Discussion

Our data clearly show that in patients presenting to the ED with chest pain and suspected ACS, GLS does not predict long term outcome, namely cardiac death, ACS, revascularization, hospitalization for heart failure or atrial fibrillation. The results were the same whether we defined worse GLS as GLS > -18.7% (median GLS) or GLS > -17% (abnormal GLS according to guidelines), or when we analyzed only patients with optimal 2D echo image quality. To the best of our knowledge, this is the first study to report the effect of GLS on long-term outcome in patients with suspected ACS. We have previously reported a significantly higher 6-month MACE in patients with ACS as compared to patients in which ACS was excluded (5.8% vs. 0.6%,P =0.0002).15 The fact that 6-month MACE was extremely rare in 2DSPER in patients in whom ACS was excluded, suggests that the likelihood of missing significant CAD in these patients was very low.
GLS has been shown to be a good predictor of outcome, better than LVEF, in patients with heart failure, myocardial infarction and severe valvular disease.7–11,25 In a meta-analysis of 16 published articles which included 5721 patients, Kalam et al concluded that GLS predicts all-cause mortality and MACE (cardiac death, hospitalization for heart failure and malignant arrhythmia).10 Most of the patients included in the meta-analysis had myocardial infarction, heart failure or severe valvular or myocardial disease. In contrast to our study population, these patients were sicker, had a wider range of GLS values and included more patients with abnormal GLS values, thus explaining the discrepancy between the other studies and ours. Even in patients with heart failure and preserved LVEF, mean GLS was worse and standard deviation wider, as compared to our group of patients with worse GLS (-15.2 ± 4.6% in the study of Park et al, compared to -16.7±1.5% in our study).9 Most of our patients had good LV function. In 2DSPER, patients without ACS had worse GLS as compared to control groups in previous studies reporting high diagnostic accuracy of 2DLS.17,19 Worse GLS, however, was not associated with poor outcome in our study.
In addition, most patients with ACS included in 2DSPER had unstable angina pectoris and one vessel disease, and good outcome is expected in such patients. In sicker patients with myocardial enzyme leak and severe CAD, 2DLS is unnecessary for the diagnosis of ACS.26
Although 2DLS and GLS are accurate, reproducible and automated measures of LV contraction, they are dependent on 2D echo image quality and loading conditions which limit their accuracy, similar to LVEF.4,27,28 Another measure of LV strain, global circumferential strain, has been shown to be a better predictor of outcome when compared to GLS.25 In 2DSPER we used a new 2DLS parameter, the peak systolic strain value identifying the worst 20% LV segments (PSS20%), since this parameter was reported to outperform GLS in diagnosing ACS, because of its ability to identify minor wall motion abnormalities.17 Unfortunately, in the 2DSPER study PSS20% was not superior to GLS in diagnosing ACS.15 Similar to most other studies assessing the utility of LV strain in predicting outcome, we used GLS, since GLS is a simple and robust diagnostic tool, readily available in most commercial echo machines.
In our study, independent predictors of long-term MACE were history of CAD, ACS diagnosis, male gender and hypertension, but not GLS. Only 14% of our study population had ACS, and recurrent ACS or revascularization accounted for the majority of long-term events included in MACE, not heart failure or cardiac death. In a study from Poland of 2731 patients with unstable angina pectoris, the predictors of death or coronary events during 3 years follow-up were age, kidney disease, hypertension, diabetes, previous stroke and previous percutaneous coronary interventions.29 In another study of 230 patients discharged from the ED with a diagnosis of chest pain of undetermined origin, 4.4% had MACE during 12 months of follow-up, and the predictors of MACE were abnormal ECG, diabetes, or preexisting CAD.30 As expected, in our study similar to the other studies, markers of CAD at baseline predicted late coronary events. GLS, contrary to previous reports, not only failed to accurately identify ACS in the ED, but failed to predict future cardiac events as well, thus extending our findings from the 2DSPER multicenter study, of a limited value for GLS in low-risk patients presenting to the ED with suspected ACS.