C 2012, Wiley Periodicals, Inc.
DOI: 10.1111/j.1540-8175.2012.01695.x
Two-Dimensional Color Doppler Echocardiography for Left Ventricular Stroke Volume Assessment: A Comparison Study with Three-Dimensional Echocardiography Cristina Da Silva, B.Sc.,∗ F´atima Pedro, B.Sc.,∗ Lizandra Deister, B.Sc.,∗ Anders Sahl´en, M.D., Ph.D., M.R.C.P. (U.K.),† Aristomenis Manouras, M.D., Ph.D.,† and Kambiz Shahgaldi, B.Sc., Ph.D.‡ ∗ Department of Clinical Physiology, †Department of Cardiology and Clinical Physiology, and ‡Department of Cardiology, Karolinska University Hospital, Huddinge, Stockholm, Sweden
Aim: Whether measurement of left ventricular outflow tract diameter (LVOTd) using color Doppler (CD) in order to more accurately define LVOTd is more accurate for determination of stroke volume (SV) than gray scale and compare it with direct measurement of LVOT area (a) using three-dimensional echocardiography (3DE) for SV determination. Methods and Results: Twenty-one volunteers were examined. LVOTa was calculated by two-dimensional echocardiography (2DE) using the following formula: π × (d/2)2 , d = LVOT diameter by gray scale and CD, respectively. Planimetry of LVOTa was performed in parasternal long axis using 3DE. Eccentricity Index was calculated using the lateral and anterior-posterior LVOTd. SV was obtained by four different methods: (1) 2D gray scale, (2) 2D color, (3) LVOTa × LVOT velocity time integral, and (4) SV by Simpson’s biplane method. Gray scale LVOTd was significantly smaller compared to LVOTd obtained with CD (P < 0.05). Significant differences occurred between LVOTa gray scale and CD (3.29 ± 0.74 cm2 vs 3.67 ± 0.70 cm2 , P < 0.05) and between LVOTa calculated by gray scale in comparison to 3DE planimetry; (3.29 ± 0.74 cm2 vs 3.61 ± 0.89 cm2 , P = 0.011). Half of the subjects had at least 17% difference between the lateral and anterior-posterior LVOTd. There were significant differences between SV by 2D gray scale and 2D CD (82.8 ± 17.1 mL vs 92.4 ± 16.8 mL, P < 0.05) and between 2D gray scale and 3DE planimetry (82.8 ± 17.1 mL vs 90.7 ± 19.8 mL, P = 0.025). Conclusion: Our study demonstrates LVOT being frequently elliptical. SV and LVOTa were found to be similar when comparing 2DE CD and 3DE planimetry and showed higher values in comparison to 2DE gray scale, which suggests 2DE CD to be an alternative approach for SV assessment. (Echocardiography 2012;29:766-772) Key words: left ventricular outflow tract, left ventricular stroke volume, color Doppler, threedimensional echocardiography
Stroke volume (SV) is an important hemodynamic parameter in clinical medicine providing a guide to the circulatory conditions as well as an evaluation of the response to therapeutic interventions. Two-dimensional echocardiography (2DE) is currently the most widely used noninvasive imaging tool for the anatomical and functional analysis of the heart including the assessment of aortic valve area (AVA) and SV.1–4 Several noninvasive 2DE methods can be used for SV calculation. The difference between enddiastolic (ED) and end-systolic chamber volumes or left ventricular outflow tract (LVOT) pulsed wave (PW) Doppler velocity time integral (VTI) multiplied by LVOT area (a) for online or offline Address for correspondence and reprint requests: Kambiz Shahgaldi, B.Sc., Ph.D., Department of Cardiology, Karolinska University Hospital Huddinge, 141 86 Stockholm, Sweden. Fax: 08-58586700; E-mail:
[email protected]
766
measurements are two of them. Volume calculations using the Simpson’s biplane formula by subtracting left ventricular (LV) ED volume (V) and LVESV for SV assessment are prone to error, since there might be difficulties in defining LV borders or abnormal ventricular geometry, foreshortening due to difficulties acquiring images from the true apex. The SV calculation from the Simpson’s biplane rule could lead to inaccurate results if there is a mitral insufficiency. SV determination from LVOT diameter (d) makes several statements, such as (1) the outflow tract has a circular configuration and (2) its dimension remains stable during systole.5 The major difficulty in assessment of SV by calculating LVOTd is that the diameter is squared in order to determine SV and thus small errors in measuring LVOTa are amplified.5–7 Three-dimensional echocardiography (3DE) assessment of the LV function has proven to
2D Color Doppler Echocardiography for LV Stroke Volume Assessment
TABLE I. Patient Characteristics (n = 21) Variables
Mean ± S D
Age (yrs) Males (%) Females (%) Height (cm) Weight (kg) BSA (m2 ) SBP (mmHg) DBP (mmHg) Resting HR (bpm) EF by RT3DE (%)
24 ± 5 38 62 171.4 ± 8.9 70.1 ± 12.0 1.8 ± 0.19 115 ± 13 78 ± 10 59 ± 7 59 ± 2
BSA = body surface area; SBP = systolic blood pressure; DBP = diastolic blood pressure; HR = heart rate; RT3DE = real time three-dimensional echocardiography.
be highly accurate when compared with magnetic resonance imaging in different patient categories.8,9 This approach might be able to eliminate a serious limitation for the calculation of LVOTa as it permits the direct measurement and assessment of the anatomic reality of this cardiac region.1,4,6,9 3DE has also proved to be superior to 2DE methods for chamber volume calculation.1,12 We tested the hypothesis that measurement of LVOTd on color Doppler (CD) superimposed images might be more accurate for determination of SV compared to gray scale determination of LVOTd. These measurements were compared to direct LVOTa planimetry by 3DE for SV assessment which was used as reference method. Methods: Patient Selection: Twenty-one volunteers (8 men and 13 women, aged 24 ± 5 years) with good to excellent acoustic window and no known cardiac disease were examined. Exclusion criteria were patients with atrial fibrillation/or flutter, poor image quality, and abnormal valvular function (regurgitation). All the participants signed an informed consent allowing the procedures. Before each examination, age, height, weight, and blood pressure were registered. Cardiovascular risk factors were collected with a questionnaire. Table I shows baseline characteristics of the included volunteers. The study was approved by the local ethical committee, Stockholm, Sweden. Imaging Acquisition: A complete 2DE and Doppler study was performed in all subjects, using a commercially available Vivid E9 (GE Healthcare, Horten, Norway) ultrasound machine equipped with M5S transducer. All acquisitions were performed by the
same operator with the subject in the left lateral position. VTI of the subvalvular flow was recorded from an apical five-chamber view by placing the sample volume (5 mm) of PW Doppler at the base of the aortic leaflets and then moving slowly away toward LVOT until a typical subvalvular flow profile was seen (0.5 mm on the LV side of the aortic valve). An optimal signal shows a smooth velocity curve with a narrow velocity range at each time point. LVOT zoomed image with and without CD were acquired. The CD (mean aliasing velocity of 60 ± 5 cm/sec) with default settings with a minimum operator adjustment was used in order to achieve standardization of image quality to better define and measure LVOTd in order to calculate SV. Apical two-chamber and four-chamber view were recorded for assessment of SV by Simpson’s biplane method. A novel 4V matrix-array transducer was used for the acquisition of the 3DE data sets from four consecutive cardiac cycles during end expiration breath-hold in parasternal long-axis view (PLAX) and in apical view. This transducer allows the use of the “full-volume” technique with a mean frame rate of 40 ± SD volumes/sec. All the above-described images were acquired twice and best image quality was chosen for further analysis. Data sets were stored for offline analysis using commercially available software (EchoPAC PC version 110.0.0, GE Healthcare). Image Analysis The image analysis involved the measurement of several parameters which were done according to the American Society of Echocardiography guidelines.3 Each parameter was measured twice by two independent readers with 1 week apart blinded to the previous reading. Each observer made three consecutive measurements and the mean value was recorded. In PLAX, gray scale and color LVOTd were measured 5 mm above the aortic valve in mid systole, according to trailing edge to leading edge5 (Fig. 1A and B). VTI was measured by tracing the spectral PW Doppler velocity signal (Fig. 1C) using a minimum gain setting and the sweep velocity was set >100 mm/sec in order to obtain a better delineation. This parameter and the approximated cross-sectional area of the LVOT were used for further calculation of SV by following equation: SV = π × (d/2)2 × LVOTVTI . Manual tracing of endocardial border using Simpson’s biplane method was performed to obtain ejection fraction (EF), LVEDV, LVESV, and SV. After modifying the transverse plane orientation of PLAX view obtained by 3DE, an idealized LVOT short-axis view (SAX) was achieved and LVOTa was measured 5 mm above the 767
Da Silva, et al.
Figure 1. A and B. Two- and three-dimensional echocardiography with and without color Doppler illustrating the LVOT diameter (d) measurement. As can be seen the gray scale image shows an LVOTd smaller than the corresponding color Doppler image (20 and 21 mm, respectively). The spectral pulsed wave Doppler of LVOT with manual tracing is shown C. Using the LVOTd in gray scale and in color Doppler image the SV is measured to 78 and 86 mL, respectively.
aortic valve by manually tracing the outline of LVOT (Fig. 2A). LVOT lateral and anteriorposterior diameters were also measured. Anteriorposterior LVOTd defined as the diameter parallel (90◦ ) to left atrium (LA) and the lateral LVOTd perpendicular to LA was calculated. These diameters were used to calculate Eccentricity Index (EI)4 , EI = 1 – (anterior-posterior LVOTd )/(lat LVOTd ) (Fig. 2B). An index of zero would represent a circular LVOT, where a higher EI describes a noncircular geometry. SV was obtained by four different methods (1) 2D gray scale SV = π × (LVOTd/2)2 × LVOTVTI , (2) 2D color SV = π × (LVOTd/2)2 × LVOTVTI , (3) 3DE LVOTa × LVOTVTI , and (4) SV assessed by Simpson’s biplane method. Statistical Analysis: Statistical analysis was performed using commercially available software (SPSS for Windows, ver-
sion 16; SPSS Inc, Chicago, IL, USA) and Microsoft Office Excel 2007. Intra- and interobserver variability were calculated according to the following formula: √ (SDdiff × 100%)/total mean × 2, where SDdiff is standard deviation of the difference between measurements.11 All variables were tested for normality using the Shapiro–Wilk test. Values of P < 0.05 were considered statistically significant. The Bland– Altman analysis was performed to determine the systemic bias and limits of agreement of all measurements by each method.13 Group comparison between continuous variables was performed using ANOVA. Paired sample t-test was performed to compare the significance of difference [LVOTd gray scale – LVOTd color] variable. Wilcoxon test was performed to compare the lateral and the anteroposterior LVOTd.
Figure 2. A. (Same patient as in Fig. 1) Planimetry of LVOT 5 mm above the aortic valve is determined by RT3DE resulting in an area of 3.5 cm2 after manipulating with the transverse plane. SV is calculated to 87 mL by multiplying LVOT area with LVOT velocity time integral. The lateral and anterior-posterior LVOT diameters are measured B.
768
2D Color Doppler Echocardiography for LV Stroke Volume Assessment
TABLE II Different Variables and Their Intra- and Interobserver Variability
Intra 1 Intra 2 Inter 1–2
Intra 1 Intra 2 Inter 1–2
2DE LVOTd Gray Scale
LVOTd Color
LVOT VTI
1.4% 1.5% 2.1%
1.4% 1.7% 2.5%
1.3% 1.7% 2.5%
RT3DE LVOT Area
LVOTd Lateral
LVOTd Antero-Posterior
SV
5.0% 1.7% 3.6%
3.2% 2.3% 3.0%
4.0% 2.1% 3.3%
2.9% 3.9% 2.8%
2DE = two-dimensional echocardiography; d = dimension; LVOT = left ventricle outflow tract; VTI = velocity time integral; SV = stroke volume.
Results: Intra and Interobserver Variability: There was good agreement within (intraobserver variability) and between (interobserver variability) the two independent observers’ measurements of all the studied parameters. LVOTVTI was the parameter with least intraobserver variability (1.3%) whereas the highest variability was found for LVOTa by 3DE planimetry (5%) (Table II). Interobserver variability ranged from 2.1% (LVOTd gray scale) to 3.6% (LVOTa) (Table II). Comparison of Color Doppler and Gray Scale LVOT Diameter: Mean LVOTd obtained by gray scale and CD was 20.4 ± 2.1 mm and 21.5 ± 1.9 mm, respectively. Paired sample t-test showed significant differences between the two LVOTd (P < 0.05). Comparison of LVOT Areas: Mean LVOTa calculated by gray scale, CD, and 3DE planimetry and the comparison of mean SV between the four different methods are presented in Table III. Significant differences were observed between LVOTa gray scale and CD (P < 0.05) and also between LVOTa gray scale and 3DE planimetry (P = 0.011). The difference between CD and TABLE III Mean Values of LVOTa and SV Obtained from 2DE and 3DE Mean Difference ± SD
2D Gray scale 2D color Doppler RT3DE planimetry 2D Simpson’s method
LVOTa (cm2 )
SV (mL)
3.29 ± 0.74 3.67 ± 0.70 3.61 ± 0.89
82.8 ± 17.1 92.4 ± 16.8 90.7 ± 19.8 79.3 ± 15.2
3DE planimetry did not reach the level of statistical significance. Comparison of Stroke Volumes: Mean difference, limits of agreement, and the corresponding r- and P-value for SV by the four different methods are illustrated in Tables III and IV and Figures 3 and 4. LVOT Eccentricity: The mean value of the lateral and the anteriorposterior LVOTd diameter was 22.8 ± 2.7 mm and 18.8 ± 2.5 mm, respectively. Significant difference occurred between these diameters with mean difference value of 4.05 ± 1.64 mm (P < 0.05). The mean EI was 0.21 and median of 0.17. Discussion: Current recommendations by the American Society of Echocardiography advocate that in order to calculate AVA by continuity equation (CE) either the aortic annular or LVOT diameters should be used.9 Since LVOT may have an elliptical shape, the LVOTd measured in 2DE gray scale is frequently smaller, resulting in significant underestimation of the true cross-sectional area.1,2,4,9,10 The results of this study show that LVOT is not a circular structure which is consistent with many other investigations.1,2,4,9,10 EI was 0.21 ± 0.13 and its median 0.17, which means that half of the subjects had at least 17% difference between the lateral and the anterior-posterior LVOTd. Similar results were founded by Doddamani et al. that showed a mean and median EI of 0.21 and 0.20, respectively.4 Direct planimetry using 3DE is increasingly popular as this technique enables actual LVOT area to be measured without relying on the assumption of circularity and avoids error induced by squaring the measured LVOT radius7 . 769
Da Silva, et al.
TABLE IV Mean Difference and Limits of Agreement of SV Assessed by 2DE and 3DE
2D gray scale vs 2D color 2D gray scale vs. RT3DE plan 2D gray scale vs. Simpson’s method 2D color vs. RT3DE plan 2D color vs. Simpson’s method RT3DE plan vs. Simpson’s method
Limits of Agreement
Mean + SD (mL) −9.7 ± 4.9
−19.5
0.2
P-Value 0.001
−7.9 ± 10.5 3.5 ± 10.9 1.7 ± 12.2 13.2 ± 11.3 11.4 ± 13.7
−29.0 −18.2 −22.8 −9.4 −15.9
13.1 25.3 26.2 35.7 38.8
0.025 1.000 1.000 0.001 0.010
2D = two-dimensional; RT3DE = real time three-dimensional echocardiography; plan = planimetry.
Our group has previously showed that direct planimetry of LVOT area is an accurate and feasible method and correlates extremely well with real time 3D echocardiography (RT3DE) full volume measurements for SV assessment.2 Ng et al. demonstrated that both multislice computed tomography and 3D transesophageal echocardiography showed an elliptical aortic annular/LVOT geometry, leading to significant underestimations of their “true” cross-sectional areas when assuming a circular geometry.9 According to our results there are significant differences between LVOTa measured by 2DE gray scale and 3DE planimetry (P = 0.011). Despite the fact that 2DE CD involves geometric assumptions, there were no significant differences observed between this method and 3DE planimetry. Even if no statistical significance was found, the limits of agreement were rather wide (–0.96 to 1.07 cm2 ) which may be of interest in the clinical setting. Nevertheless our results suggest that 2DE CD is in better agreement with 3DE planimetry compared to the corresponding of 2DE gray scale in LVOTd assessment. According to this finding another question could be raised, if it is possible that LVOTd assessed by 2DE CD overestimate the diameter. In that case the calculated area by CE using this parameter is similar to direct planimetry in elliptical LVOT but could lead in overestimation in circular LVOT. Determination of LV volumes and EF from 2DE images may be relatively inaccurate because of foreshortened views and the use of geometric as-
sumptions.2,13 Regarding SV measurement by the different methods, higher values were obtained with 3DE LVOT planimetry and 2DE CD. In fact, 2D gray scale resulted in a systematic underestimation of SV compared with the corresponding 3D and 2D CD measurements. This is not surprising considering the fact that 2D gray scale is based on geometric assumptions of LVOT that do not necessarily correspond to LVOT anatomy. On the other hand, color guidance provide better delineation of the LVOT and yields consistently larger LVOT diameter in better agreement with measurements performed with direct planimetric 3D analysis. As provided in Table V, not only the mean difference between 3D planimetry and 2D CD measurements was lower than when these measurements were compared to 2D but the limits of agreement were narrower as well. Additionally there was no significant differences between 3D planimetry and 2D CD (P = 1.000). However, despite this, regarding SV measurements the limits of agreement were relatively wide between measurements obtained by these two novel approaches, i.e., 3D and color superimposition on 2D images, which should be considered in clinical praxis especially in patients with low output states. In a previous study it was shown that SV measurements with 3D planimetry were in good correlation with full 3D volume measurements and the limits of agreement between these two methods were narrow.2 Furthermore, Khaw et al.
TABLE V Mean Difference and Limits of Agreement of LVOTa Analysis with 2DE and 3DE Gray Scale vs. Color Doppler (cm2 ) Mean difference ± SD Limits of agreement P-value r2 -value
–1.2 ± 0.6 −2.4 0.001 0.92
0.02
Gray Scale vs. RT3DE Plan (cm2 ) –0.3 ± 0.5 −1.3 0.011 0.72
0.7
RT3DE = real time three-dimensional echocardiography; plan = planimetry; SD = standard deviation.
770
Color Doppler vs. RT3DE Plan (cm2 ) 0.06 ± 0.5 −0.94 1.000 0.62
1.06
2D Color Doppler Echocardiography for LV Stroke Volume Assessment
Figure 3. Bland–Altman graphs displaying differences of average values between 2DE and 3DE derived SV between 2D gray scale and 2D color Doppler A., between 2D gray scale and biplane Simpson’s (BPS) method B., and between 2D color Doppler and 3D planimetry C. The central line represents mean difference and the dashed lines represent ±SD from the mean (measurements in mL).
and Blot-Souletie et al. support these findings demonstrating the direct measurement of the LVOTa from 3DE provides accurate assessment of SV.14,15 However, SV measurements using 3DE require expertise, expensive equipments, and dedicated software which may limit its use in clinical
Figure 4. A and B. Correlation plots between SV assessed by 2D gray scale and 2D color Doppler against 3D planimetry (A and B) and correlation plot between 2D gray scale against 2D color Doppler (C) are illustrated.
practice. Taken into consideration this, LVOTd assessed by 2DE CD is easier to perform in everyday clinical practice in most echocardiography laboratories and could be considered as an alternative approach for hemodynamic measurements of cardiac output and SV. In addition this study was performed in patients with structurally normal aortic valves. In 771
Da Silva, et al.
clinical practice, pathological findings as valvular calcification, with consequent LVOT shadowing or interventricular basal septal hypertrophy, could lead to difficulty in order to measure LVOTd.16 In these cases the use of CD might improve the LVOTd assessment since it can provide better delineation of LVOT borders. This is especially important given the current interest in conditions such as paradoxical low-flow, low-gradient severe aortic stenosis.16 Future studies are needed in order to elucidate end validate the use of the method described in this report in this patient category. Limitations: All measurements from the current study presented a very low variability. This finding could be due to the good to excellent image quality in our young and healthy study population. Thus, it is expected that in clinical practice, variability will assume higher values.5 The study sample was small in dimension and included only normal subjects. The echocardiography image acquisition and the offline analysis were performed by inexperienced observers which may affect the results. Nevertheless, interand intra-observer variability was low for all the measured variables. The same acquired images were used to make the analysis of the multiple parameters which does not account for variability that would be observed if the same subject was reimaged. Measurements with different modalities were not performed in the same cardiac cycle. So, minor variations from beat to beat cannot be excluded. However, all individuals in the present study were in sinus rhythm without any significant variations in heart frequency. Conclusion: Our study showed that LVOT shape is usually not round. Incorrectly assuming a circular LVOT geometry using the CE based on 2DE gray scale can underestimate LVOTa and consequently SV. Our results suggest that 2DE CD is similar to 3DE planimetry regarding LVOTa and SV calculation than 2DE gray scale. The use of 2D CD in LVOTd measurement in order to calculate SV and AVA could be considered as an option in daily practice, whenever the use of direct planimetry is not possible. References 1. Poh KK, Levine RA, Solis J, et al: Assessing aortic valve area in aortic stenosis by continuity equation: A novel approach using real-time three-dimensional echocardiography. European Heart J 2008;29(20):2526–2535.
772
2. Shahgaldi K, Manouras A, Brodin LÅ, et al: Direct measurement of left ventricular outflow tract area using threedimensional echocardiography in biplane mode improves accuracy of stroke volume assessment. Echocardiography 2010;27(9):1078–1085. 3. Lang RM, Bierig M, Devereux RB, et al: Recommendations for chamber quantification: A report from the American Society of Echocardiography’s Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005;18:1440–1463. 4. Doddamani S, Bello R, Friedman MA, et al: Demonstration of left ventricular outflow tract eccentricity by real time 3D echocardiography: Implications for the determination of aortic valve area. Echocardiography 2007;24(8):860– 866. 5. Baumgartner H, Hung J, Bermejo J, et al: Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice. Eur J Echocardiogr 2009;10:1– 25. 6. Guti´errez-Chico JL, Zamorano JL, Prieto-Moriche E, et al: Real-time three-dimensional echocardiography in aortic stenosis: A novel, simple, and reliable method to improve accuracy in area calculation. Eur Heart J 2008;29(10):1296–306. 7. Adegunsoye A, Mundkur M, Nanda NC, et al: Echocardiographic evaluation of calcific aortic stenosis in the older adult. Echocardiography 2011;28(1):117–129. 8. Van den Bosch AE, Robbers-Visser D, Krenning BJ, et al: Real-time transthoracic three-dimensional echocardiographic assessment of left ventricular volume and ejection fraction in congenital heart disease. J Am Soc Echocardiogr 2006;19(1):1–6. 9. Ng AC, Delgado V, van der Kley F, et al: Comparison of aortic root dimensions and geometries before and after transcatheter aortic valve implantation by 2- and 3dimensional transesophageal echocardiography and multislice computed tomography. Circ Cardiovasc Imaging 2010;3(1):94–102. 10. Perez de Isla L, Zamorano J, Perez de la Yglesia R, et al: Quantification of aortic valve area using three-dimensional echocardiography. Rev Esp Cardiol 2008;61:494–500. 11. Dahlberg G: Statistical methods for medical and biological studies. New York: Interscience Publications, 1940. 12. Bland JM, Altman DG: Statistical methods for assessing agreements between two methods for clinical measurements. Lancet 1986;1:307–310. 13. Jacobs LD, Salgo IS, Goonewardena S, et al: Rapid online quantification of left ventricular volume from realtime three-dimensional echocardiographic data. Eur Heart J 2006;27:460–468. 14. Khaw AV, von Bardeleben RS, Strasser C, et al: Direct measurement of left ventricular outflow tract by transthoracic real-time 3D-echocardiography increases accuracy in assessment of aortic valve stenosis. Int J Cardiol 2009;136:64–71. 15. Blot-Souletie N, Hebrard A, Acar P, et al: Comparison of accuracy of aortic valve area assessment in aortic stenosis by real time three-dimensional echocardiography in biplane mode versus two dimensional transthoracic and transesophageal echocardiography. Echocardiography 2007;24:1065–1072. 16. Hachicha Z, Dumesnil JG, Bogaty P, et al: Paradoxical lowflow, low-gradient severe aortic stenosis despite preserved ejection fraction is associated with higher afterload and reduced survival. Circulation 2007;115(22):2856–2864.
