Baseline LV ejection fraction by cardiac magnetic resonance and 2D echocardiography after ST-elevation myocardial infarction – influence of infarct location and prognostic impact

The patients in this study were participants in a prospective analysis enrolling patients who underwent CMR after acute STEMI treated with primary percutaneous coronary intervention (PCI) [4, 17, 18]. Inclusion criteria were the diagnosis of first STEMI according to the redefined ESC/ACC committee criteria [19] and primary PCI within 24 h of symptom onset. Exclusion criteria were renal dysfunction with an estimated glomerular filtration rate < 30 ml/min/1.73 m², Killip class > 2 at time of CMR acquisition and contraindications for CMR.

Demographic data and clinical profile of patients were acquired with the help of a standardised questionnaire during hospitalisation for STEMI. Blood samples were collected as previously reported [20]. The study complies with the Declaration of Helsinki and the local ethics committee approved the study protocol. Written informed consent was obtained from all participants.

Echocardiography was performed in routine clinical practice in the Echocardiography Department of the University Hospital Innsbruck according to current guidelines [21]. Native exams were reloaded from the PACS and analysed twice by a single consultant cardiologist (JPS) at a time interval of 2 months and measurements averaged. Measurements were used to calculate intraobserver variabilities. JPS has significant experience and is certified in transthoracic and transoesophageal echocardiography by the European Society of Cardiology. JPS was blinded to CMR and clinical results. Echocardiograms were analysed on the IMAGE-COM® Software of TOMTEC Imaging Systems.

Determination of end-diastolic and end-systolic volumes was performed on four-chamber view (4CV) and two-chamber view (2CV) according to the Simpson method. Similar to others, we have used a 5-point scale to assess image quality [22]. Each 4CV and 2CV was assessed separately to increase accuracy of quality assessment. Therefore, a maximum of 10 points was allowed if there was complete endocardial definition and typical configuration (e.g. no foreshortening in 4CV or posterolateral papillary muscle visible in 2CV) in both views. Four points were allowed for each view when picture quality was mildly reduced, 3 when moderate and 2 when moderately reduced and 1 point if image quality was of borderline quality. Interobserver variabilities were calculated using 50 subjects which were selected evenly across the spectrum of 2D echocardiography quality by a second observer accredited in transthoracic and transoesophageal echocardiography by the European Society of Cardiology (GK). Observer variability analyses were performed without knowledge of patient identity or previous results.

All scans were performed on a 1.5-Tesla Magnetom AVANTO scanner (Siemens). Briefly, cine CMR images in short-axis were acquired using breath-hold, retrospective ECG-triggered TrueFISP bright-blood sequences. Evaluation of images was performed using a standard software (ARGUS, Siemens). End-diastolic and end-systolic volumes as well as stroke volume and LVEF were obtained from manual delineation of the endocardial borders. The most basal slice in end-systolic views was discarded if myocardium was not present in less than a half of the ventricular circumference. Trabeculation and papillary muscles were included into the LV volume. Late gadolinium enhancement (LGE) images were acquired by using an ECG-triggered phase-sensitive inversion recovery single-shot TrueFISP sequence with consecutive short-axis slices as described in detail previously [23‐26]. Infarct localisation was assessed LGE CMR. Cardiac magnetic resonance images were acquired under the supervision of a EuroCMR level II–certified radiologist (AM) with a long experience in CMR (> 10 years). Analysis of CMR cine volumetric data was performed by individuals trained in volumetric dataset analysis and reviewed by at least level I–certified individuals with a long experience in CMR (> 10 years GK, AM; > 5 years SJR).

A LVEF of < 52% was chosen as the cut-off for reduced LVEF in both image modalities because it represented both the mean and median in the 2DE population and it corresponds to the lower limit of normal in current echocardiography guidelines for male patients [21], which fits the male predominance in our study. To ensure intermodality comparability, the same cut-off was chosen for CMR.

Patients were followed for major adverse cardiac events (MACE), defined as a composite of death, myocardial re-infarction and new congestive heart failure as previously described [27]. Time to MACE was defined as time from PCI to the first end-point.

Statistical analysis was performed using SPSS for Windows, release 22.0.0.1 (IBM). The Shapiro–Wilk test was used to test for normal distribution. Normally distributed continuous data are presented as mean ± standard deviation (SD) and comparisons were performed by using paired t test. Non-normally distributed continuous data are presented as median with interquartile range (IQR) and comparisons were performed with non-parametric Wilkoxon signed-rank test. Comparisons of normally distributed continuous data in patients with or without MACE were performed using unpaired t test or Mann-Whitney U Test in the case of non-normally distributed data. Correlation of variables was performed using Pearson correlation coefficient. P values < 0.05 were considered to indicate statistical significance. Intermodality agreement was studied using the Bland–Altman method, whereby the mean difference was presented as the bias and 95% limits of agreement around the bias expressed as the mean difference ± 1.96 SDs. Interobserver and intraobserver variations of LVEF by 2DE were computed as the root of the mean squared differences between corresponding observations, divided by the number of observations. Intra- and interobserver correlations were calculated using intraclass correlation coefficient.

Kaplan–Meier survival curves were used to describe the cumulative incidence of event-free survival over time, and log-rank test was used to test for differences. Multivariate Cox regression analysis, adjusted for sex and age, was applied to evaluate the effect of LVEF on MACE.

During the study period, 323 STEMI patients were enrolled into the CMR database. In 271 of these, transthoracic echocardiograms were performed in addition to CMR and pictures archived in the hospital’s PACS. In 221 of these, appropriate 4CV and 2CV were available and analysed. One patient was excluded due to severe breathing artefacts in CMR. Echocardiograms were performed a median of 3 (IQR 2–5), and CMR a median of 2.6 days after STEMI (IQR 2–4). The mean difference between exams was 0.4± 1.8 days. Mean heart rate during MR was 74 ± 16; three patients had heart rates > 110 beats/min.

The mean age of the study participants was 58 ± 11 years; 34 patients (15.4%) were female. Detailed patient characteristics on admission are shown in Table 1.

Echocardiography image quality was evenly distributed, with the majority of echocardiograms being of moderate quality (5–7 points) (Fig. 1). Echocardiograms of lower quality (≤ 4 points; n = 39; 17.7%) were excluded, leaving 181 patients for subsequent analysis.

LVEF measurements were moderately correlated (r = 0.589; p < 0.0001) and not statistically different from each other (CMR 53 ± 11% vs. 2DE 52 ± 8%; p = 0.16). Only a minority of patients (12) demonstrated a severely reduced LVEF of < 35% during CMR. Figure 2 (panel a) shows the Bland–Altman plot demonstrating a small mean difference though relatively wide limits of agreement (− 17.6 to + 19%).

In 161 patients (89%), follow-up was completed, and 24 patients experienced a MACE after a median follow-up of 41 months (IQR 28.1–56). In total, 40.3% of patients were classified as reduced LVEF with CMR compared with 52% when assessed with 2DE.

In univariate analysis, patients with MACE demonstrated significantly lower LVEF in both imaging modalities when compared with patients without MACE (Table 2). Figure 3 demonstrates Kaplan–Meier curves for each imaging modality showing significantly lower event-free survival in patients with reduced LVEF of < 52% (2DE p = 0.03; CMR p = 0.025; log-rank).

In multivariate Cox regression analysis, adjusted for sex and age, reduced LVEF (< 52%) as assessed by either 2DE or CMR was similarly predictive of MACE (2DE HR = 2.57 (95% CI 1.1–6.2), p = 0.036; CMR HR 2.51 (95% CI 1.1–5.7), p = 0.028).

Several limitations of our study need to be mentioned. Firstly, a significant number of echocardiograms needed to be excluded upfront in our analysis due to the lack of appropriately stored views and low imaging quality; however, strict quality control is important in an operator-dependent technique like 2D echocardiography. Early measurements of LVEF after STEMI can be misleading because of increase in contractility in uninvolved territories [38‐40] and improvement in LVEF may occur in patients who are reperfused [41]. Other factors can influence measurements in the phase of acute cardiac disease, like different loading conditions, stress, arrhythmia or tachycardia. In this study, CMR and 2DE were performed very close to each other (mean 0.4 days) but not truly sequentially; therefore, haemodynamic differences between exams cannot be fully ruled out. However, mean LV function was relatively preserved in our population. Secondly, a majority of patients were receiving beta blockers (86%) and only a minority were receiving diuretic therapy (16%). These factors as well as the normal mean heart rate during CMR (74) all indicate haemodynamic stability.

Furthermore, due to the small number of patients with severely reduced LVEF, our data cannot be extrapolated to this particularly important subgroup. Pellika et al have recently compared echocardiography with CMR in the STICH Trial, a large trial including patients with previous myocardial infarction with significant left ventricular dysfunction and have found only moderate correlation of LVEF by CMR and 2DE, although in 46% of patients only single-plane echocardiography was performed [42]. Similarly, this applies to other acute cardiomyopathies with regional wall motion abnormalities or healthy probands.

We are unfortunately unable to offer data about myocardial strain in our patients, as assessed by myocardial tagging or CMR feature tracking [43]. Myocardial strain imaging is a promising tool for better quantification of global and regional left ventricular functions in a broad range of cardiovascular diseases, including myocardial infarction or myocarditis [44‐46]. Several CMR methods have been described and there is growing evidence that these techniques offer robust information after STEMI [47]. However, the improvement of risk stratification beyond traditional CMR indexes such as left ventricular ejection fraction remains controversial. Recent clinical outcome studies in STEMI patients using feature tracking or displacement encoding with stimulated echoes (DENSE) CMR indicated a strong prognostic role of global longitudinal strain or circumferential strain [48‐50].

Despite these factors, our data suggest that LVEF measurements by 2DE at baseline after STEMI are generally robust provided adequate image and reporting quality and the infarct location are taken into consideration.

In summary, 2D echocardiography significantly underestimated LVEF in comparison with CMR at baseline after non-anterior STEMI, whereas after anterior infarction, measurements were within acceptable limits of agreement. Despite these differences, a reduced LVEF of < 52% by both imaging modalities had similar prognostic values for MACE over a mean follow-up of 41 months. Further study is warranted to investigate if this phenomenon is confined to the acute post-infarct period as it would have significant impact on further device therapy.