BACKGROUND We aimed to determine the accuracy of the echocardiographic assessment of cardiac index (CI) in subjects with preserved left ventricular ejection fraction (LVEF). METHODS Thirty-three subjects with LVEF >50%, normal sinus rhythm, and a broad spectrum of hemodynamic profiles underwent echocardiography immediately followed by right heart catheterization. As gold standards, CI was assessed using thermodilution [CI (TD)] and the Fick method [CI (F)]. Echocardiographic CI was assessed by four methods: from the left ventricular outflow tract (LVOT) velocity time integral and the LVOT diameter as measured [CI (LVOTm)] as well as estimated from body surface area [CI (LVOTe)], and from stroke volume indices assessed using the biplane [CI (BP)] and monoplane [CI (MP)] methods. RESULTS The mean CI (TD), CI (F), CI (LVOTm), CI (LVOTe), CI (BP), and CI (MP) were 3.0 ± 0.9, 3.1 ± 0.7, 2.8 ± 0.6, 3.3 ± 0.6, 2.0 ± 0.6, and 2.2 ± 0.7 L/min/m(2) . There were modest correlations between CI (TD) and CI (F) and all four noninvasive measures of CI with r(2) values ranging from 0.09 to 0.30. CI (LVOTm) underestimated CI (TD) and CI (F) by 0.3 and 0.3 L/min/m(2) , CI (LVOTe) overestimated CI (TD) and CI (F) by 0.3 and 0.2 L/min/m(2) , and CI (BP) and CI (MP) underestimated CI (TD) and CI (F) by 1.1 and 1.1 L/min/m(2) and 0.9 and 0.9 L/min/m(2) , respectively, with large limits of agreement for all comparisons. CONCLUSIONS In subjects with nondilated left ventricles with preserved LVEF, flow- or volume-based measures of CI by 2D echocardiography may not accurately reflect CI (TD) and CI (F). Further larger studies are required to verify our findings and to evaluate the accuracy of contrast and 3D echocardiography in this setting.