Variability and Resource Utilization of Bedside Three-dimensional Echocardiographic Quantitative Measurements of Left Ventricular Volume in Congenital Heart Disease

Abstract

This study evaluated the variability and time resource utilization of bedside 3-dimensional echocardiographic left ventricular volume analysis (3D-LVVA) in congenital heart disease (CHD). Background. There are currently limited data on the resource utilization and variability of 3D-LVVA in the CHD. Four reviewers of varying experience levels were timed performing 15 on-scanner 3D-LVVAs. Inter- and intraobserver variability for left ventricular end-diastolic volume (LVEDV), end-systolic volume (LVESV), and ejection fraction (LVEF) was evaluated. Median age was 12.7 years (0.6-33 years). Diagnoses were: normal (n = 4), cardiomyopathy (n = 4), ventricular septal defect (n = 2), and atrioventricular canal, tricuspid atresia, bicuspid aortic valve, left ventricular hypertrophy, and heart transplant (n = 1 each). For interobserver variability, intraclass correlation coefficients (ICCs) for all possible combinations of reviewers were: LVEDV, 0.991-0.999 (P < .01); LVESV, 0.98-0.99 (P < .01); LVEF, 0.95-0.98 (P < .01). Bland-Altman plot mean differences (+/-2SD) were: LVEDV, -3 +/- 14%; LVESV, -5.4 +/- 21.4%; LVEF, 1.2 +/- 14.7%. Interobserver variability of LVESV was not dependent on ventricular volumes (P = .25; r(2) = 0.01) or heart rate (P = .43; r(2) = 0.003). For intraobserver variability, ICCs for 2 reviewers were LVEDV, 0.99, 0.99 (P < .01); LVESV, 0.99, 0.99 (P < .01); and LVEF, 0.94, 0.94 (P < .01), respectively. Bland-Altman plot mean differences (+/-2SD) were: LVEDV, -1 +/- 9.2%; LVESV, 0 +/- 19.6%; LVEF, -2.2 +/- 24%. Reviewers with varying experience levels can accomplish 3D-LVVA at the bedside with acceptable inter- and intraobserver reproducibility, providing the rationale for integrating 3D-LVVA into the care of CHD patients.