Assessment of myocardial longitudinal function has proved to be a sensitive marker of deteriorating myocardial function in aortic stenosis, demonstrated by both color Doppler tissue imaging and recently by two-dimensional speckle-tracking echocardiography. The aim of this study was to compare velocity (color Doppler tissue imaging) and deformation (two-dimensional speckle-tracking echocardiography) in relation to global and regional longitudinal function in asymptomatic and severe symptomatic aortic stenosis.
In a cross-sectional design, 231 patients with aortic stenosis were divided into four groups: asymptomatic moderate aortic stenosis (aortic valve area, 1.0–1.5 cm 2 ; n = 38), asymptomatic severe aortic stenosis (aortic valve area < 1.0 cm 2 ; n = 66), and symptomatic severe aortic stenosis with preserved ( n = 68) and reduced (<50%) left ventricular ejection fraction ( n = 59).
Among all global (peak systolic s′, diastolic e′ and a′, longitudinal displacement, and global longitudinal strain and strain rate) and regional longitudinal (basal, middle, and apical longitudinal strain and strain rate) parameters, only diastolic e′, longitudinal displacement, and basal longitudinal strain (BLS) remained significantly associated with symptomatic status, independent of age, gender, heart rate, aortic valve area, stroke volume index, left ventricular mass index, left atrial volume index, and tricuspid annular systolic plane excursion. Furthermore, in a model with the aforementioned parameters, including e′, longitudinal displacement, and BLS, only BLS remained significantly associated with symptomatic status in the entire study population (BLS per one-unit decrease: odds ratio, 1.23; 95% CI, 1.04–1.46; P = .017). Furthermore, patients with BLS < 13% were more likely to be symptomatic (odds ratio, 4.97; 95% CI, 2.6–9.4; P < .001), and no patients with asymptomatic severe aortic stenosis with BLS ≥ 13% were admitted with myocardial infarction or heart failure during follow-up of 1,462 days.
Among the many echocardiographic measures of longitudinal velocity and deformation, BLS has the strongest association with symptomatic status in aortic stenosis, and BLS < 13% is related to adverse outcomes in severe asymptomatic aortic stenosis.
Cardiac long-axis function has been intensively investigated over the more than three decades since the 1979 study by Dumesnil et al. demonstrated that longitudinal shortening, measured by M-mode echocardiography, was reduced in patients with aortic stenosis compared with normal subjects and patients with aortic or mitral regurgitation. Since the era of M-mode echocardiography, many new methods have been added to the echocardiographic toolbox, and today we can assess myocardial performance with both tissue Doppler velocities and two-dimensional (2D) speckle-tracking echocardiography in relation to systolic and diastolic function and global, regional, and multidirectional performance.
In severe asymptomatic aortic stenosis, the single echocardiographic parameter included in the guidelines is left ventricular ejection fraction (LVEF), which when <50% provides a clear indication for aortic valve replacement. However, because of progressive changes and myocardial hypertrophy in aortic stenosis, LVEF may be preserved until late in the disease despite decreasing stroke volumes and impaired contractility. It seems intuitively correct to monitor cardiac function to identify patients with worsening aortic stenosis, and measures of longitudinal cardiac function, evaluated by color Doppler tissue imaging (DTI) and 2D STE, seem especially appealing. In contrast to conventional echocardiographic measures of systolic function, global and regional longitudinal function have proved to be sensitive markers of deteriorating myocardial function in various heart diseases, including aortic stenosis.
In previous studies, tissue velocities measured by color DTI have proved to be more robust measures of longitudinal function compared with the newer technique with 2D STE, but regional function may not be reflected accurately with color DTI, because of tethering. In contrast, 2D STE is angle independent, and regional function is not influenced by adjacent segments (tethering). Thus, the aim of this study was to compare global and regional velocity and deformation in patients with asymptomatic and symptomatic aortic stenosis, divided into four groups according to severity (aortic valve area), symptomatic status, and LVEF.
The study population of 231 patients with aortic stenosis was recruited from March 2008 to January 2012 from two parallel studies investigating the use of multidetector computed tomography and echocardiography in asymptomatic patients ( n = 104) and symptomatic patients scheduled for aortic valve replacement, with or without coronary bypass grafting ( n = 127).
Inclusion criteria for both studies were aortic valve stenosis and an echocardiographic peak velocity by continuous-wave Doppler of >2.5 m/sec, and patients with phosphocreatinine > 130 mmol/L, allergies to iodine contrast media, and other severe heart valve disease were excluded.
Group 1: Asymptomatic Cohort
The patients in the asymptomatic cohort were recruited from six hospitals in the greater Copenhagen area by screening for patients with the diagnosis registry code for aortic stenosis (DI35.0) in clinical records, hospital registries, and echocardiographic databases. One hundred twenty-four patients with asymptomatic aortic stenosis and peak velocities > 2.5 m/sec and none of the aforementioned exclusion criteria gave informed consent to participate in the study. At the time of inclusion at the University Hospital of Copenhagen, Rigshospitalet, all patients underwent transthoracic echocardiographic examinations, and asymptomatic status was confirmed by obtaining their medical histories. Twenty patients were excluded from this study because of atrial fibrillation ( n = 9), reduced LVEF (<50%; n = 4), poor image quality ( n = 5), mild aortic stenosis with aortic valve area > 1.5 cm 2 ( n = 1), and symptomatic status at the time of examination at Rigshospitalet ( n = 1). One hundred four patients with moderate to severe asymptomatic aortic stenosis, defined as an aortic valve area ≤ 1.5 cm 2 , were ultimately included in this study.
Group 2: Symptomatic Cohort
The symptomatic cohort was recruited by screening all patients referred for elective aortic valve replacement, with or without concomitant coronary bypass grafting, at University Hospital of Copenhagen, Rigshospitalet, from March 2008 to March 2010. The treating physician was blinded to the results of the echocardiographic examination, and referral for aortic valve replacement was performed independently by the clinical heart team.
One hundred eighty-five patients had data sets available for 2D speckle-tracking echocardiographic and color DTI analysis. Patients with atrial fibrillation ( n = 31), ventricular arrhythmias ( n = 5), aortic valve area ≥ 1.0 cm 2 ( n = 14), or poor image quality ( n = 8) were excluded. Ultimately, 127 patients were included in the study.
The entire study cohort of 231 patients with aortic stenosis consisted of 104 patients from the asymptomatic cohort (moderate, n = 38; severe, n = 66) and 127 patients from the symptomatic cohort (severe with preserved LVEF, n = 68; severe with reduced LVEF [<50%], n = 59).
The study was approved by the local research and ethics committee (H-B-2009-027), and all patients gave individual informed consent. Data from the two cohorts concerning measurement of aortic valve area by multidetector computed tomography have been published.
Examinations that took place from March 2008 to July 2009 ( n = 66) were performed using a Vivid 7 Dimension system (GE Vingmed Ultrasound AS, Horten, Norway) with a 3.5-MHz transducer. From August 2009 to January 2012, examinations ( n = 165) were performed using a Vivid E9 ultrasound system (GE Vingmed Ultrasound AS), with a 3.5-MHz transducer. All participants were examined with conventional 2D echocardiography and pulsed-wave and color DTI. Echocardiographic cine loops were obtained by recording three consecutive heart cycles. All examinations were performed by two experienced operators. Offline analyses were performed by a single experienced operator, who was blinded to all other patient data, using EchoPAC BT version 11.1.0 (GE Vingmed Ultrasound AS).
Left ventricular diameters, wall thicknesses, relative wall thickness, and tricuspid annular plane systolic excursion were assessed and calculated as recommended. Left ventricular mass was calculated from left ventricular linear dimensions using the Devereux formula and divided by body surface area (calculated using the Du Bois formula) to obtain left ventricular mass index. Left ventricular hypertrophy was defined as a left ventricular mass index > 115 g/m 2 for men and > 95 g/m 2 for women. Left ventricular geometry (normal, concentric remodeling, or eccentric and concentric hypertrophy) was assessed according to relative wall thickness (increased if ≥0.43) and left ventricular mass index. Left ventricular volumes and left atrial volume index were estimated using the biplane method of disks (modified Simpson’s rule). LVEF was considered reduced when <50%. Stroke volume was calculated with the Doppler method; stroke volume = cross-sectional area in left ventricular outflow tract × velocity-time integral in left ventricular outflow tract.
Left ventricular outflow tract diameter was measured in midsystole in the parasternal long-axis view.
Pulsed Doppler recordings were performed in the apical three- and five-chamber views by placing the sample volume 0.5 to 1.0 cm below the aortic valve, where a clear and smooth signal could be obtained. Peak flow velocity across the aortic valve was determined in the three- or five-chamber view, where the highest velocity could be obtained. To avoid underestimation, caution was made to align the continuous-wave curser parallel with the stenotic jet. The flow signal was traced to obtain peak velocity, velocity-time integral, and mean gradient across the valve, and aortic valve area was calculated with the continuity equation. Peak transvalvular gradient was estimated using the simplified Bernoulli equation.
Severity of the aortic valve stenosis was defined with regard to aortic valve area as mild (>1.5 cm 2 ), moderate (1.0–1.5 cm 2 ), or severe (<1.0 cm 2 ), to also include patients with low gradients. Furthermore, a subanalysis was performed with regard to four flow patterns: normal flow with low gradient (stroke volume index ≥ 35 mL/m 2 and mean gradient < 40 mm Hg), normal flow and high gradient (stroke volume index ≥ 35 mL/m 2 and mean gradient ≥ 40 mm Hg), low flow and low gradient (stroke volume index < 35 mL/m 2 and mean gradient < 40 mm Hg), and low flow and high gradient (stroke volume index < 35 mL/m 2 and mean gradient ≥ 40 mm Hg). All measurements, except from measures of the maximum peak velocity, were averaged over three cardiac cycles.
Color DTI loops were obtained in three apical views at the highest possible frame rate (median, 150 frames/sec; range, 133–168 frames/sec). The sector width and depth were optimized to include the left ventricle only. For the offline analysis, temporal smoothing was set to 30 msec. Event timing was performed to define aortic valve opening and aortic valve closure, by color DTI in the four-chamber view, using a 2- to 4-cm straight M-mode line through the septal half of the anterior mitral leaflet in end-systole.
Regional longitudinal peak systolic (s′), early diastolic (e′), and late diastolic (a′) myocardial velocities and longitudinal displacement during ejection were measured by placing a 6-mm circular region of interest at the end of ejection at the most basal myocardium of the six mitral annular sites: septal, lateral, anterior, inferior, posterior, and anteroseptal. Systolic longitudinal displacement during ejection time is the absolute distance moved by that particular point and was assessed by using the Tissue Tracking modality of the EchoPAC software and calculated as the integral of the velocity curves during ejection. Regional measures from the six mitral annular sites were averaged to provide global estimates of s′, e′, a′, and longitudinal displacement.
Deformation: 2D Speckle-Tracking Analysis
Two-dimensional speckle-tracking echocardiographic analysis was performed on grayscale images from the three apical views and the short-axis midventricular view, with the frame rate optimized to 50 to 90 frames/sec (median, 74 frames/sec; range, 57–81 frames/sec). The endocardial borders were tracked automatically when two mitral annular points and a third point in the apex were manually positioned in each apical view. Tracking quality was confirmed visually, and segments that failed to be tracked by the software were adjusted manually as needed. Segments that could not be tracked properly after manual adjustment were rejected. Regional longitudinal strain values from the 18-segment model were averaged to obtain global longitudinal strain (GLS) and global longitudinal strain rate. Regional strain and strain rate values from the six basal, middle, and apical segments were averaged to obtain basal longitudinal strain rate (BLS) and middle and apical longitudinal strain and strain rate. Regional strain values from the six short-axis midventricular segments were averaged to obtain circumferential strain and strain rate and radial strain and strain rate.
Clinical Follow-Up in Patients with Severe Asymptomatic Aortic Stenosis
By September 2013, information on mortality and hospital was obtained from the electronic health record by a systematic review of hospital contacts (outpatient visits and acute admissions) after the baseline examination in the 66 patients with severe asymptomatic aortic stenosis. The reviewer was blinded to all echocardiographic data. The combined endpoint was defined as admissions for heart failure, myocardial infarction, and all-cause mortality. Follow-up was 100% complete. During the follow-up period (1,462 days), 12 patients reached the combined endpoint (heart failure, n = 6; myocardial infarction, n = 1; and all-cause mortality, n = 5).
Inter- and Intraobserver Variability Analysis
Inter- and intraobserver variability of GLS, longitudinal displacement, and BLS were assessed independently by two experienced echocardiographers, blinded to clinical data and previous measurements, in 40 randomly selected patients (20 asymptomatic and 20 symptomatic patients). Intraobserver analysis was performed with >1 month between the repeated measures.
All strain measures are reported as absolute values throughout the text.
Normal distribution was demonstrated by Q-Q plots and normality tests (Kolmogorov-Smirnov and Shapiro-Wilk). Continuous variables with normal distributions are expressed as mean ± SD, and those not normally distributed are presented as median (interquartile range).The four groups were compared for statistical significance using one-way analysis of variance and post hoc Bonferroni (parametric) and Kruskal-Wallis (nonparametric) tests.
Comparison of groups was performed using Student’s t tests and Mann-Whitney U tests. Categorical variables are presented as percentages and were compared by using Fisher exact tests. Logistic and linear regression analysis was performed to evaluate the associations of global and regional longitudinal parameters with symptomatic status and aortic valve area. Cumulative percentage plots were constructed to illustrate the different distributions of longitudinal displacement, e′. and BLS across three groups (severe asymptomatic aortic stenosis and symptomatic with or without preserved LVEF) graphically.
Variability is expressed as mean difference ± SD and coefficient of variation (CV), calculated as the SD × 100 divided by the mean average of the repeated measures. P values < .05 were considered to indicate statistical significance. All analyses were performed with SPSS for Windows version 20.0 (SPSS, Inc, Chicago, IL).
The study population of 231 was divided into four groups: asymptomatic moderate aortic stenosis (aortic valve area, 1.0–1.5 cm 2 ), asymptomatic severe aortic stenosis (aortic valve area < 1.0 cm 2 ), and symptomatic severe aortic stenosis with preserved or reduced (<50%) LVEF. Clinical characteristics of the four groups are displayed in Table 1 .
|Moderate ( n = 38)||Severe ( n = 66)||Severe, LVEF ≥ 50% † ( n = 68)||Severe, LVEF < 50% ‡ ( n = 59)|
|Age (y)||71 (64–76)||76 (67–80)||69 (65–77)||74 (67–79)||.08|
|Men||58% (22)||74% (49)||53% (36)||53% (31)||.038|
|Body mass index (kg/m 2 )||27 ± 3.6||26 ± 4.2||26 ± 5.6||27 ± 4.8||.66|
|Heart rate (beats/min)||58 ± 8 §||60 ± 10 §||62 ± 11 ||||67 ± 11||<.001|
|Hypertension||68% (26)||68% (45)||65% (44)||59% (35)||.72|
|Hypercholesterolemia||58% (22)||59% (39)||65% (44)||56% (33)||.82|
|Current smokers||26% (10)||12% (8)||13% (9)||19% (11)||.23|
|Diabetes||8% (3)||15% (10)||13% (9)||12% (7)||.75|
|Previous MI, PCI, or CABG||13% (5)||11% (7)||13% (9)||14% (8)||.96|
|Previous stroke||13% (5)||15% (10)||19% (13)||7% (4)||.25|
|Peripheral artery disease||8% (3)||6% (4)||6% (4)||15% (9)||.21|
|Chronic lung disease||11% (4)||5% (3)||6% (4)||19% (11)||.11|
Echocardiographic analyses revealed a pattern of increasing severity across the four groups ( Table 2 ). Compared with the asymptomatic group, the group with symptoms and reduced LVEFs showed significant structural (higher mass and altered geometry), functional (decreased stroke volume and reduced tricuspid annular systolic plane excursion), and diastolic (left atrial volume index and E/e′ ratio) differences. In addition, more patients had low flow and low gradient and low flow and high gradient flow patterns in the group with symptoms and reduced LVEFs ( Table 3 ).
|P < .001||Asymptomatic||Symptomatic|
|Severe ( n = 66)||Severe, LVEF ≥ 50% ( n = 68)||Severe, LVEF < 50% ( n = 59)|
|Normal flow/low gradient||42.4% (28)||27.9% (19)||16.9% (10)|
|Normal flow/high gradient||39.4% (26)||55.9% (38)||35.6% (21)|
|Low flow/low gradient||0% (0)||2.9% (2)||18.6% (11)|
|Low flow/high gradient||18.2% (12)||13.2% (9)||28.8% (17)|
The group with symptoms but preserved LVEFs did not display any differences in left or right ventricular function or structure but had slightly larger volumes of the left atrium.
Global Measures of Longitudinal Function: Velocity, Displacement, and Deformation
Global measures of systolic (s′) and diastolic velocities (e′ and a′), longitudinal displacement, and deformation (GLS and global longitudinal strain rate) all decreased significantly in relation to symptomatic status, with the highest values in patients with moderate asymptomatic aortic stenosis and the lowest in symptomatic patients with reduced LVEFs ( Table 4 ). However the differences in s′ and e′, longitudinal displacement, and global longitudinal strain rate were already evident within the asymptomatic group between moderate and severe aortic stenosis.
|Moderate ( n = 38)||Severe ( n = 66)||Severe, LVEF ≥ 50% † ( n = 68)||Severe, LVEF < 50% ‡ ( n = 59)|
|Velocity (color DTI)|
|s′ (cm/sec)||5.1 ± 0.8 §||4.5 ± 0.6||4.2 ± 0.9||3.6 ± 0.8 §||<.001|
|e′ (cm/sec)||5.5 ± 1.2 §||4.6 ± 1.3||3.9 ± 1.2 §||3.5 ± 1.0 §||<.001|
|a′ (cm/sec)||6.6 ± 1.3||6.7 ± 1.3||5.6 ± 1.8 §||5.4 ± 1.9 §||<.001|
|Displacement (color DTI)|
|Longitudinal displacement (mm)||11.2 ± 1.5 ||||10.1 ± 1.6||8.8 ± 1.9 §||7.6 ± 2.2 §||<.001|
|Global deformation (2D STE)|
|Global longitudinal strain (%)||16.5 ± 2.8||15.1 ± 2.6||13.9 ± 3.9 ¶||11.5 ± 2.6 § ¶||<.001|
|Global longitudinal strain rate (1/sec)||0.99 ± 0.2 ||||0.88 ± 0.2||0.88 ± 0.2 ¶||0.76 ± 0.2 || ¶||<.001|
|Circumferential strain (%)||14.3 ± 3.8||15.1 ± 4.4||14.5 ± 5.6||11.7 ± 5.2 ||||.06|
|Circumferential strain rate (1/sec)||1.62 ± 0.4||1.51 ± 0.4||1.39 ± 0.3||1.21 ± 0.4 ¶||.036|
|Radial strain (%)||30 ± 12||27 ± 9||25 ± 13||20 ± 12 #||.07|
|Radial strain rate (1/sec)||1.89 ± 0.4||2.01 ± 0.5||1.90 ± 0.7||1.52 ± 0.5 §||.012|
|Regional deformation (2D STE)|
|Basal longitudinal strain (%)||16.0 ± 3.6 ||||14.0 ± 2.7||11.4 ± 4.5 § ¶||10.6 ± 3.6 § ¶||<.001|
|Basal longitudinal strain rate (1/sec)||1.08 ± 0.2 ||||0.96 ± 0.2||0.88 ± 0.2 ¶||0.79 ± 0.2 § ¶||<.001|
|Midlongitudinal strain (%)||16.3 ± 2.9||15.2 ± 2.8||13.9 ± 3.6 ¶||11.8 ± 3.4 § ¶||<.001|
|Midlongitudinal strain rate (1/sec)||0.88 ± 0.2||0.80 ± 0.1||0.81 ± 0.2||0.67 ± 0.2 § ¶||<.001|
|Apical longitudinal strain (%)||16.9 ± 3.7||15.9 ± 4.6||16.7 ± 6.3||12.6 ± 4.7 § ¶||.001|
|Apical longitudinal strain rate (1/sec)||0.98 ± 0.2||0.87 ± 0.2||1.01 ± 0.3||0.82 ± 0.3 ∗∗||.010|
Among the regional measures of deformation, BLS and basal longitudinal strain rate were significantly lower in asymptomatic patients with severe compared with moderate asymptomatic aortic stenosis ( Table 4 ) and also compared with the symptomatic group with reduced LVEFs. We found reduced deformation of the basal segments only in symptomatic patients with preserved LVEFs compared with asymptomatic patients with severe aortic stenosis ( Table 4 ). Conversely, regional deformation was reduced from base to apex in patients with severe symptomatic aortic stenosis with reduced LVEFs compared with asymptomatic patients with severe aortic stenosis ( Figure 1 , Table 4 ).