Patients

For this prospective study, we screened 437 unselected patients with known chronic ICM (defined as known coronary artery disease and LVEF ≤ 50%) who were hospitalized for heart failure and underwent echocardiography and CMR within 3 ± 1 days between 2007 and 2012 ( Figure 1 ). All patients received optimal medical treatment for ≥3 months and coronary revascularization (all vessels were classified as open by recorded Thrombolysis In Myocardial Infarction grade 3 flow) according to current guidelines before inclusion in the study. Patients with recent myocardial infarctions (<90 days), atrial fibrillation, and relevant valvular heart disease were excluded. All patients gave informed written consent. Detailed follow-up was performed using recordings in a central database (including device interrogation) and by telephone contact using a scripted interview annually.

Flowchart of patient study inclusion. CABG , Coronary artery bypass graft; PCI , percutaneous coronary intervention; Pts , patients.

Echocardiography

Echocardiography was performed with a Vivid 7 and E9 system (GE Vingmed Ultrasound AS, Horten, Norway) equipped with a 2.5-MHz transducer. The frame rate for these studies was between 56 and 92 frames/sec, using tissue harmonic imaging. Three LV parasternal short-axis views at the basal, midventricular, and apical levels and three views from an apical window (two-chamber, four-chamber, and long-axis views) were acquired. An 18-segment model was used to divide the left ventricle. Wall motion was assessed by visual interpretation for each LV segment as normokinetic, mildly hypokinetic, severely hypokinetic, akinetic, or dyskinetic. WMSI was calculated for each patient as the average of the analyzed segmental values. LVEF was assessed using the biplane Simpson method using manual tracing of digital images.

Myocardial Deformation Imaging

Analysis was performed offline with the aid of a commercially available software package (EchoPAC 113 1.0; GE Vingmed Ultrasound AS). It allows calculation of mean strain values for total wall thickness and additionally for each of three myocardial layers (endocardial, midmyocardial, and epicardial), as described previously, within a few minutes ( Figure 2 ). Circumferential strain (CS), longitudinal strain (LS), and radial strain myocardial deformation parameters were determined for each myocardial segment and averaged considering all 18 segments to obtain one summarized data point. Following actual echocardiographic definitions, the label “global” was used only for strain data regarding the total myocardial wall thickness. All echocardiographic data were analyzed by two blinded experienced cardiologist.

Layer-specific CS with corresponding CMR image. All given strain-time curves exhibit the circumferential deformation separated into the three myocardial layers: endocardial (endo), midmyocardial (mid), and epicardial (epi). The different color-coded lines in each curve exhibit the strain-time run of one myocardial segment. Regarding the endocardial scar (CMR image), the deformation of the inferior ( dark blue line ) and inferolateral ( purple line ) segments is reduced, mainly in the endocardial-layer analysis. The lines were retraced for better visualization.

Cardiac Magnetic Resonance Imaging

Cardiac magnetic resonance imaging (CMR) was performed on a 1.5-T whole-body magnetic resonance scanner (Achieva; Philips Medical Systems, Best, the Netherlands), as described previously, within 3 ± 1 days after echocardiography. Each myocardial segment was evaluated for the presence of hyperenhancement, defined as an area of signal enhancement ≥2 SDs of the signal intensity of nonenhanced myocardium. The total myocardial area and the contrast-enhanced area of each segment were traced manually. TMS was calculated, defined as the percentage contrast-enhanced area of the total myocardial area (area _{hyperenhancement} /area _myocardium × 100). All CMR data were analyzed by two blinded, experienced cardiologists.

Clinical Follow-Up

Clinical follow-up was performed annually by telephone contact and a centrally recorded database (including reports of implantable cardioverter-defibrillator device interrogation). All verification of database recordings and interviews were performed by an experienced nurse or physician. During the interview, the patient or a family member was queried for the occurrence of cardiac events, defined as readmission for worsening of heart failure, ventricular arrhythmias, or death of any cause. If such an event was identified, the referring general practitioner was contacted for detailed information. Ventricular arrhythmias were ranked as clinical events only if documented by device interrogation or definite electrocardiography.

Statistical Analysis

Clinical characteristics of the study population are presented as frequencies or as mean ± SD. Continuous variables were compared using Student’s t test, the Wilcoxon test, or analysis of variance as appropriate. We used the χ ² test or Fisher exact test to test the null hypothesis that two dichotomous variables were independent.

To define intraobserver and interobserver variability, the same observer and a second independent observer repeated the analysis in 40 randomly selected patients using the same two-dimensional echocardiographic or CMR loop. Intraclass coefficients were calculated.

Univariate analysis was performed to establish the relationship between baseline clinical data and cardiac events. Cox proportional hazards analysis was used to determine significant predictors for cardiac events. Variables with univariate significance of <0.1 were selected for inclusion in the multivariate model. A series of nested models with the separate addition of all myocardial deformation parameters, LVEF, and TMS were undertaken. The ability of myocardial deformation parameters as well as TMS to predict cardiac events was explored with a random-effects model to address the issue of repeated observations (data on multiple segments per patient used in the analysis). The output of a generalized estimating equation approach with a binomial distribution, a logit link, and a working correlation matrix with exchangeable correlation assumption allowed the derivation of receiver operating characteristic curves, which were used to designate cutoffs. Using these cutoffs, we compared the sensitivity and specificity of all parameters for the prediction of cardiac events according to the McNemar test.

To express uncertainty concerning the true parameters, 95% CIs were calculated using 1,000 bootstrap samples from the original data, with sample size equal to that of the study presented here. These bootstrap CIs allowed a formal statistical comparison of areas under the curve.

Survival without cardiac events was expressed using Kaplan-Meier analysis. All tests were two sided, and P values < .05 were considered to indicate statistical significance. All statistical analyses were performed using SPSS version 19.0 (SPSS, Chicago, IL).

Patients

Forty-seven patients were excluded: 18 because of valvular heart disease, 13 because of recent myocardial infarctions, five patients because of poor echocardiographic image quality, and five patients because there was a delay between echocardiography and CMR (2 and 3 months). Finally, six patients showed such difficult and distinctive coronary calcification that the heart team considered the risk for complete revascularization too high and constituted a medical approach because of age and comorbidities. Three hundred ninety patients (mean age, 63 ± 4 years; 69% men) formed the study population; baseline characteristics are summarized in Table 1 . The mean LVEF was 41 ± 7%. Three hundred forty-three patients (88%) underwent percutaneous coronary intervention, and 14 (4%) underwent coronary artery bypass graft surgery to achieve optimal myocardial revascularization before inclusion in the study. In the remaining 33 patients, there was no actual need for revascularization of the coronary arteries (coronary stenoses < 50%). Patients with cardiac events were older and more often had diabetes mellitus and renal insufficiency (defined as an estimated glomerular filtration rate < 60 mL/min/1.73 m ² ). Besides this, no significant differences in comorbidities and medication use at admission were observed between patients with and those without adverse events. Echocardiography and CMR were performed within 3 ± 1 days.

Echocardiographic and CMR Findings

Comparing the echocardiographic data, patients with cardiac events had significantly lower LVEFs (38 ± 7% vs 43 ± 6%, P = .002), global CS (GCS) (−14 ± 5% vs −22 ± 6%, P = .007), endocardial CS (−20 ± 7% vs −26 ± 6%, P = .001), GLS (−12 ± 4% vs −17 ± 4%, P = .002), endocardial LS (−15 ± 4% vs −18 ± 5%, P = .022), and global radial strain (24 ± 6% vs 30 ± 10%, P = .005) and significantly higher WMSIs (1.7 ± 0.2% vs 1.5 ± 0.2%, P = .004) than those without cardiac events ( Table 2 ). Regarding the CMR data, patients with cardiac events had significantly higher TMS (31 ± 11% vs 23 ± 7%, P = .001) compared with those without cardiac events ( Table 2 ).

Prediction of Cardiac Events

There were 133 cardiac events (34%) over 4.9 ± 2.2 years of follow-up: 95 cases of heart failure, 11 deaths, and 27 cases of documented ventricular arrhythmias (89% by device interrogation and 11% by definite electrocardiography). The significant multivariate predictors of cardiac events from baseline variables were age (hazard ratio, 1.27; 95% CI, 1.06-1.73), diabetes (hazard ratio, 1.52; 95% CI, 1.11-2.41), and renal insufficiency (hazard ratio, 1.77; 95% CI, 1.02-2.56) ( Table 3 ). The overall model χ ² value was 23.1. As a second step to the baseline model, several separate models were determined to evaluate the additional prognostic power by adding the imaging parameters. Diabetes and renal insufficiency remained strong predictors; age was not significant in any of these models. Regarding the imaging parameters, GCS, endocardial CS, GLS, endocardial LS, LVEF, and TMS were strong predictors, and their addition resulted in a significant increase in model χ ² value over baseline. Endocardial CS showed the greatest increase (χ ² = 39.2, P < .001), followed by endocardial LS (χ ² = 36.3, P < .001), GLS (χ ² = 34.8, P < .001), TMS (χ ² = 31.6, P = .004), GCS (χ ² = 27.2, P = .013), and LVEF (χ ² = 25.8, P = .032) ( Figure 3 ).

Table 3

Analyses to identify baseline predictors of cardiac events

Variable	Univariate			Multivariate
	P	HR	95% CI	P	HR	95% CI
Age	.02	1.33	1.11-1.53	.04	1.27	1.06-1.73
Male sex	.25	0.86	0.32-1.03
Heart rate	.18	0.95	0.91-1.07
Hypertension	.83	1.02	0.78-1.32
Dyslipidemia	.76	0.89	0.80-1.25
Diabetes	.001	1.57	1.17-2.23	.001	1.52	1.11-2.41
Renal insufficiency	.001	1.81	1.21-2.97	.001	1.77	1.02-2.56
Smoker	.66	1.06	0.74-1.38
Family history	.93	0.97	0.66-1.64
CRT/ICD device	.001	1.63	1.54-1.81	.001	1.59	1.51-1.78

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Quantification of Right Ventricular Volume and Function Using Single-Beat Three-Dimensional Echocardiography: A Validation Study with Cardiac Magnetic Resonance
2. Right Ventricular Deformation Analyses Using a Three-Dimensional Speckle-Tracking Echocardiographic System Specialized for the Right Ventricle
3. The Impact of Function-Flow Interaction on Left Ventricular Efficiency in Patients with Conduction Abnormalities: A Particle Image Velocimetry and Tissue Doppler Study
4. ASE Scientific Sessions 2016: Sleepless In Seattle

Stay updated, free articles. Join our Telegram channel

Join