Cardiac Strain as a Predictor of Adverse Events and Ventricular Remodeling: A Cohort Study

Background: It remains controversial whether cardiac strain accurately predicts adverse events after acute ST-segment elevation myocardial infarction (STEMI). The aim of the present study was to evaluate the effects of cardiac strain revealed on cardiac magnetic resonance (CMR) imaging on cardiac events and adverse left ventricular (LV) remodeling. Methods: Between February 2015 and September 2016, we conducted a prospective two-center cohort study of patients with STEMI treated with primary percutaneous coronary intervention comprising stent implantation. All included patients underwent CMR imaging before discharge. Major adverse cardiac events (MACE) and LV remodeling were assessed during 6 months of follow-up. Results: Seventy-six patients were available for the nal analysis. The MACE rate was 23.7%, using cardiac death, reinfarction, unplanned revascularization, and heart failure as combined events during 6 months of follow-up. The global longitudinal strain (GLS) was an independent predictor of MACE (OR=1.21 (1.07–1.36), P=0.002) and LV remodeling (OR=2.06 (1.14–3.73), P=0.017). Conclusion: In patients with STEMI treated with primary percutaneous coronary intervention, the GLS determined on CMR imaging performed before discharge is a predictor of MACE and adverse LV remodeling during 6 months of follow-up.

The objective of the present study was to assess the effects of CMR-determined cardiac strain on cardiac events and adverse LV remodeling. We hypothesized that cardiac GLS measured before discharge is a good predictor of cardiac events and adverse LV remodeling after revascularization for STEMI. To test this hypothesis, the present prospective two-center study evaluated patients with STEMI treated with primary percutaneous coronary intervention (PCI) comprising stent implantation. These patients underwent CMR imaging before discharge. Follow-up data at 6 months after treatment were obtained from hospital records or face-to-face visits. The study ndings will aid in the prediction of adverse events after revascularization in patients with STEMI. Participants: The study cohort comprised 86 patients with STEMI treated with primary PCI comprising stent implantation in one of the two participating centers from February 2015 to September 2016. All patients underwent CMR imaging before discharge.

Methods
Exclusion criteria were: 1) atrial brillation, frequent premature contractions, persistent ventricular tachycardia, or other tachyarrhythmia; 2) previous cardiac surgery or myocardial infarction; 3) severe liver and/or kidney dysfunction; 4) malignancy; 5) life expectancy of less than 1 year; 6) pregnancy; and 7) contraindications to magnetic resonance imaging (e.g. contrast agent allergy, ferromagnetic objects in the body, claustrophobia). CMR Imaging Measurements: CMR imaging was performed before discharge (generally 5-7 days after the index event). All patients were examined with a 1.5 Tesla GE magnetic resonance imaging scanner. Three long-axis views (four-, three-, and two-chamber orientation) and short-axis stacks were acquired using a balanced steady-state free-precession imaging technique for functional cardiac analyses. Native T2, T2-weighted, and post-contrast T1-weighted image sequences were used for the assessment of edema, infarction size, microvascular obstruction (MVO), and intramyocardial hemorrhage. T1-weighted images were obtained 15 minutes after the administration of gadolinium-based contrast agent.
CMR Imaging Analysis: The analysis was performed o ine by two experienced radiologists. Infarct size, edema, MVO, and intramyocardial hemorrhage were quanti ed using CVI 42 software (Circle Cardiovascular Imaging Calgary, Canada) (13). CMR-FT strains (GLS, global circumferential strain (GCS), and global radial strain (GRS), LV end-diastolic volume (LVEDV), LV end-systolic volume, and LV ejection fraction (LVEF) were determined using the TomTec Imaging System (2D CPA MR, Cardiac Performance Analysis, version 1.1.2, TomTec Imaging Systems, Germany) (14,15). Brie y, LV contours were drawn semi-automatically at the end of diastole and systole. Subsequently, image features throughout an entire cardiac cycle were determined by the automatic border tracking algorithm of the software. Accurate tracking was con rmed by visual review of all borders and manual adjustments with consequent reapplication of the algorithm if necessary.
Follow-up Examination: The incidence of major adverse cardiac events (MACE), including cardiac death, reinfarction, unplanned revascularization, and heart failure within 6 months after STEMI was obtained from hospital records or face-to-face visits. Heart failure manifestations were defined as the exacerbation of exertional dyspnea or pulmonary edema requiring hospital admission, initiation of diuretics, or an increase in an existing diuretic regimen. Follow-up CMR imaging was performed at 6 months after STEMI. Adverse LV remodeling was de ned as an LVEDV of > 15% greater than the LVEDV before discharge from the hospital.
Statistical Analysis: Variables are denoted as mean ± standard deviation, and the independent t test or Fisher exact test was used to compare differences between groups. Variables that were not normally distributed (as determined by Kolmogorov-Smirnov tests) were expressed as medians with 25th and 75th percentiles, and were compared using the Mann-Whitney U test. Based on the ratio of the infarcted myocardial mass to the LV mass (IM%LV), patients were divided into group A (IM%LV < 10%), group B (10% ≤ IM%LV < 20%), and group C (IM%LV ≥ 20%).
A comparison of multiple variables was performed between patients with LV remodeling and patients without LV remodeling, and between patients who did and did not develop MACE during follow-up. Uniand multivariate logistic backward stepwise regression analyses were performed to evaluate the potential correlations between clinical parameters and CMR imaging parameters to MACE and LV remodeling.
Because of the small sample size, the parameters with a p value < 0.05 in comparisons between the MACE and no MACE groups (and between the LV remodeling and no LV remodeling groups), sex, and age were selected for the logistic regression analysis. If there were multiple parameters with high correlations, only the most clinically signi cant parameter was selected for the analysis. For example, as LVEDV, LV end-systolic volume, and LVEF were highly correlated, only LVEF was used in the analysis. Receiver operating-characteristic (ROC) curve analysis was used to determine the cutoff values of the GLS for predicting MACE. All statistical analyses were performed with a test signi cance level of 0.05 using SAS version 9.4 (SAS Institute, Inc., Cary, NC).

Results
The follow-up analysis was carried out in 76 patients (age 55.5 ± 10.7 years; 88% male) who were treated with primary PCI for STEMI and underwent CMR imaging examination before discharge, as shown in Fig.   1. Baseline patient characteristics are presented in Table 1. The most commonly accessed vessel was the left anterior descending artery, followed by the right coronary artery and left circum ex artery. The baseline characteristics that differed between patients with different degrees of myocardial infarction were the peak brain natriuretic peptide (BNP) level, peak cardiac troponin I level, and symptom-to-balloon time. Table 2 lists the cardiac characteristics obtained from CMR imaging at baseline. Group C had the lowest absolute GLS, GCS, and GRS values and the lowest LVEF, while group A had the highest values. Group C also had the worst features regarding the other variables. In 6 months of follow-up, MACE occurred in 18 patients (23.7%), including one patient with cardiac death, one with non-fatal reinfarction, four with unplanned revascularization, and 12 with heart failure. Patients with a higher IM%LV had a higher incidence of MACE than those with a lower IM%LV. Table 3 summarizes the baseline clinical and CMR imaging characteristics of the patients with MACE compared with those without MACE. Patients with MACE had a higher peak BNP level, higher peak cardiac troponin I level, longer hospital stay, longer symptom-to-balloon time, and worse CMR parameters than those with no MACE.
Univariate logistic regression analysis revealed that the variables predicting MACE were peak BNP level, LVEF, IM%LV, MVO, GLS, and GCS. Backward stepwise multivariate analysis confirmed that GLS was an independent predictor of MACE (OR=1.21 (1.07-1.36); P=0.001; Table 4). Figure 2 shows the ROC curve of the GLS. The area under the ROC curve was 0.763. The best cutoff value of GLS for predicting MACE was -14.6%, with a diagnostic sensitivity of 72.2% and a diagnostic specificity of 74.2%.
CMR imaging was performed at 6 months after STEMI in 24 patients (age 54 ± 11 years; 88% male). CMR imaging parameters showed improved cardiac function at 6 months after STEMI treatment compared with baseline (Table 5). Table 6 shows a comparison of the clinical and CMR imaging characteristics at baseline in patients with LV remodeling versus patients without LV remodeling on follow-up examination.
Univariate logistic regression analysis revealed that the variables predicting LV remodeling were symptom-to-balloon time and GLS. Backward stepwise multivariate analysis confirmed that GLS was an independent predictor of LV remodeling (OR=2.06 (1.14-3.73); P=0.017; Table 7).

Discussion
The present study showed the usefulness of GLS in the prediction of MACE and LV remodeling after PCI in patients with STEMI. After adjustments for clinical and morphometric parameters, the CMR-determined GLS before discharge was independently associated with adverse remodeling and outcomes at 6 months after STEMI treatment.

Relationship between GLS and MACE
The present study reported a MACE rate of 23.7%, using cardiac death, reinfarction, unplanned revascularization, and heart failure as combined events. Previous studies have reported similar MACE rates of 22% using cardiac death and heart failure as combined events (16), and 21% using cardiac death, acute myocardial infarction, and heart failure as combined events (17).
When measured soon after revascularization, the LVEF is a proven predictor of poor outcome in patients with myocardial infarction (18)(19)(20). However, the LVEF is a global parameter that represents the entire LV function and is thus a weak predictor of late myocardial dysfunction (21,22). Assessments of the myocardial strain in the circumferential, longitudinal, and radial directions (i.e., GLS, GCS, and GRS, respectively) are sensitive markers of intrinsic myocardial function, providing an improved analysis of cardiac dysfunction early after myocardial infarction on local and global levels (7,23). Myocardial strain assessed using the speckle tracking echocardiography technique accurately predicts adverse events (24)(25)(26). Based on the gold standard CMR imaging measurements (27)(28)(29), the present study showed that GLS was an independent predictor of MACE, with an optimal cutoff value of -14.6%. Similarly, previous studies have reported that GLS is a strong and independent predictor of adverse events (30,31), and a study of 659 patients with acute myocardial infarction demonstrated that a GLS value of greater than -15.1% was an independent predictor of cardiovascular events, either by combining all events or separating these events into mortality, reinfarction, revascularization, and hospitalization for heart failure (30).

Relationship between GLS and LV remodeling
Even after PCI, adverse LV remodeling occurs in 30-35% of patients with STEMI (32), and is an important predictor of arrhythmias, heart failure, and mortality (33,34). In the present study, the rate of adverse LV remodeling at 6 months after STEMI treatment was 29%, which is similar to the rate reported in a previous study (35). Although the LVEF is routinely used to assess LV systolic function, it cannot predict the development of LV remodeling. GLS is reportedly an independent predictor of adverse LV remodeling (16,30,35). The present study con rmed that GLS was an independent predictor of adverse LV remodeling. Similarly, a previous prospective study identi ed GLS as a strong predictor of clinical outcomes and an independent predictor of MACE and LV remodeling in multivariable logistic regression analysis after adjustment for other established prognostic risk factors, including the LVEF and infarct size (11,12).

Study Limitations
The present study has several limitations. First, the sample size was relatively small and was limited to patients with STEMI treated with primary PCI comprising stent implantation in two centers. Hence, selection bias and low statistical power should be considered when interpreting the ndings. Second, patients with cardiogenic shock and those requiring mechanical ventilation or intra-aortic balloon counter-pulsation therapy were not included. Third, the echocardiographic GLS was not measured, and so could not be compared with the CMR-derived GLS. Finally, heart rate and blood pressure, which influence strain computation, were not available for all patients when underwent CMR imaging.

Conclusions
In patients with STEMI treated with primary PCI, the CMR-determined GLS before discharge was a predictor of MACE and adverse LV remodeling at 6 months after STEMI treatment. Hence, GLS has potential as a risk factor to quantify ventricular dysfunction.

Declarations
Ethics approval and consent to participate: The study was approved by the Institutional Review Board of each participating center, and conformed to the Declaration of Helsinki and Good Clinical Practice Guidelines of the China Food and Drug Administration. All patients provided written informed consent for study participation.

Consent for publication: Not applicable.
Availability of data and materials: The datasets used and/or analyzed in the current study are available from the corresponding author on reasonable request.
Competing interests: The authors declare that they have no competing interests.
Authors' contributions: Yanjun Gong and Yuan Lu assessed the patients for study eligibility, performed data analysis, and wrote and edited the manuscript. Jessica C. Huo polished the language of the manuscript. Zhi Wang, Fan Yang, and Lin Qiu helped perform the patient follow-up and data collection. Shu Fang edited the manuscript. Jianxing Qiu performed cardiac magnetic resonance imaging. Yong Huo helped in designing the study, performing data analysis, and editing the manuscript.       Figure 1 Flowchart of the study. CMR: cardiac magnetic resonance Receiver operating-characteristic curve for the prediction of major adverse cardiac events within 6 months after acute ST-segment elevation myocardial infarction, using the global longitudinal strain as the independent variable.