Left Ventricular End Diastolic Volume and Ejection Fraction Calculation: Correlation between Three Echocardiographic Methods

Abstract

Background: Left ventricular ejection fraction (LVEF) and end diastolic volume (EDV) are measured using Simpson's biplane (SB), 3-dimensional method (3DE), and speckle tracking (STE). Comparisons between methods in routine practice are limited. Our purpose was to compare and to determine the correlations between these three methods in clinical setting.

Methods: LVEF and EDV were measured by three methods in 474 consecutive patients and compared using multiple Bland-Altman (BA) plots. The correlations (R) between methods were calculated.

Results: Median (IQR) LVEF_SB, LVEF_STE, and LVEF_3DE were 63.0% (60-69)%, 61% (57-65)%, and 62% (57-68)%. Median (IQR) EDV_SB, EDV_STE, and EDV_3DE were 85 ml (71-106) ml, 82 ml (69-100) ml, and 73 ml (59-89) ml. R between LVEF_SB and LVEF_3DE was 0.65 when echogenicity was good and 0.43 when poor. R for EDV_SB and EDV_3DE was 0.75 when echogenicity was good and 0.45 when poor. On BA analysis, biases were acceptable (<3.5% for LVEF) but limits of agreement (LOA) were large: 95% of the differences were between -15.4% and +18.8% for LVEF as evaluated by SB in comparison with 3DE, with a bias of 1.7%. In the comparison EDV_SB and EDV_3DE, the bias was 14 ml and the LOA were between -24 ml and +53 ml. On linear regressions, LVEF_3DE = 17.92 + 0.69 LVEF_SB and EDV_3DE = 18.94 + 0.63 EDV_SB.

Conclusions: The three methods were feasible and led to acceptable bias but large LOA. Although these methods are not interchangeable, our results allow 3DE value prediction from SB, the most commonly used method.

Figure 1

LVEF (a) and EDV (b) values according to the method used. Box-plot representation showing the distribution of LVEF values according to the method which was used. The inbox line represents the mid value. The inbox circle represents the mean value. The edges of the box represent the 25th and 75th percentiles (Q1 and Q3), and the ends of the whiskers represent the upper and lower adjacent values, which are the most extreme values within Q3 + 1.5 ∗ (Q3 − Q1) and Q1 − 1.5 ∗ (Q3 − Q1), respectively. At first sight, 3DE values seem to be the lowest. Box-plot representation showing the distribution of EDV values according to the method which was used. At first sight, STE values seem to be the lowest.

Figure 2

Simple linear regression analysis between LVEF_3DE and LVEF_SB (a) and LVEF_STE and LVEF_SB (b). LVEF_3DE = 17.92 + 0.69 LVEF_SB. LVEF_STE = 33.84 + 0.42 LVEF_SB.

Figure 3

Simple linear regression analysis between EDV_3DE and EDV_SB (a) and EDV_STE and EDV_SB (b). EDV_3DE = 18.94 + 0.63 EDV_SB. EDV_STE = 10.51 + 0.84 EDV_SB.

Figure 4

Bland–Altman plots for the comparisons between (a) LVEF_SB and LVEF_3DE, (b) LVEF_SB and LVEF_STE, and (c) LVEF_3DE and LVEF_STE. Bias: 1.66% (95% CI 0.86–2.47); bias: 3.49% (95% CI 2.77–4.20); and bias: 1.76% (95% CI 0.86–2.65), respectively.

Figure 5

Bland–Altman plots for the comparisons between (a) EDV_SB and EDV_3DE, (b) EDV_SB and EDV_STE, and (c) EDV_3DE and EDV_STE. Bias: 14.19 ml (95% CI 12.38–16.00); bias: 3.49 ml (95% CI 2.23–4.75); and Bias: 10.58 ml (95% CI -12.46 to -8.71), respectively.