Intramyocardial hemorrhage (IMH) and microvascular obstruction (MVO) are two major mechanisms of reperfusion injury of the left ventricle after acute ST-segment elevation myocardial infarction (STEMI). The aim of this study was to assess the impact of IMH and MVO on left ventricular (LV) cardiac mechanics using two-dimensional speckle-tracking echocardiography during the acute phase of STEMI and on LV functional recovery.
Eighty-one patients with STEMI who received primary reperfusion therapy were prospectively studied. Infarct segments were classified by cardiac magnetic resonance according to infarct transmurality and the presence or absence of IMH and/or MVO. Segmental systolic longitudinal strain, circumferential strain (CS), and radial strain were measured by two-dimensional speckle-tracking echocardiography. Adverse LV remodeling and major adverse cardiovascular events were assessed at 1 year.
MVO without IMH was much less frequent in nontransmural infarct segments than in transmural infarct segments (6.0% vs 19.1%, P = .000), while IMH was present only in transmural infarct segments. In nontransmural infarct segments, MVO was not associated with any significant changes in strain ( P > .5). In transmural infarct segments, there were no differences in all types of strain between segments without reperfusion injury and those with MVO alone ( P > .20). IMH was evident in the midmyocardial layer within the infarct zone in 196 segments (46.1%). The presence of IMH in addition to MVO decreased CS significantly ( P = .004), but not longitudinal and radial strain ( P > .5). A receiver operating characteristic curve analysis with cross-validation by k-folding showed that the sensitivity and specificity of CS using a cutoff of >−11.66% to diagnose IMH were 78.00% and 79.45%, respectively (area under the curve = 0.86; P = .0001). At 1 year, patients with major adverse cardiovascular events and LV remodeling had significantly lower baseline measurements of all types of global strain ( P < .05).
In the acute phase of STEMI, reperfusion MVO and IMH injury have differential effects on cardiac mechanics. IMH preferentially affects CS, presumably related to its location in the midmyocardial layer.
Intramyocardial hemorrhage (IMH) is an important mechanism of reperfusion injury after acute ST-segment elevation myocardial infarction (STEMI). It is closely related to the phenomenon of microvascular obstruction (MVO). Both MVO and IMH are associated with adverse left ventricular (LV) remodeling, diminished recovery of LV function, and poor long-term prognosis after revascularization treatment for acute STEMI. However, the interaction of IMH and MVO on contractile function and recovery of the infarct segments with reperfusion injury has not been clearly defined.
Both IMH and MVO can be detected by cardiac magnetic resonance (CMR). Two-dimensional (2D) speckle-tracking echocardiography (STE) is a novel imaging technique that detects and tracks acoustic markers frame by frame on grayscale 2D images, allowing relatively angle-independent analysis of myocardial deformation in multiple directions, including longitudinal strain (LS), circumferential strain (CS), and radial strain (RS), both globally and of individual myocardial segments. Previous studies using CMR assessment showed that global CS and LS in the acute phase of STEMI are closely related to infarct size. More recent studies have demonstrated that 2D speckle-tracking echocardiographic assessment of CS is useful in distinguishing transmural from subendocardial infarction. Moreover, 2D speckle-tracking echocardiographic evaluation of the presence of MVO and its impact on myocardial function was feasible. Despite the prognostic significance of IMH and MVO after reperfusion injury, their interactions in relation to their impact on LV mechanics have not been systematically evaluated. We hypothesized that in the acute phase of STEMI, (1) reperfusion injury, which was defined as MVO or IMH in this study, is associated with altered segmental and global strains, in addition to myocardial infarction (MI) transmurality; (2) MVO and IMH have different influences on various strain components; and (3) reduced strain at baseline is associated with adverse myocardial functional recovery and outcome.
We prospectively screened patients admitted to our hospital for acute STEMI from September 2012 to December 2013. STEMI was defined as (1) chest pain and (2) new ST-segment elevation at the J point in two anatomically contiguous leads using the following diagnostic thresholds: ≥0.1 mV (1 mm) in all leads other than V 2 and V 3 , where the following diagnostic thresholds apply: ≥0.2 mV (2 mm) in men ≥40 years of age, ≥0.25 mV (2.5 mm) in men <40 years of age, and ≥0.15 mV (1.5 mm) in women. A total of 260 patients were screened. The inclusion criteria were as follows: (1) symptom onset < 12 hours and (2) eligibility for primary reperfusion therapy. The exclusion criteria were as follows: (1) age > 75 years, (2) left bundle branch block and/or atrial fibrillation, (3) cardiomyopathy, (4) significant valvular heart disease, (5) prior MI, (6) contraindication to reperfusion therapy or CMR, (7) poor echocardiographic image quality, and (8) refusal to participate in the study. A total of 81 patients were included in the final analysis for this study. Selection for our study population is presented in the flowchart shown in Figure 1 . At 1 year after the index STEMI, occurrences of major adverse cardiovascular events (MACEs), including death, resuscitated cardiac arrest, and acute heart failure (with typical manifestations of pulmonary edema), were followed and recorded. Written informed consent was obtained from all patients. The study was approved by the ethics committee of the institution.
CMR Imaging and Analysis
All patients underwent CMR within 8 days (median, 4.9 days) after reperfusion therapy. Electrocardiographically gated CMR imaging was performed using a 3.0-T scanner (Achieva TX; Philips Healthcare, Best, The Netherlands). All sequences were acquired in breath-hold, with a field of view of 350 × 350 mm 2 .
Cine CMR was performed using a balanced steady-state free precession sequence in the short-axis view to cover the whole left ventricle without gap (repetition time, 3.2 msec; echo time, 1.6 msec; 30 phases; voxel size, 2.0 × 1.6 × 8 mm 3 ).
Three black-blood T2 short-tau inversion-recovery images were acquired at the apical, mid, and basal levels (repetition time, two R-R intervals; echo time, 75 msec; voxel size, 2.0 × 1.6 × 8 mm 3 ). Myocardial edema was defined as high-signal myocardium within the territory of the culprit vessel (signal intensity > 2 SDs above the mean signal in remote noninfarcted myocardium), and the hyposignal area within the edema area was recognized as IMH ( Figures 2 A–2C).
Immediately after a bolus intravenous administration of contrast agent (0.2 mmol/kg) (Magnevist; Bayer HealthCare Pharmaceuticals, Munich, Germany), first-pass perfusion was performed in the same slice locations and planes. Myocardial defect within the territory of the culprit vessel was defined as first-pass perfusion defect.
Late gadolinium enhancement using a three-dimensional inversion recovery segmented gradient echo sequence was performed 10 min after contrast injection in short-axis and two- and four-chamber views covering the whole left ventricle (repetition time, 3.5 msec; echo time, 1.7 msec; temporal resolution, 190 msec; voxel size, 1.5 × 1.7 × 10 mm 3 interpolated into 0.74 × 0.74 × 5 mm 3 ). Infarction was defined as hyperenhanced myocardium (signal intensity > 5 SDs of normal myocardium). An experienced reader who was blinded to the clinical data analyzed the CMR data using customized software (QMass MR version 7.5; Medis Medical Imaging, Leiden, The Netherlands). Infarct size was expressed as percentage necrosis of segmental volume for each of the LV segments. Nontransmural infarct (NTI) was defined as 1% to 50% extent of infarction in the radial direction at the segmental level, and transmural infarct (TI) was defined as 51% to 100% infarct extent. The total scar percentage was reported as the percentage of infarction to the total LV mass. MVO was defined as a hypoenhanced region in the infarct-related myocardium ( Figures 2 D–2F).
Myocardial segments were divided into six groups on the basis of CMR findings: (1) normal segments without infarction, (2) segments with NTI but no CMR evidence of reperfusion injury (NTI MVO−/IMH−), (3) NTI segments with MVO but no IMH (NTI MVO+/IMH−), (4) segments with TI but no reperfusion injury (TI MVO−/IMH−), (5) TI segments with MVO but no IMH (TI MVO+/IMH−), and (6) TI segments with IMH in addition to MVO (TI MVO+/IMH+). Patients were categorized into three groups according to the presence or absence of MVO and/or IMH in their infarct segments.
For CMR parameter analysis, including LV ejection fraction (LVEF), scar percentage, and the presence or absence of MVO and IMH, reproducibility was determined by repeating analysis on 20 randomly selected patients ≥1 week after the initial measurement by the same observer (for intraobserver variability) and a different observer (for interobserver variability).
Standard Echocardiography and STE
All patients underwent standard 2D and Doppler echocardiography at baseline using a 1.5- to 4.5-MHz probe (M5S, E9; GE Healthcare, Milwaukee, WI) on the same day as CMR. Left atrial volume, LV end-systolic volume, LV end-diastolic volume, and LVEF were assessed using the Simpson method, and segmental wall motion was graded as (1) normokinetic, (2) hypokinetic, (3) akinetic, or (4) dyskinetic according to the recommendations of the American Society of Echocardiography. Diastolic function was assessed using Doppler echocardiography of mitral inflow and Doppler tissue imaging of mitral annular velocity. At 1-year follow-up, transthoracic echocardiography was repeated to assess LV function. The infarct segment was considered to demonstrate functional improvement at 1-year follow-up if the segmental wall motion index improved by at least one point. LV remodeling was defined as an increase in LV end-systolic volume of >15% from baseline to 1-year follow-up.
Two-dimensional grayscale images of the left ventricle acquired from the three parasternal short-axis (basal, mid, and apical) and the three apical (four-chamber, two-chamber, and long-axis) views were used for offline speckle strain analysis. Images were optimized from frame rate (≥50 Hz) for three consecutive sinus beats during breath-hold. Digitally stored images were transferred to a workstation for offline analysis using dedicated software (EchoPAC PC version 110; GE Vingmed Ultrasound AS, Horten, Norway). For each of the three short-axis views, the sampling points were placed manually along the endocardium at the LV base, middle, and apex during end-systole. For apical two-, three-, and four-chamber views, three sampling points were placed manually at the septal mitral annulus, lateral corner, and apical endocardium. A region of interest was then generated by the software to cover the myocardial thickness along the entire LV wall. The region of interest was adjusted manually to ensure that the inner margin conformed to the whole LV endocardial border and that it included the entire thickness of the LV myocardium. The software subsequently identified the tissue speckles and tracked their movement frame by frame throughout the cardiac cycle. LV segmental LS, CS, and RS assessed by 2D STE were measured using the 18-segment LV model ( Figure 3 ). Global strain values were calculated from the average strain parameters of 18 segments. Patients were excluded from the calculation of global strain if more than two segments were unavailable for the acquisition of strain parameters.
The reproducibility of 2D speckle-tracking echocardiographic measurements was determined by repeating strain analysis on stored 2D data sets of 20 randomly selected studies ≥1 week after the initial measurement by the same observer (intraobserver) and a different observer (interobserver).
Statistical analyses were performed using SPSS version 17.0 (SPSS, Chicago, IL). Data are expressed as mean ± SD or as number of patients (percentage) as appropriate. Between-group comparisons used analysis of variance or the Fisher exact test as appropriate. Within-group comparisons of continuous variables between baseline and 1 year used paired t tests. Receiver operating characteristic curve analysis was used to assess the diagnostic performance of strain parameters in detecting IMH. The k-fold method was used to cross-validate receiver operating characteristic curves. Reproducibility was calculated as the absolute difference of strain measurements between the two observers divided by the mean of both measurements, expressed as a percentage. Intraobserver and interobserver intraclass correlation coefficients (ICCs) for quantitative measurements, including strain parameters and the measurement of CMR LVEF and scar percentage, were calculated in a two-way mixed model with 95% CIs. A two-sided P value < .05 was considered to indicate statistical significance. The reproducibility for the definition of MVO and IMH by CMR was tested using κ statistics.
Patient characteristics are summarized in Table 1 . Among the 81 subjects included in the final analysis, CMR demonstrated reperfusion injury in 58 patients. Twelve patients had MVO without IMH (MVO+/IMH− group), and 46 had MVO with IMH (MVO+/IMH+ group). Patients with reperfusion injury had higher levels of peak creatine kinase–MB, more infarct segments, larger scar percentages, lower LVEFs, higher values of wall motion score index, and larger left atrial volumes at presentation ( P < .05 for all), while other indexes of diastolic function did not differ among groups.
|MVO−/IMH− ( n = 23)||MVO+/IMH− ( n = 12)||MVO+/IMH+ ( n = 46)||P|
|Age (y)||57 ± 8||59 ± 6||57 ± 10||.719|
|Men||21 (91%)||12 (100%)||43 (93%)||.591|
|Hypertension||13 (57%)||6 (50%)||26 (57%)||.916|
|DM||8 (35%)||6 (50%)||18 (39%)||.68|
|Smoking||8 (35%)||6 (50%)||18 (39%)||.522|
|Hypercholesterolemia||9 (39%)||4 (33%)||30 (65%)||.041|
|CRF||1 (4%)||0 (0%)||4 (9%)||.49|
|LAD||12 (52%)||6 (50%)||27 (59%)|
|RCA||8 (35%)||5 (42%)||13 (28%)|
|LCX||3 (13%)||1 (8%)||6 (13%)|
|Multivessel disease||9 (39%)||4 (33%)||24 (52%)||.384|
|Time of ischemia (h)||5.44 ± 1.74||4.74 ± 2.28||5.81 ± 2.80||.429|
|Post-PCI TIMI flow grade 3||23 (100%)||11 (92%)||38 (83%)||.295|
|Peak CK-MB||258.00 ± 177.06||350.08 ± 233.15||418.13 ± 209.07||.013|
|Number of infarct segments||4.61 ± 2.46||6.33 ± 2.50||8.59 ± 2.80||<.001|
|Scar percentage (%)||11.26 ± 7.88||23.67 ± 11.95||30.87 ± 12.42||<.001|
|LVEF (%)||58.98 ± 9.37||51.08 ± 9.44||48.20 ± 10.48||<.001|
|Echocardiography and STE|
|WMSI||1.23 ± 0.26||1.39 ± 0.28||1.55 ± 0.37||.001|
|LA volume (mL)||60.70 ± 17.79||64.67 ± 15.29||72.78 ± 19.24||.032|
|E peak (cm/sec)||70.50 ± 13.18||77.75 ± 15.89||77.88 ± 18.92||.240|
|A peak (cm/sec)||72.00 ± 16.50||60.42 ± 16.72||71.51 ± 19.40||.150|
|E/A ratio||1.04 ± 0.34||1.41 ± 0.58||1.20 ± 0.53||.110|
|DT (msec)||162.26 ± 47.17||142.92 ± 38.73||138.83 ± 45.68||.140|
|E/e′ ratio||7.56 ± 2.53||8.09 ± 1.40||9.27 ± 3.66||.100|
|LVEF (%)||56.73 ± 10.68||52.08 ± 8.94||48.42 ± 9.10||.005|
|Global CS (%)||−20.45 ± 4.66||−14.92 ± 4.89||−14.14 ± 4.99||<.001|
|Global RS (%)||39.26 ± 15.22||35.08 ± 8.22||26.48 ± 8.23||<.001|
|Global LS (%)||−16.26 ± 4.91||−11.25 ± 5.84||−11.60 ± 5.76||.004|
CMR analysis of infarct transmurality and reperfusion injury was feasible in all 1,485 segments of 81 patients. Excellent intra- and interobserver reproducibility was achieved for CMR analysis: for LVEF, intraobserver ICC = 0.99 (95% CI, 0.98–1.00), interobserver ICC = 0.99 (95% CI, 0.97–0.99); for scar percentage, intraobserver ICC = 0.99 (95% CI, 0.98–1.00), interobserver ICC = 0.98 (95% CI, 0.96–0.99); for MVO, intraobserver κ = 1.00, interobserver κ = 0.983; and for IMH, intraobserver κ = 1.00, interobserver κ = 1.00. The mean infarct size was 24.6 ± 14.3% of LV mass. Of the 1,458 segments analyzed, there were 882 remote noninfarct segments (60.5%) (remote segments), 151 NTI segments (10.4%), and 425 TI segments (29.1%). In NTI segments, MVO was present in nine segments (NTI MVO+/IMH− group), and the rest ( n = 142) showed no reperfusion injury (NTI MVO−/IMH− group). IMH was absent in all NTI segments. In TI segments, 148 (34.8%) showed no reperfusion injury (TI MVO−/IMH− group), 81 (19.1%) had MVO without IMH (TI MVO+/IMH− group), and 196 (46.1%) showed MVO with IMH (TI MVO+/IMH+ group). Both MVO and IMH were located in the midmyocardial layer within the infarct zone. IMH was noted only in segments with MVO.
Analysis of LS was feasible in 1,430 of 1,458 segments (98%), while analysis of CS and RS was feasible in 1,324 of 1,458 segments (91%). The feasibility of global LS assessment was 95%, while that of global CS and global RS was 83%. At the patient level, the amplitudes of global strain measurements were lower in the MVO+/IMH− and MVO+/IMH+ groups compared with MVO−/IMH− group ( Table 1 ). At the segment level, peak systolic segmental CS in the normal, NTI, and TI groups was −20.83 ± 7.35%, −15.26 ± 7.03%, and −7.00 ± 8.46%, respectively ( P < .0001). Peak systolic segmental LS in the three groups was −17.38 ± 6.41%, −12.52 ± 5.82%, and −6.25 ± 6.91%, respectively ( P < .0001). The respective measurements for RS were 35.64 ± 20.62%, 31.78 ± 18.59%, and 19.00 ± 15.12% ( P < .0001). For segment-level subgroup analysis, peak systolic segmental CS in normal, NTI MVO−/IMH−, NTI MVO+/IMH−, TI MVO−/IMH−, TI MVO+/IMH−, and TI MVO+/IMH+ segments was −20.93 ± 7.45%, −15.47 ± 7.12%, −11.83 ± 4.24%, −10.16 ± 7.77%, −7.97 ± 7.12%, and −4.24 ± 8.58%, respectively. The corresponding figures for segmental LS were −17.53 ± 6.38%, −12.50 ± 5.92%, −12.46 ± 4.24%, −8.01 ± 6.80%, −6.12 ± 7.22%, and −4.99 ± 6.62%, and those for segmental RS were 36.01 ± 20.85%, 31.69 ± 18.79%, 33.21 ± 16.22%, 19.05 ± 14.75%, 19.12 ± 18.23%, and 18.92 ± 14.03% ( Figure 4 ). Within the NTI segment group, reperfusion MVO injury was uncommon (nine of 151 segments [6%]) and was not associated with any significant changes in strains ( P > .50 for all comparisons between NTI MVO+/IMH− and NTI MVO−/IMH− segments). Within the TI segment group, reperfusion injury was common (277 of 425 segments [65%]). Notably, there were no differences in all types of strain between TI segments without reperfusion injury and those with MVO as the only manifestation of reperfusion injury ( P > .20 for all comparisons between TI MVO−/IMH− and TI MVO+/IMH− segments). However, the occurrence of IMH decreased CS significantly ( P = .004 for TI MVO+/IMH+ vs TI MVO+/IMH− segments) without affecting LS and RS ( P > .50). The receiver operating characteristic curve analysis showed that the sensitivity and specificity of CS using a cutoff of >−11.66% to diagnose IMH were 85.16% and 75.11%, respectively (area under the curve = 0.86, P = .0001). After cross-validation of our data set by k-folding, the sensitivity and specificity were 78.00% and 79.45%, respectively. The results for intra- and interobserver reproducibility of strain measurements are listed in Table 2 .
|Absolute mean difference (mean ± SD)||Intraclass correlation||95% CI||P|
|Intraobserver||0.26 ± 2.53||0.98||0.98||0.99||<.001|
|Interobserver||0.28 ± 4.23||0.94||0.92||0.95||<.001|
|Intraobserver||1.26 ± 10.06||0.93||0.92||0.95||<.001|
|Interobserver||0.30 ± 8.2||0.93||0.89||0.95||<.001|
|Intraobserver||0.19 ± 1.89||0.99||0.98||0.99||<.001|
|Interobserver||−0.07 ± 2.34||0.97||0.96||0.98||<.001|