LAUSR

Background: Left ventricular (LV) volumes and ejection fraction (EF) from real-time three-dimensional echocardiography (RT3DE) hasbeen shown to be superior to measurements obtained from two dimensional echocardiography (2DE). However, most of these studies were conducted in research setting with high volume 3DE use, selected cohort of patients and experienced sonographers. Objectives: This study aimed to determine the accuracy and reproducibility of RT3DE and 2DE in measuring LV volumes and EF in daily clinical practice. Methods: 30 patients (age 52 ± 12 years, 24 men, 29 in sinus rhythm, 23with good acousticwindow) undergoing clinically indicated cardiac magnetic resonance (CMR) imagingwere prospectively recruited to have transthoracic 2DE and RT3DE performedwithin 4 hours after CMR. To assess inter-observer variability, 2 sonographers performed the same set of measurements independently on the same day. A subgroup of patients (n=10) was studied for intra-observer variability. CMR was the reference standard. Results: The LV end-diastolic volume (EDV), end-systolic volume (ESV), and EF measured from CMR were 194.3 ± 72.5 ml, 125.7± 69.0 ml and 37.8 ± 19.2% respectively. The biases ± SD for RT3DE were -72.7 ± 45.7 ml, - 47.6 ± 38.5 ml and 2.3 ± 9.8% for EDV, ESV and EF respectively. The biases ± SD for 2DE were - 70.5 ± 46.6 ml, - 50.8 ± 42.4 ml and 5.7± 9.5% for EDV, ESV and EF respectively. The difference in bias between RT3DE and 2DE volumeswas statistically not significant (p = 0.54 and p = 0.47 for EDV and ESV respectively). However, the difference in bias between RT3DE and 2D EF was marginally significant (p= 0.05). EF measured by CMR was similar by RT3DE (P= 0.21) but not by 2DE (P = 0.003). The inter- and intra-observer variation in volumes and EF were similar for RT3DE and 2DE. Conclusions: In daily clinical practice, RT3DE and 2DE underestimates LV volumes. Compared to 2DE, RT3DE is more accurate for EF measurement. The reproducibility of RT3DE measurements is similar to that of 2DE.