Page 134 - The role of advanced echocardiography in patients with ischemic heart disease - Rachid Abou

Chapter seven. LVMD in STEMI and prognosis
significance for variables to be included in the multivariable linear regression analysis was set at p<0.05. Furthermore, cumulative event rates were estimated with Kaplan- Meier analysis. The study population was divided into two groups according to the median LVMD and the cumulative event rates were compared between groups with log-rank tests. The association of clinical and echocardiographic variables with all-cause mortality was tested using the Cox proportional hazards analysis. The hazard ratio (HR) and 95% confidence interval (CI) were calculated. Statistically significant associates of all-cause mortality in the univariable Cox regression analysis (p<0.05) were included in multivariable Cox regression models. To illustrate the change in relative risk across the spectrum of baseline LVMD, a spline curve was built. The spline model for LVMD versus the risk of all-cause mortality was plotted with overlaid confidence intervals. Finally, after adjusting for multiple parameters (age, hypertension, diabetes mellitus, previous myocardial infarction, QRS duration, multi-vessel disease, thrombolysis in myocardial infarction (TIMI) flow ≥2, Killip class ≥ 2, peak troponin levels, estimated glomerular filtration rate (eGFR), body mass index, angiotensin converting enzyme inhibitor (ACEi) or angiotensin receptor blocker (ARBs), β-blockers, heart rate at discharge, WMSI, mitral regurgitation ≥ 2, LVEF, LV mass (indexed) and LV GLS, a multivariate spline model for all-cause mortality versus LVMD was built. To evaluate the incremental value of LVMD over clinical and conventional echocardiographic variables, a baseline regression model was created and the change in global χ2 values was calculated. The intra- and inter- observer reproducibility of LVMD was assessed using the Bland-Altman method and the intraclass correlation coefficient (ICC). In addition, the beat-to-beat variability was assessed. Twenty-five randomly selected subjects were analyzed by 2 separate readers. An excellent agreement was defined as ICC >0.75, whereas strong agreement was defined as ICC 0.60 to 0.74. Statistical analysis was performed on SPSS for Windows v23.0 (IBM, Armonk, New York). A 2-tailed p-value <0.05 was considered statistically significant.
RESULTS
A total of 1,124 STEMI patients treated with primary PCI were evaluated. In 124 (11%) patients, STE analysis was not feasible and therefore were excluded. The final study population consisted of 1,000 patients (77% male, mean age 61±12 years) (Table 1). For the overall study population, mean LVEF was 47±10%, mean LV GLS was -13.7 ± 3.7% and median LVMD 54 (IQR 43-66) ms (Table 2).
128