Scholars@Duke publication: Two-dimensional global longitudinal strain is superior to left ventricular ejection fraction in prediction of outcome in patients with left-sided infective endocarditis.

Two-dimensional global longitudinal strain is superior to left ventricular ejection fraction in prediction of outcome in patients with left-sided infective endocarditis.

Publication , Journal Article

Lauridsen, TK; Alhede, C; Crowley, AL; Kisslo, J; Sørensen, LL; Hansen, TF; Risum, N; Larsen, CT; Hassager, C; Søgaard, P; Dahl, A; Bruun, NE

Published in: Int J Cardiol

June 1, 2018

Published version (DOI) Link to item

BACKGROUND: Impaired cardiac function is the main predictor of poor outcome in infective endocarditis (IE). Global longitudinal strain (GLS) derived from two-dimensional strain echocardiography has proven superior in prediction of long-term outcome as compared to left ventricular ejection fraction (LVEF) in valvular disease and heart failure in general. Whether measurements of cardiac deformation can predict survival in patients with IE has not previously been investigated. METHODS: The study included consecutive patients with Duke definite IE who underwent transthoracic and transesophageal echocardiography within 7 days. Clinical and echocardiographic markers associated with 1-year survival were identified using a Cox-proportional hazards model that included propensity adjustment for surgery. Reclassification statistics including receiver operating characteristic curves and net reclassification improvement were applied to LVEF and GLS, respectively. RESULTS: A cohort of 190 patients met eligibility criteria. LVEF and GLS were both prognostic markers of mortality. Independent markers of 1-year mortality were S. aureus IE (HR:2.02; 95%CI 1.11-5.72, p = .022), diabetes (HR:2.05; 95%CI 1.12-3.75, p = .020), embolic stroke (HR:3.95; 95%CI 1.93-8.10, p < .001) and LVEF<45% (HR: 3.02; 95% CI 1.70-5.38, p < .001), GLS> -15.4% (HR:2.95; 95%CI 1.52-5.72, p < .001). Adding LVEF<45% to a model with known risk factors of IE did not significantly improve risk classification, whereas addition of GLS to the model resulted in significant increase (AUC = 0.763, p < .001). CONCLUSIONS: When treatment was taken into account, LVEF<45% and GLS > -15.4% were both associated with adverse long-term outcome in left-sided IE. GLS >-15.4 % was significantly associated with 1-year mortality in the multivariate analysis. Further, GLS was superior to LVEF in risk prediction and risk discrimination of long-term outcome in patients with left-sided IE.