17

Abstract

Introduction: Troponin strongly predicts mortality in both acute coronary syndrome (ACS) and non-ACS patients. However, the mechanisms causing troponin elevation in non-ACS and sepsis in particular are yet unclear. Hypothesis: We hypothesized that myocardial dysfunction as identified by advanced echocardiography can explain troponin elevation and mortality in severe sepsis and septic shock. Methods: A cohort of 106 patients with severe sepsis or septic shock underwent 225 echocardiograms and high-sensitivity troponin-T (hs-cTnT) measurements on the same day (2.1 ± 1.4 /patient). All clinical and laboratory data were collected and patients were followed for mortality for up to 1 year. All echocardiography parameters including mitral annular tissue Doppler imaging, left ventricular (LV) longitudinal and circumferential strain and strain-rate by speckle-tracking, and 3-dimentional volume measurements of the LV and right ventricle (RV), were correlated with concurrent hs-cTnT concentrations and with survival. Univariate and multivariate mixed linear model and logistic regression analyses were used to investigate the correlations of all clinical and echocardiographic parameters with log-transformed hs-cTnT (log(hs-cTnT) and with in-hospital mortality. Results: Ischemic heart disease (IHD), glomerular filtration rate (eGFR) and APACHE-II score were the clinical independent predictors of log(hs-cTnT). RV dilatation (increased RV end-systolic volume index -RVESVi) and diastolic dysfunction (increased mitral E-wave to strain-rate e’-wave - E/SRe’ ratio) were the independent echocardiographic correlates of log(hs-cTnT). Combining clinical and echocardiographic parameters, APACHE-II score, eGFR, RVESVi and E/SRe’-wave independently correlated with log(hs-cTnT) (mixed linear model: t= -2.1, 2.8, 3.3 and 3.8, p=0.007, 0.006, 0.001 and 0.0002, respectively). Fifty-one (48%) patients died within one year and 41 (39%) died in-hospital. RVESVi, SRe’-wave and APACHE-II were the independent predictors of in-hospital mortality (odds ratio= 1.06, 0.02 and 1.4, p=0.004, 0.010 and 0.001, respectively). Hs-cTnT, a significant univariate predictor of mortality (odds rati0=4.8, p=0.004), lost its predictive value when included in the multivariate analysis with RVESVi and SRe’-wave (odds ratio= 2.6, 1.05 and 0.43, p= 0.13, 0.032 and 0.012, respectively). Conclusions: LV Diastolic dysfunction and RV dilatation linearly correlate with hs-cTnT concentrations and predict mortality in severe sepsis and septic shock.