Comparison of left ventricular stroke volume assessment by two- and three-dimensional echocardiography in a swine model of acute myocardial infarction validated by thermodilution method.

Background: Accurate left ventricular stroke volume (LVSV) measurement is clinically important in patients presenting with acute myocardial infarction. Three-dimensional echocardiography (3DE) is expected to overcome limitations of two-dimensional echocardiography (2DE). However, inaccuracy in volumetry by 3DE has often been reported hindering further clinical application. This study aimed at comparing agreement and correlation with the thermodilution method (TDM) between 2DE and 3DE measurement of LVSV.

Methods: Swine model of myocardial infarction was created and LVSV was measured by 3DE by subtracting end-systolic from end-diastolic volume (3DE-method). Pulsed Doppler ultrasound and left ventricular outlet tract area were used to measure LVSV by 2DE (2DE-method). TDM was performed by the Swan-Ganz catheter. Bland-Altman analysis followed by assessment of intraclass correlation coefficient (ICC) were performed between 2DE-method and TDM as well as 3DE-method and TDM.

Results: A total of 25 comparisons revealed a significant overestimation of LVSV by the 2DE-method (bias = 6.5 mL; 95% confidence interval [CI], 3.9-9.0 mL; P < 0.0001), whereas there was no significant bias by the 3DE-method (bias =-1.6; 95% CI, -4.3 to 1.1 mL; P = 0.22). The ICC between 2DE and TDM was 0.49 (95% CI, 0.14-0.74) whereas ICC between 3DE and TDM was 0.75 (95% CI, 0.51-0.88).

Conclusions: This study elucidated that LVSV is better estimated by 3DE-method compared to the conventional 2DE-method. This investigation will provide a more accurate, quick and noninvasive way of LVSV and cardiac output assessment at bedside by further application of 3DE.