ST-segment elevation myocardial infarction patients who visited our institution, a tertiary referral center from January 2008 to September 2013, were eligible if they had: a) ST-segment elevation in 2 or more contiguous leads on ECG as standard definition; and b) had undergone successful primary PCI and CMR imaging. Patients with the following were excluded: admission for more than 12hours after symptom onset, previous myocardial infarction, or previous coronary revascularization (coronary artery bypass graft or PCI). Baseline clinical data from the dedicated registry, including past medical history, presence of risk factors, medications, angiographic and procedural data, and clinical outcomes, were recorded prospectively by research coordinators. This study was approved by the local institutional review board and waived the requirement for informed consent.

Percutaneous coronary intervention was performed using standard intervention techniques.11 The decision to pursue invasive treatment, as well as determining the access site, stent type for intravascular ultrasound, and use of glycoprotein IIb/IIIa receptor inhibitors, was left to the physician's discretion. Myocardial blush grade was evaluated using a final angiogram, as described previously.12 All interventions and procedural anticoagulations were performed in accordance with current standard guidelines.13

The primary outcome was myocardial infarct size (% of the left ventricle [LV] mass) measured by CMR. Secondary outcomes were AAR (% of LV mass), MSI (%), microvascular obstruction, and hemorrhagic infarction. Major adverse cardiovascular events were defined as a composite of all-cause death, myocardial infarction, and repeat coronary revascularization.

A STEMI diagnosis was made when ECG change fit the standard definition of new ST elevation at the J-point in at least 2 contiguous leads of at least ≥ 2mm (0.2mV) in men or ≥ 1.5mm (0.15mV) in women in leads V2-V3 and/or of ≥ 1mm (0.1mV) in other contiguous chest leads or the limb leads plus one of the following 2 criteria: duration of chest pain more than 30minutes or elevation of serum cardiac enzyme markers.14 A reciprocal change was defined as the presence of a 0.1-mV depression 80 msec after the J-point in 2 or more adjacent leads far from the leads exhibiting ST-segment elevation.15 Patients were stratified according to the presence or absence of reciprocal ECG changes.

Cardiac magnetic resonance was performed using a 1.5-T scanner (Achieva; Philips Medical Systems, Best, The Netherlands) with a SENSE cardiac coil according to our laboratory protocol.16 Images were acquired using electrocardiographic gating and expiratory breath holds. The CMR protocol consisted of cine, T2-weighted images, first-pass perfusion, and late-gadolinium enhancement imaging. Cine imaging was carried out based on balanced steady-state free precession sequences along the long and short axes from the apex to the base of the LV. Next, T2Ws were acquired in the cardiac short-axis direction using a dark-blood T2W inversion-recovery fast-spin echo sequence. First-pass perfusion imaging was obtained with the T1-weighted dynamic sequence (turbo field echo with SENSE, repetition time/echo time=2.6/1.3ms) after intravenous infusion of gadolinium-diethylenetriamine penta-acetic acid (Gd-DTPA, Magnevist; Bayer Schering Pharma, Berlin, Germany; 0.15 mmol/kg weight in total amount at 3mL/sec). Slice thickness was set to 6mm with a field of view of 40cm × 40cm and an image matrix of 128 × 128. Images of 4 locations for every 2 heartbeats were acquired for 40 phases.

The CMR images were analyzed using validated software (ARGUS; Siemens Medical System, Erlangen, Germany) at our MRI core laboratory by 2 experienced radiologists who were blinded to the clinical information. After acquisision of the short-axis images at the end of diastole and at the end of systole, endocardial borders were traced manually. Left ventricle, end-diastolic volume, end-systolic volume, and ejection fraction were calculated using the Simpson rule. The infarct volume was quantified as the sum of the area with late-gadolinium enhancement within each segment of the short-axis images multiplied by slice thickness to cover the entire LV.17 The infarct area was traced by the visual border detection via a manual drawing method using commercialized analysis software. Microvascular obstruction was defined and planimetered manually as a hypointense core within an infarcted myocardium on late-gadolinium enhancement images. The AAR mass was defined from T2W images as the myocardial mass with signal intensity ≥ 2 standard deviations above remote. The AAR mass was then normalized to the LV myocardial mass to estimate the AAR percentage. Intramyocardial hemorrhage was identified and planimetered manually as a hypointense core within the AAR on T2W images. Myocardial salvage was calculated as (AAR mass-infarct mass), and MSI was calculated as (myocardial salvage/AAR mass) and was expressed as a percentage.9,18Figure 1 shows a representative CMR image of reperfused nonanterior STEMI with or without reciprocal ST change.

Figure 1.

Representative cardiac magnetic resonance images from a study of nonanterior ST-segment elevation myocardial infarction with and without reciprocal change. A: a short-axis slice of a T2-weighted image with reciprocal change. B: a short-axis slice of a late-gadolinium enhancement image with reciprocal change. C: a short-axis slice of a T2-weighted image without reciprocal change. D: a short-axis slice of a late-gadolinium enhancement image without reciprocal change.

(0.3MB).

Continuous variables were compared using Student t test or the Wilcoxon rank-sum test where applicable and are presented as mean±standard deviation or median with interquartile range [IQR]. Categorical data were tested using the Fisher exact test or the chi-square test, as appropriate. Covariates that appeared to be relevant by virtue of having P-values < .2 on univariate analysis or were otherwise clinically relevant were included in our risk factor adjustment model, using the multiple linear regression model. To balance the underlying demographic and clinical characteristics, we also considered the propensity score method. Propensity scores were estimated using a multiple logistic regression model that included all variables presented in Table 1. The covariate balance was assessed by comparing the absolute standardized differences in covariates between the 2 groups (Table of the supplementary material). In the propensity score-matched patients, continuous variables as primary outcome and secondary outcomes of CMR were compared with a paired t test or the Wilcoxon signed-rank test. Categorical variables were compared with the McNemar or Bowker test of symmetry, as appropriate. We checked the residual plots thoroughly, including Cook's distance and the normal probability plots, and confirmed them to be satisfactory before adopting any final model results.

All tests were 2-tailed, and P < .05 was considered statistically significant. All analyses were performed using the Statistical Analysis Software package (SAS version 9.2, SAS Institute, Cary, North Carolina, United States).

Of the 301 registered patients, 39 who had chest pain before the 12-hour evaluation period and 18 who underwent previous PCI and coronary artery bypass grafting were excluded from our analysis. A total of 244 patients met our inclusion criteria for undergoing successful primary PCI and CMR imaging. These patients were divided into 2 groups: those with reciprocal change (n=133) and those without reciprocal change (n=111; Figure 2).

Figure 2.

The flowchart of the study population. CABG, coronary artery bypass graft; CMR, cardiac magnetic resonance imaging; ECG, electrocardiogram; STEMI, ST-segment elevation myocardial infarction; PCI, percutaneous coronary intervention.

(0.2MB).

The baseline clinical and angiographic characteristics by patient group are summarized in Table 1. Patients with reciprocal change had a higher frequency of diabetes mellitus (27.8% vs 16.2%; P=.032) and nonanterior infarction (72.2% vs 28.8%; P < .001) than those without. Pain onset to balloon time was shorter in patients with reciprocal change (221.5±169.8 vs 289.7±337.3 min; P=.042). Patients with reciprocal change had a higher incidence of baseline Thrombolysis In Myocardial Infarction flow of grade 0, thrombus aspiration, and no reflow during PCI compared with those without reciprocal change. Coronary angiography a found similar distribution of contralateral disease between the 2 groups (38.3% vs 35.1%; P=.69). The baseline clinical and angiographic characteristics in propensity score-matched patients are shown in Table 2.

Patients underwent CMR at a median of 3 days [IQR 3-5] after primary PCI, and there was no a significant difference between 2 groups (4.9±5.7 vs 5.0±5.5 days; P=.93). Infarct size, the primary outcome, was not significantly different between patients without and with reciprocal change (median [IQR]: 18.33% [12.34–27.65] vs 17.95% [11.44–25.49]; P=.35). Although patients with reciprocal changes had significantly higher AAR than those without (37.24% [27.83–46.82] vs 33.28% [20.07–44.27]; P=.05), MSI was not significantly different between the groups (48.57% [34.65–58.69] vs 44.63% [33.15–56.88]; P=.27). The other CMR parameters, including LV ejection fraction (P=.46), LV end-diastolic volume (P=.89), LV end-systolic volume (P=.83), LV stroke volume (P=.46), LV cardiac output (P=.56), LV mass (P=.66), the presence of microvascular obstruction (P=.15), and hemorrhagic infarction (P=.12), were not significantly different between the groups (Table 3). In a propensity score-matched subset data, CMR outcome parameters were not significantly different (Table 4).

After the performance of multiple linear regressions that included propensity scores and other adjustments for confounding variables, patients with reciprocal change had a larger extent of AAR (slope estimate=6.97, standard error [SE]=2.20, P=.002), and a greater MSI (slope estimate=5.47, SE=2.64, P=.04) compared with those without. Consequently, myocardial infarct size was not significantly different between the 2 groups (slope estimate=2.11, SE=1.43, P=.14).

The overall reciprocal change in ST depression was associated with infarct size (slope estimate=0.57, SE=0.25, P=.021) and AAR (slope estimate=1.02, SE=0.35, P=.004), according to our analysis.

The cumulative clinical outcomes of the study patients are summarized in Table 5. The median length of follow-up was 26.7 [IQR: 12.8–40.9] months and was not significantly different between patients without reciprocal change (28.2 [IQR: 11.1–42.8] months) and those with reciprocal change (25.9 [IQR: 13.3–38.9] months, P=.54).

The log-rank test indicated no significant differences in event times to clinical outcomes between the 2 groups (Table 5). The major adverse cardiovascular events occurred in 9% of patients with or without reciprocal change and were not significantly different (HR, 1.05; 95%CI, 0.45–2.44; P=.92; Table 5).

This study has several limitations. This was a nonrandomized observational study, and our results may have been significantly affected by confounding factors. Our study included data from a single center and had a limited number of patients, despite being the largest CMR study to date that assessed AAR values, infarct size, myocardial salvage, and MSI. Based on the inclusion and exclusion criteria of this single-center study and the treatment protocols used in specialized high-volume hospitals, our results may not be not generalizable to other STEMI populations, such as late STEMI presenters (> 12hours). Additionally, the predominantly male study populations limited the generalizability of the results. The patients were included over a relatively long time period during which therapeutic and guidelines changes could have occurred that could have influenced the results.

Additionally, the 2 groups were qualitatively quite different, and as a result, our participant loss was fairly significant when we attempted to match the groups. The number of pairs matched was extremely limited. Therefore, matched analysis was used to compare the results with those of the original unmatched cases. Had the number of unmatched cases been small, we would have attempted to identify the clinical reasons why they became unmatched.

Whether the real AAR is accurately delineated by T2-weighted CMR has been a matter of intense debate.35,36 Indeed, recent experimental data suggest that T2 may represent the actual infarct size rather than the real AAR,37 or may be influenced by the time between the acute event and image acquisition.38 It might be that hyperintense areas in T2-weighted sequences may not reflect the real AAR, and therefore provide biased estimates of myocardial salvage and inaccurate adjustments in statistical models. It has been recently suggested that edema might a dynamic phenomenon, and therefore unstable, during the first week after myocardial infarction. As opposed to early assessment alone, it might be crucial to perform CMR in a systematic narrow time window. For myocardial salvage measurement, we considered that early assessment of the AAR was crucial because myocardial edema was maximal and constant during the first week after infarction and decreased thereafter.39,40 Despite the lack of difference in the time interval from primary PCI to CMR between the groups (4.9±5.7 vs 5.0±5.5 days; P=.93), AAR size may vary significantly if measured 1 to 40 days after primary PCI, and the patients were not scanned at such a consistent narrow time window.

We could not completely rule out possible overestimation of the infarct size caused by the presence of edema on CMR within first 5 days. However, it would not have affected the results of our study because there was no significant difference in the time interval from primary PCI to CMR between the 2 groups.

The low event rate in our study population and the wide confidence intervals do not provide convincing evidence of poorer outcomes. In addition, the low incidence of events could be caused by a recruitment bias, with the inclusion of patients with nonsevere myocardial infarction in this CMR registry, as it showed a higher than expected incidence of Killip class I (87.2% and 91.9% in patients with and without reciprocal change). While adverse clinical events were not centrally adjudicated in our registries, all events were identified by the patients’ physicians and confirmed by the hospital's principal investigator.