were of suboptimal quality ( n = 17 [7.4%]), those who were lost to follow-up ( n = 3 [1.3%]), and those who died of sepsis and bleeding-related complications, a total of 208 patients (mean age, 63.8 ± 10.2 years; 161 men) were entered into the analysis ( Supplemental Figure 1 ). Among them, the incidence rates of left main, left coronary artery, and right coronary artery (RCA) lesions were 20.2%, 76.4%, and 62.5%, respectively. The median follow-up duration for the entire patient population was 23.1 ± 6.6 months. During follow-up, 27 patients (12.98%) died of cardiovascular causes, including myocardial infarction ( n = 12 [44.4%]), heart failure ( n = 7 [25.9%]), and arrhythmia ( n = 8 [29.62%]). These patients had significantly higher percentages of left main or RCA lesions, hospitalization for exacerbated heart failure, and new-onset arrhythmia compared with survivors ( Table 1 ). In echocardiographic analysis, cardiovascular mortality was significantly associated with a lower LVEF and a greater reduction in the amplitude of LV GLS or RVLS_FW ( Table 2 ). In addition, compared with the gender- and age-matched control subjects ( n = 50), patients with non-ACS angina (both survivors and nonsurvivors) had reduced amplitudes of both LV GLS and RVLS_FW.
