© 2014, Wiley Periodicals, Inc. DOI: 10.1111/echo.12864

Echocardiography

ORIGINAL INVESTIGATION

Prognostic Value of Speckle Tracking Echocardiography in Patients with ST-Elevation Myocardial Infarction Treated with Late Percutaneous Intervention Tao Cong, M.D., Yinghui Sun, M.D., Zhijuan Shang, M.D., Ke Wang, M.D., Dechun Su, M.D., Lei Zhong, M.D., Shulong Zhang, M.D., and Yanzong Yang, M.D. The First Affiliated Hospital of Dalian Medical University, Liaoning, China

Background: Left ventricular remodeling (LVr) is common after ST-segment elevation myocardial infarction (STEMI). The aim of this study was to evaluate the prognostic value of speckle tracking echocardiography (STE) for predicting LVr 6–9 months after late percutaneous coronary intervention (PCI) in patients with STEMI. Methods: Patients with first STEMI who accepted late PCI were enrolled. Echocardiography was performed within 48 hours of admission. Six to nine months after MI, an echocardiography examination was repeated. LVr was defined as >15% increase in LV end-systolic volume (LVESV) after 6 months. Results: One hundred and twenty-seven patients were divided into two groups: 86 patients without LVr and 41 patients with LVr. There were significant differences in the global longitudinal strain (GLS), SD of time to peak longitudinal systolic strain (longitudinal Ts-SD), longitudinal postsystolic index, radial strain (RS), and SD of time to peak radial systolic strain (Radial Ts-SD). In multivariate logistic regression analysis, the GLS(odds ratio [OR] = 0.39, 95% confidence interval [CI] = 0.26–0.57, P < 0.01), and RS(OR = 1.07, 95% CI = 1.02–1.13, P = 0.01) were determinants of LVr. A receiver operating characteristic curve showed that the GLS predicted LVr with an optimal cutoff value of 10.85 (sensitivity: 89.7%, specificity: 91.7%). During clinical follow-up for 16.9  1.6 months, death or congestive heart failure developed in 12 patients (9.4%), and the baseline ejection fraction (OR = 1.91, 95% CI = 1.18–3.1, P = 0.009) and GLS (OR = 0.56, 95% CI = 0.34– 0.91, P = 0.02) were independent predictors. Conclusion: In patients with STEMI treated with late percutaneous coronary intervention, the GLS as measured by STE is a strong predictor of LVr and adverse events. (Echocardiography 2014;00:1–8) Key words: speckle tracking echocardiography, remodeling, myocardial, infarction, strain

Acute myocardial infarction (AMI) results in complex alterations in ventricular architecture involving both infarcted and noninfarcted zones, which can lead to remodeling of LV, particularly in patients who received late percutaneous coronary intervention (late PCI).1 Left ventricular remodeling (LVr) has profound effects on the function of the ventricle and thus the prognosis of patients.2,3LVr develops in a considerable proportion (30%) of patients after AMI despite successful treatment with primary percutaneous coronary intervention (pPCI),4 and LVr begins in the acute phase of myocardial ischemia. There is a change in the myocardial wall structure, includAddress for correspondence and reprint requests: Yanzong Yang, M.D., The First Affiliated Hospital of Dalian Medical University, Dalian, Liaoning 116000, China. Fax: 0860411-84394085; E-mail: [email protected]

ing wall thinning, elongation, and progression toward hypertrophy and dilatation.2 Changes in LV geometry may lead to heart failure and lifethreatening arrhythmias, thus increasing of the mortality rate.5 Therefore, the early identification of patients at risk for developing LVr after AMI has important prognostic and therapeutic implications. Traditional echocardiography is widely used for the prediction of LVr, but its accuracy is low.6–8 Perfusion echocardiography is a noninvasive tool used to assess myocardial viability in the region of infarction, but it cannot be widely used because of the expensive contrast media and extensive sonographer experience required.9 In recent years, speckle tracking echocardiography (STE) was used to predict LVr or heart failure after AMI, and although its value is reliable, studies continue to investigate which indicator is more 1

Cong, et al.

powerful.10–13 Until now, few researches have examined the value of postsystolic motion for predicting LVr after AMI. In China, only 30% of patients of AMI can accept pPCI because of the limited medical condition (part of the municipal hospitals have not the qualification for PCI)14; thus, many patients undergo late PCI, and there is no known manner in which to predict LVr for AMI patients who underwent late PCI. The aim of this study was to investigate the value of STE for predicting LVr after AMI treated with late PCI and to assess which indicator, including postsystolic motion, is more powerful for predicting LVr in patients with AMI who underwent late PCI. Methods: This is an observational prospective cohort study in which we enrolled patients diagnosed with primary ST-segment elevation myocardial infarction (STEMI) (59.9  11.6 years; 103 men). The criteria for enrollment were as follows: STEMI with an onset of chest pain at 120 msec, other than sinus heart rhythm, active inflammatory diseases including pneumonia, hepatitis, nephritis, etc., abnormal hormone status including hyperthyroidism or hypothyroidism, etc., and poor echocardiographic conditions for analyzing the results of STE were excluded. All patients were given loading doses of aspirin, 300 mg of clopidogrel, and 100 IU/kg of heparin (maximum 5000 IU). The use of abciximab was optional, and the decision was left to the invasive cardiologist. Blood samples for cardiac enzymes and troponin I were repeated using immunoinhibition and immunoenzymatic quantitative methods every 6 hours until peak levels were identified. Twelve-lead electrocardiography was performed before and 60 minutes after late PCI for the infarct-related artery, and we analyzed the sum of the elevation in ST segments in all leads. Echocardiography: Baseline echocardiography was performed within 24 hours of admission, and follow-up echocardiography was performed after 6–9 months. Twodimensional echocardiography was performed, and the STE results were assessed. Echocardiography was performed using the Vivid 7 system (GE Vingmed Ultrasound AS, Horten, Norway). Three apical scans of the left ventricle in the four-chamber, three-chamber, and two-chamber 2

views were obtained according to the guidelines of the American Society of Echocardiography.15 A 16-segment model of the left ventricle was used for wall-motion score,16 and segments were graded (normokinetic = 1, hypokinetic = 2, akinetic = 3, dyskinetic = 4) according to wallmotion amplitude and changes in LV thickness at systole. The wall-motion score index was defined as the sum of the segment score ratings divided by the number of segments scored. The LV mass was calculated by M-mode in the parasternal long-axis view. The LV end-diastolic volume, endsystolic volume, and ejection fraction (EF) were calculated using a modified Simpson’s method. The ratio of the mitral inflow E velocity to the Doppler e’ velocity from the lateral annulus (E/e’) was measured. The mitral regurgitant jet area was measured with Color Doppler using the apical four-chamber views. STE: 2D speckle tracking analysis was performed on a quantitative workstation using commercial software (EchoPAC version 6.00, GE Medical Systems, Milwaukee, WI, USA). Images from apical four-chamber, three-chamber, and two-chamber views and the parasternal short-axis view at the papillary muscle level were analyzed. The endocardial and epicardial LV borders were manually traced initially at end-diastole with avoidance of papillary muscles and trabeculations. The software then automatically tracked and delineated the endocardial and epicardial borders throughout the cardiac cycle. The LV myocardium was automatically divided into 6 regions of interest representing 6 segments in each view. Manual adjustment of segmental tracking points was performed if necessary. After achieving satisfactory tracking approved by the observer, the software provided the strain value for each segment and the global strain (longitudinal strain from the apical view and radial strain (RS) from the parasternal view). The LV global longitudinal strain (GLS) was calculated as an averaged value from the apical four-chamber, three-chamber, and twochamber views (Fig. 1), and left ventricular RS curves were acquired by the mid-short-axis views. The frame rate of images for speckle tracking analysis was in the range of 50–90 frames per second. The time to peak systolic strain for each segment was automatically measured, and the SD of time to peak systolic strain for the 18 longaxis segments or 6 short-axis segments was calculated to assess left ventricular synchrony. The postsystolic index (PSI) was defined as the postsystolic strain, the peak longitudinal systolic strain was defined as the maximal negative peak in the strain curve during systole, and the longitudinal postsystolic strain was measured as

STE for ST-Elevation Myocardial Infarction

Figure 1. Curves of the longitudinal strain of the left ventricle in four-chamber projections of a patient with acute myocardial infarction (AMI).

the negative peak during isovolumic relaxation time up to 50–70 msec after mitral valve opening.17 Long-Term Follow-up: Six to nine months after STEMI, two-dimensional echocardiography was performed. LV remodeling was defined as a >15% increase in LVESV. Analysis of all of the echocardiographic data was performed by the same cardiologist blinded to the patient identity and clinical status. Clinical follow-up data were obtained 16.9  1.6 months (13.0–21.0 months) from AMI, the occurrence of clinical events of death or heart failure was recorded by attending cardiologists. Development of heart failure was defined as worsening of exertional dyspnea with typical chest radiograph findings of progressive cardiomegaly or pulmonary edema that required hospital admission or the administration of new or different diuretics. Statistical Analysis: Continuous variables are expressed as the mean  SD, and categorical variables are expressed as absolute values and percentages. Normal distribution of the data was verified by the Kolmogorov–Smirnov test. Differences in continuous variables between groups were analyzed

by the unpaired or paired Student’s t-test. Categorical variables were compared using the Chisquare or Fisher’s exact tests where indicated. All tests were two-sided, and statistical significance was defined as P < 0.05. Variables with a P-value less than 0.05 between patients who did and did not develop LV remodeling or adverse clinical events were included in multivariable logistic regression analysis to identify factors with independent associations with LVr or adverse events (anterior infarction, time from AMI to intervention, ΣST before PCI, ΣST post-PCI, peak CKMB, peak troponin I, baseline LVESV, baseline LVEF, wall-motion score index (WMSI), GLS, longitudinal Ts-SD, longitudinal PSI, RS, and radial Ts-SD). We performed the multiple logistic regression analyses using stepwise forward elimination. A ROC curve was used to determine the optimal GLS, RS value, and baseline EF for predicting LVr after 6–9 months or clinical adverse events. To evaluate the reliability of echocardiographic results, intra-observer and inter-observer variability was assessed. Twenty subjects were randomly chosen for that analysis, The intraclass correlation coefficient (ICC) was calculated. ICC was calculated for GLS and RS assessment by STE. Intra-observer ICC was 82% (95% CI, 59.9– 92.4%) for GLS and 85% (95% CI, 66–94%) for RS, and inter-observer ICC was 83% (95% CI, 3

Cong, et al.

63–93%) and 77% (95% CI, 50–90%), respectively. Data were collected and analyzed using statistical software (SPSS 13.0 for Windows, SPSS Inc., Chicago, IL, USA). Results: One hundred thirty patients satisfied the baseline inclusion criteria. Six patients were excluded, three patients were disqualified from STE because of technically inadequate image quality, and three patients did not have greater than 6 months follow-up; thus, echocardiographic re evaluation could not be performed in these patients. In total, 127 patients with complete clinical and echocardiographic follow-up constituted the study population. The baseline demographic and clinical characteristics of patients included in this study are presented in Table I. Anterior wall STEMI was diagnosed in 52 patients (40.9%), and the mean time from AMI to late percutaneous coronary intervention was 6.6  2.0 days. TABLE I Characteristics of the Study Population (n = 127) Parameter Male (%) Age (years) Diabetes (%) Hyperlipidemia (%) Hypertension (%) Smoking (%) Anterior infarction (%) Time from AMI to intervention (days) Maximum troponin (ug/L) BUN (mmol/L) Cr (umol/L) AST (U/L) ALT (U/L) ΣST before PCI (mm) ΣST 60 min after PCI (mm) LVEDV (ml) LVESV (ml) LVEF (%) GLS (%) Longitudinal Ts-SD (msec) Longitudinal PSI RS (%) Radial Ts-SD (msec) WMSI

Value 81.1 59.9  11.6 26.8 30.7 47.2 67.7 40.9 6.6  2.0 73.9 6.97 78.69 209.01 56.43 3.2 2.6 104.4 50.2 51.8 12.3 59.8 1.6 33.1 44.1 1.5

               

102.2 2.0 33.86 150.15 14.04 1.4 1.1 18.5 10.4 5.1 3.5 18.6 0.8 17.7 36.1 0.2

Data are expressed as percentages or mean  SD. BUN = blood urea nitrogen; Cr = creatinine; AST = aspartate aminotransferase; ALT = alanine aminotransferase; LVEDV = LV end-diastolic volume; LVESV = LV end-systolic volume; LVEF = LV ejection fraction; ΣST = sum of ST-segment elevations in all leads; GLS = global longitudinal strain; Ts-SD = standard deviation of time to peak systolic strain; PSI = postsystolic index; RS = radial strain; WMSI = wall-motion score index.

4

Variables Associated with LV Remodeling: After 6–9-month follow-up, the study population was divided into two groups according to increases in the LVESV. Those developing a >15% increase in LVESV were described as the remodeling group, and the remaining patients were described as the nonremodeling group. A total of 41 (32%) patients developed LV remodeling according to the definition. A comparison of the remodeling and nonremodeling groups is shown in Table II. There were no significant differences between the two groups in demographic variables. A hypertension history was more common in the remodeling group (58.5% vs. 41.9%, P = 0.08), but the difference did not reach statistical significance. Patients in the remodeling group had a greater incidence of anterior infarction (56.1% vs. 33.7%, P = 0.02), longer time from AMI to intervention (7.1  2.2 vs. 6.3  1.8 days, P = 0.02), more ΣST before and post-PCI (3.9  1.3 vs. 2.8  1.3 mm, P < 0.01; 3.2  1.0 vs. 2.3  1.0 mm, P < 0.01), higher CKMB and maximum troponin (303.8  252.1 vs. 141.2  119.5, P < 0.01; 121.7  145.7 vs. 51.1  62.3, P < 0.01) compared with the nonremodeling group. About LV function, Patients in the remodeling group had lower LVEF, higher LVESV and WMSI (48.4  4.6 vs. 53.4  4.5, P < 0.01; 53.4  11.4 vs. 48.7  9.6, P = 0.02; 1.6  0.1 vs. 1.5  0.1, P = 0.03). About STE, Patients in the remodeling group had lower GLS, RS (8.7  2.0 vs. 14.0  2.7, P < 0.01; 21.4  13.1 vs. 38.5  17.0, P < 0.01) and higher Radial Ts-SD, Longitudinal PSI, Longitudinal Ts-SD (62.8  41.7 vs. 35.6  29.8, P < 0.01; 2.3  0.9 vs. 1.2  0.6, P < 0.01; 70.4  18.9 vs. 54.7  16.2, P < 0.01). Predictors of Left Ventricular Remodeling: To determine independent predictors of LVr, stepwise logistic regression analysis was performed. Variables that were statistically significant (incidence of anterior infarction, time from AMI to intervention, ΣST before PCI, ΣST postPCI, CKMB, maximum troponin, baseline LVESV, LVEF, WMSI, GLS, longitudinal Ts-SD, longitudinal PSI, RS, and radial Ts-SD) upon analysis were entered into the equation. Independent predictors of remodeling were found to be GLS, RS (Table III). A ROC curve was used to calculate the optimal value for the GLS, RS to predict LV remodeling: GLS: 10.85% (sensitivity: 89.7%, specificity: 91.7%); RS: 28.46% (sensitivity: 82.1%, specificity: 66.7%) (Fig. 2). Clinical events during follow-up: During the clinical follow-up of 16.9  1.6 months, 2 patients died as a result of heart failure 10 months after the initial event. The

STE for ST-Elevation Myocardial Infarction

TABLE II Characteristics of Patients with and without LVr Parameter

Nonremodeling

Remodeling

P-Value

Number (%) Age (years) Male (%) BMI (Kg/m2) Hypertension (%) Diabetes (%) Dyslipidemia (%) Smoking (%) Anterior infarction (%) Time from AMI to intervention (days) ΣST before PCI (mm) ΣST post-PCI (mm) CKMB (ug/L) Maximum troponin (ug/L) BUN (mmol/L) Cr (umol/L) AST (U/L) ALT (U/L) ACE-inhibitor/ARB (%) b-blocker (%) Statin (%) Antisterone (%) LVEDV (ml) LVESV (ml) LVEF (%) LV mass E/e’ MR(cm2) WMSI GLS (%) Longitudinal Ts-SD (msec) Longitudinal PSI RS (%) Radial Ts-SD (msec)

86 (68) 59.4  11.7 83.7 25.4  3.5 41.9 27.9 33.7 70.9 33.7 6.3  1.8 2.8  1.3 2.3  1.0 141.2  119.5 51.1  62.3 6.54  1.53 74.72  18.38 185.59  151.65 56.22  13.30 98 100 100 9 104.3  18.0 48.7  9.6 53.4  4.5 201.7  40.6 10.5  2.8 1.3  0.7 1.5  0.1 14.0  2.7 54.7  16.2 1.2  0.6 38.5  17.0 35.6  29.8

41 (32) 60.9  11.6 75.6 25.9  2.9 58.5 24.4 24.4 61.0 56.1 7.1  2.2 3.9  1.3 3.2  1.0 303.8  252.1 121.7  145.7 7.89  2.52 87.14  53.92 259.2  138.59 56.87  16.0 100 98 100 20 104.7  19.7 53.4  11.4 48.4  4.6 215.4  32.7 11.3  2.0 1.5  0.5 1.6  0.1 8.7  2.0 70.4  18.9 2.3  0.9 21.4  13.1 62.8  41.7

0.51 0.28 0.36 0.08 0.68 0.29 0.26 0.02 0.02