Residual ischemia detection after ST-segment elevation myocardial infarction (STEMI) during dobutamine stress echocardiography (DSE) using visual analysis is challenging. The aim of the present study was to investigate the feasibility and accuracy of two-dimensional speckle-tracking strain DSE to detect significant coronary artery disease (CAD) after STEMI.
First STEMI patients ( n = 105; mean age, 60 ± 11 years; 86% men) treated with primary percutaneous coronary intervention undergoing full-protocol DSE at 3 months and repeat coronary angiography within 1 year were retrospectively included. Using two-dimensional speckle-tracking echocardiography, segmental and global left ventricular peak longitudinal systolic strain (PLSS) at rest and peak stress and change (Δ) in PLSS were measured. Significant CAD was defined as detection of >70% diameter stenosis at coronary angiography.
In total, 1,653 (93%) and 1,645 (92%) segments were analyzable at rest and peak stress, respectively. At follow-up, 38 patients (36%) showed significant angiographic CAD. These patients demonstrated greater worsening in global PLSS from rest to peak (−16.8 ± 0.5% to −12.6 ± 0.5%) compared with patients without significant CAD (−16.6 ± 0.4% to −14.3 ± 0.3%; group-stage interaction P < .001). The optimal cutoff of ΔPLSS for the detection of significant CAD on receiver operating characteristic curve analysis was ≥1.9% (area under the curve, 0.70; sensitivity, 87%; specificity, 46%; accuracy, 60%). Using a sentinel segment approach (apex, midposterior, and midinferior for the left anterior descending, left circumflex, and right coronary artery territories, respectively), larger segmental ΔPLSS was also independently associated with significant CAD (odds ratio, 1.1; 95% CI, 1.1–1.2).
Two-dimensional speckle-tracking echocardiographic strain analysis is feasible on DSE after STEMI and represents a promising new technique to detect significant angiographic CAD at follow-up.
Patients who recover from ST-segment elevation myocardial infarction (STEMI) remain at high risk for new ischemic events and premature death. For optimal risk stratification in post-STEMI patients, the recently updated European Society of Cardiology guidelines include stress testing or imaging, such as dobutamine stress echocardiography (DSE), in patients with multivessel disease or in whom revascularization of other vessels is being considered as a class IA recommendation. DSE is frequently used for the detection, localization, and assessment of the extent of ischemia in patients with suspected or established coronary artery disease (CAD) because of its wide availability, low cost, and lack of ionizing radiation. However, assessment of regional myocardial function on DSE relies on the semiquantitative evaluation of endocardial excursion and wall thickening and is therefore highly subjective and image quality dependent, even in the hands of expert observers. Moreover, detection of residual and/or new ischemia during DSE after STEMI is particularly challenging because of the presence of existing wall motion abnormalities.
The need for more quantitative techniques to objectively evaluate regional left ventricular (LV) myocardial performance during DSE has led to the incorporation of deformation indices such as two-dimensional (2D) speckle-tracking echocardiography (STE)–derived longitudinal strain, a highly sensitive alternative method of quantifying regional myocardial function on the basis of grayscale ultrasound imaging. Several experimental studies have already validated 2D strain techniques against sonomicrometry during dobutamine infusion and/or ischemia. Recently, clinical studies have investigated the diagnostic value of 2D strain and related parameters during DSE for inducible ischemia detection in patients with suspected CAD. However, despite the known prognostic value of residual ischemia detection in this population, few clinical studies have investigated 2D STE–derived strain parameters on DSE after STEMI.
We hypothesized that 2D STE may be incremental to conventional visual analysis in characterizing the complex ischemic substrate(s) in patients with previous STEMI. Therefore, the aim of this study was to compare the feasibility and accuracy of both wall motion score (WMS) and 2D speckle-tracking echocardiographic strain and its derived parameters on DSE after STEMI for the detection of angiographically significant CAD at follow-up.
All first STEMI patients presenting to our institution between November 2010 and February 2012 and treated according to the MISSION! protocol were evaluated for inclusion in this retrospective study. This protocol is designed to improve care around all aspects of STEMI and is based on the most recent American College of Cardiology Foundation and American Heart Association and European Society of Cardiology guidelines. Diagnosis of acute myocardial infarction was made on the basis of typical electrocardiographic changes and/or ischemic chest pain associated with elevation of cardiac enzymes. DSE is also performed as part of this protocol 3 months after STEMI in all patients for risk stratification and to optimize management through detection of new or residual ischemia and/or determination of myocardial viability. Repeat coronary angiography is performed during follow-up as part of this protocol if ischemia is demonstrated on DSE and/or if patients present with symptoms and signs suggestive of ischemia, as clinically appropriate.
For inclusion in the study, patients were required to have undergone both full-protocol DSE as recommended by our institutional guidelines and repeat coronary angiography (clinically and/or DSE driven) within 1 year. Patients were excluded from the study if they had undergone coronary artery bypass grafting or experienced repeat myocardial infarction up to the date of repeat coronary angiography. Additionally, patients were excluded if they reached an end point other than completion of the DSE protocol (see below) or if image quality at any stage of DSE was suboptimal for 2D speckle-tracking echocardiographic analysis ( Figure 1 ).
Patient data were prospectively collected in the departmental cardiology information system (EPD-Vision; Leiden University Medical Center, Leiden, The Netherlands) and retrospectively analyzed. For this retrospective analysis of clinically acquired data, the institutional review board waived the need for patient written informed consent.
DSE was performed according to a standard protocol. The decision of whether to instruct patients to continue or suspend β-blocker therapy 48 hours before the study was clinically directed. After a baseline electrocardiogram and echocardiogram were recorded, dobutamine was administered, beginning at a dose of 5 μg/kg/min. The dose was increased at 3-min intervals up to 40 μg/kg/min, and if the target heart rate (85% of age-predicted maximal heart rate) was not achieved, intravenous atropine in divided doses of 0.25 to 0.5 mg (up to 2 mg) was given. End points of our dobutamine stress echocardiographic protocol were completion of the test, significant arrhythmias, hypotension, severe hypertension (systolic blood pressure > 240 mm Hg) or intolerable symptoms. Stress echocardiographic images were obtained using a commercially available ultrasound system (iE33; Philips Medical Systems, Bothell, WA) equipped with a broadband S5-1 transducer with the patient in the left lateral decubitus position and were acquired from apical (four-, two-, and three-chamber) views at rest, low-dose, and peak-stress phases of dobutamine. Two-dimensional grayscale images were obtained at frame rates ranging from 60 to 100 frames/sec for all stages. Images were saved in cine-loop format from three consecutive beats, and analysis was subsequently performed offline using QLAB version 9.0 (Philips Medical Systems). Low-dose images were acquired at 20 μg/kg/min; peak-stress images were acquired upon achievement of target heart rate. Development of ischemia was defined by conventional visual analysis as a new or worsening wall motion abnormality or, in the case of segments with preexisting wall motion abnormalities, as a biphasic response if there was augmentation at low dose followed by further deterioration at peak stress.
Conventional Dobutamine Stress Echocardiographic Analysis
Wall motion was independently assessed by an experienced observer blinded to the coronary angiographic results according to the 17-segment model of the American Society of Echocardiography. Each segment was scored individually on the basis of its motion and systolic thickening (1 = normal, 2 = hypokinesia, 3 = akinesia, 4 = dyskinesia). Global WMS index (WMSI) was then calculated for each patient as the sum of the segment scores divided by the number of segments scored. Change in WMSI from rest to peak was also assessed (ΔWMSI).
Two-Dimensional Speckle-Tracking Echocardiographic Strain Analysis
Quantitative strain analysis using 2D STE was performed independently of conventional visual analysis by experienced observers blinded to the coronary angiographic results. In particular, longitudinal strain was measured for each LV segment using the apical four-, two-, and three-chamber views. The LV endocardial border was traced using the optimal frame for endocardial identification in all three apical views, and the automatically created region of interest was manually adjusted to the thickness of the myocardium. Segments were discarded if tracking was of persistent poor quality following readjustment of the region of interest. Subsequently, numeric and graphical displays of deformation parameters were automatically generated for all LV segments. Aortic valve closure was defined in the apical long-axis view, and the interval between the R wave and this time point was then automatically measured to serve as a reference for identification of end-systole.
LV peak longitudinal systolic strain (PLSS) was measured for each segment as maximal longitudinal shortening in systole. The presence of further segmental shortening occurring in diastole beyond maximal systolic shortening (postsystolic shortening [PSS]) and its magnitude (peak shortening in diastole minus PLSS) were assessed. PSS index (PSI) was then calculated as PSS divided by maximum shortening, expressed as a percentage for those segments with PSS present. The magnitude of change in each parameter between stages (peak to rest) was also calculated (ΔPLSS, ΔPSS, and ΔPSI). Notably, longitudinal shortening is denoted by convention using a negative sign; therefore, a positive value of ΔPLSS indicates more impaired global PLSS from rest to peak. For all three parameters, global values were also obtained by averaging the original segmental data.
All patients underwent selective coronary angiography using the Judkins technique, and images were read by an experienced interventionalist. Significant CAD was defined as >70% luminal diameter stenosis in one or more of the three major epicardial vessels, as assessed by computer-assisted quantitative coronary angiography using multiple planes. Segments were assigned to a specific coronary territory on the basis of a standardized perfusion model: left anterior descending coronary artery (LAD) (basal and middle anteroseptal, basal and middle anterior, apical anterior and septal, and the apex), left circumflex coronary artery (LCx) (basal and middle posterior and all lateral segments), and right coronary artery (RCA) (basal and middle septal and all inferior segments).
Continuous variables are presented as mean ± SD or mean ± SE or as median (interquartile range) as appropriate. Categorical data are presented as absolute numbers and percentages.
For global (per patient) analysis, Student’s paired t test was performed first to assess differences across stages for individual dobutamine stress echocardiographic parameters in the total population. Subsequently, a linear mixed-effects modeling approach was used to compare differences in conventional dobutamine stress echocardiographic (WMSI, heart rate, and blood pressure) and 2D speckle-tracking echocardiographic (PLSS, number of segments with PSS and PSI) parameters between patients with and those without significant CAD across both stages (group-stage interaction). Differences within groups at each stage were also evaluated. Subsequently, the relationship between the presence of significant CAD on follow-up angiography and a range of putative predictor variables (clinical, conventional dobutamine stress echocardiographic and 2D speckle-tracking echocardiographic parameters) was evaluated using univariate and multivariate logistic regression analysis. Candidate variables with P values < .05 on univariate analysis were included in the multivariate analysis. Because of multicollinearity, individual conventional dobutamine stress echocardiographic and quantitative 2D speckle-tracking echocardiographic parameters achieving this significance level were entered into separate multivariate models together with the other “fixed” clinical parameters achieving univariate significance at this same level. Thereafter, receiver operating characteristic (ROC) curve analysis was performed for the 2D speckle-tracking echocardiographic parameter achieving the strongest independent significance on multivariate analysis (ΔPLSS) to evaluate its predictive ability for the presence of significant CAD. Sensitivity, specificity, and diagnostic accuracy were calculated using cutoff values based on the principle of optimal sensitivity (with reasonable specificity) for the detection of significant CAD at follow-up. Finally, the likelihood ratio test was performed to evaluate the incremental value of quantitative strain analysis (using this cutoff) over conventional visual analysis (using the cutoff of two or more stress-induced wall motion abnormalities ) and clinical parameters for the prediction of significant follow-up CAD.
For segmental analysis, ROC curves were performed to investigate the accuracy of ΔPLSS across each segment for the detection of a significant stenosis (>70%) in the corresponding territory. For further segmental analysis involving WMS, all segments were included in the analyses. In contrast, to facilitate increased feasibility of 2D strain assessment on DSE for routine clinical use, representative segments (“sentinel segments”) were chosen for further analysis, selected as segments displaying the highest area under the curve (AUC) on ROC curve analysis for the detection of a significant stenosis (>70%) in the corresponding territory. Sensitivity, specificity, and accuracy were also calculated for each sentinel segment for the detection of significant CAD in that territory using the cutoff previously derived at the global level. Generalized estimating equations (allowing adjustment for the fact that segments in each patient are correlated) were then used to test the association between segmental parameters and the presence of a significant stenosis in the corresponding coronary territory (both unadjusted and adjusted for significant clinical parameters at that level). Finally, subgroup analyses were performed using these sentinel segments for the risk for >70% stenosis in the corresponding territory among relevant clinical (age, gender, diabetes) and infarct (infarct-related artery [IRA], single or multivessel disease, LAD territory or not, peak troponin T level according to median) subgroups.
All statistical tests were two sided, and P values < .05 were considered to indicate statistical significance. The statistical software program SPSS version 20.0 (SPSS, Inc, Chicago, IL) was used for all analyses.
The reproducibility of PLSS was assessed at both the global and segmental levels at both stages of DSE and expressed as both interclass correlation coefficient and as a percentage of the absolute difference divided by the mean of the pair-repeated observations (absolute difference). For global intra- and interobserver variability, 10 patients (340 segments in total) were selected at random, and measurements were repeated by the same observer on the same echocardiographic images ≥2 weeks apart and by another independent observer. Similarly, intra- and interobserver variability was calculated for each sentinel segment (60 segments in total).
Of 135 first STEMI patients meeting the initial inclusion criteria, 7% ( n = 10) were excluded because of the occurrence of a dobutamine stress echocardiographic end point other than completion of the protocol ( Figure 1 ). A further 15% ( n = 20) were excluded because of either inadequate image quality at rest or peak dose or inaccurate tracking involving a full regional wall or more than two segments within the same coronary territory. Clinical characteristics of the remaining 105 patients (mean age, 60 ± 11 years; 86% men) at the time of admission for STEMI are shown in Table 1 . The LAD was the IRA in 33% of patients, and just over 50% of the patient population had multivessel disease (defined as angiographic stenosis ≥ 50% in two or more vessels, including the IRA).
|Variable||Total patient population ( n = 105)|
|Age (y)||60 ± 11|
|Current or previous smoking||60 (57%)|
|Family history of CAD||41 (39%)|
|Killip class ≥ II||4 (4%)|
|Multivessel disease||59 (56%)|
|Final TIMI flow grade ≥ 2||103 (98%)|
|Peak CPK level (U/L)||1,478 (661–2, 611)|
|Peak TnT level (μg/L)||3.3 (1.2, 6.3)|
|Dual-antiplatelet therapy||103 (98%)|
|ACE inhibitors/ARBs||101 (96%)|
Median time from DSE to follow-up coronary angiography was 6 months (interquartile range, 5–6 months). In total, 36% of patients ( n = 38) had significant CAD (>70%) in one or more epicardial vessels at follow-up (LAD, 26%; LCx, 11%; RCA, 14%). There were no differences in dual-antiplatelet (97% vs 99%, P = .68), statin (100% vs 100%, P = 1.00), angiotensin-converting enzyme inhibitor or angiotensin receptor blocker (95% vs 97%, P = .56), or β-blocker (95% vs 94%, P = .88) therapy at the time of discharge after STEMI between those with significant CAD on follow-up angiography and those without significant CAD.
Conventional Dobutamine Stress Echocardiographic Analysis
The majority of patients (55% [ n = 76]) had suspended β-blocker therapy before DSE. Mean heart rates at rest and peak dose were 66 ± 13 and 138 ± 13 beats/min, respectively ( P < .001). In total, WMS assessment was feasible in 1,784 (100%), 1,768 (99%), and 1,785 (100%) segments at rest, low dose, and peak stress, respectively. The majority (52% [ n = 55]) of patients had abnormal WMSI values at rest, consisting of 220 abnormal segments (12%). Stress-induced wall motion abnormalities were seen in 11% of patients ( n = 51 segments). A biphasic response was observed in nine (0.5%) segments. Mean WMSI values at rest and peak dose were 1.16 ± 0.24 and 1.09 ± 0.20, respectively ( P < .001 for change).
Dividing the patient population according to patients with or without significant CAD at follow-up, the increases in heart rate ( P = .43) and systolic ( P = .75) and diastolic ( P = .59) blood pressure between groups from rest to peak and at each stage were similar ( Supplemental Figure 1 ). The change in WMSI during DSE was significantly different from rest to peak stress between patients with and those without significant CAD at follow-up (group-stage interaction P = .02) ( Figure 2 ). Within-group analysis showed a significant decrease in WMSI throughout DSE in patients without significant CAD (from 1.20 ± 0.03 to 1.10 ± 0.03, P < .001) and no significant change in WMSI (from 1.10 ± 0.04 to 1.09 ± 0.03, P = .63) from rest to peak stress in those with significant CAD at follow-up.
Quantitative 2D Speckle-Tracking Echocardiographic Analysis
Mean frame rates at rest and peak stress were 71 ± 12 and 71 ± 11 frames/sec, respectively. In total, 1,653 (93%) and 1,645 (92%) segments were analyzable at rest and at peak stress, respectively. The least feasible segment was the midanterior segment (82%) at rest and the midanteroseptal segment (76%) at peak stress ( Table 2 ). In the total population, mean global PLSS decreased significantly from −16.6 ± 3.2% at rest to −13.6 ± 2.9% at peak dose ( P < .001), while both mean global number of segments with PSS (from 6.1 ± 2.5 to 6.5 ± 3.2, P = .29) and global PSI (from 8.9 ± 6.7% to 9.9 ± 6.1%, P = .33) showed a nonsignificant trend toward an increase.
|Segment||Segmental ΔPLSS||Feasibility PLSS ∗|
|AUC (95% CI)||Rest||Peak|
|Basal anteroseptal||0.50 (0.35–0.64)||98||87|
|Apical septum||0.50 (0.37–0.63)||100||105|
|Apical anterior||0.51 (0.37–0.65)||93||98|
|Basal anterior||0.41 (0.29–0.54)||98||98|
|Basal posterior||0.42 (0.24–0.60)||97||96|
|Apical lateral||0.52 (0.32–0.73)||103||102|
|Basal lateral||0.48 (0.27–0.69)||100||102|
|Basal inferoseptum||0.46 (0.29–0.62)||102||103|
|Basal inferior||0.49 (0.33–0.64)||99||105|
|Apical inferior||0.47 (0.32–0.62)||103||103|
Differences in 2D speckle-tracking echocardiographic global parameters between patients with and those without significant CAD across stages and groups are also illustrated in Figure 2 . The only parameter to exhibit a significant group-stage interaction was global PLSS (significant CAD, −16.8 ± 0.5% to −12.6 ± 0.5%; no significant CAD, −16.6 ± 0.4% to −14.3 ± 0.3%, P < .001). Although global PSI only showed a trend toward overall group-stage significance ( P = .08), it did increase significantly from rest to peak in patients with significant CAD (from 8.4 ± 1.1% to 11.5 ± 1.0%, P = .02), leading to a significant difference at peak stress between the two groups (11.5 ± 1.0% vs 8.9 ± 0.7%, P = .04).
Conventional versus Quantitative Dobutamine Stress Echocardiographic Analysis: Global Assessment
Table 3 shows the results of univariate and multivariate analyses of global conventional and 2D speckle-tracking echocardiographic parameters on DSE associated with >70% stenosis in one or more epicardial vessels at follow-up after STEMI. Although ΔWMSI was associated with significant CAD at the univariate level (odds ratio [OR], 38; 95% CI, 1.2–1,178; P = .04), after adjusting for clinical parameters, no independent association was demonstrated (OR, 14; 95% CI, 0.32–643; P = .17). Both peak PLSS (OR, 1.2; 95% CI, 1.0–1.5; P = .03) and a larger positive value of global ΔPLSS (representing greater impairment of global PLSS from rest to peak) (OR, 1.3; 95% CI, 1.1–1.6; P = .01) were the only dobutamine stress echocardiographic (and 2D speckle-tracking echocardiographic strain) parameters independently associated with significant CAD on multivariate analysis ( Table 3 ). Given the greater strength of the association for ΔPLSS, this parameter was chosen as the primary 2D speckle-tracking echocardiographic strain parameter to perform further analyses. Of note, for this multivariate model, age (OR, 1.0; 95% CI, 1.0–1.1; P = .03) and multivessel disease (OR, 4.0; 95% CI, 1.4–11; P = .009) were also independent predictors of significant CAD at follow-up.
|Variable||Univariate analysis||Multivariate analysis ∗|
|OR (95% CI)||P||OR (95% CI)||P|
|Rest WMSI||0.12 (0.01–1.0)||.05|
|Peak WMSI||0.69 (0.09–5.2)||.72|
|ΔWMSI||38 (1.2–1,178)||.04||14 (0.32–643)||.17|
|Rest PLSS (%)||0.98 (0.87–1.1)||.76|
|Peak PLSS (%) †||1.3 (1.1–1.5)||.006||1.2 (1.0–1.5)||.03|
|ΔPLSS (%) ‡||1.4 (1.1–1.6)||.001||1.3 (1.1–1.6)||.01|
|Rest PSS segments||0.85 (0.71–1.0)||.06|
|Peak PSS segments||0.99 (0.88–1.1)||.89|
|ΔPSS segments||1.1 (0.96–1.2)||.24|
|Rest PSI (%)||0.98 (0.92–1.0)||.54|
|Peak PSI (%)||1.1 (1.0–1.1)||.04||1.1 (0.98–1.1)||.15|
|ΔPSI (%)||1.0 (1.0–1.1)||.08|
∗ Because of multicollinearity, separate multivariate analyses for each parameter achieving significance of P < .05 on univariate analysis were performed, adjusted for the clinical characteristics achieving univariate significance (age, LAD as the IRA, and multivessel disease).
‡ For this multivariate model, age (OR, 1.0; 95% CI, 1.0–1.1; P = .03) and multivessel disease (OR, 4.0; 95% CI, 1.4–11; P = .009) were also independent associates of the presence of significant CAD at follow-up.