Aortic stenosis (AS) leads to remodeling of the left heart. Strain measurements enable the assessment of left atrial (LA) mechanics. The goal of this study was to evaluate the short-term effects of transcatheter aortic valve implantation (TAVI) on LA myocardial deformation as well as left ventricular (LV) diastolic function.
Thirty-two patients with severe AS were prospectively enrolled and examined before and 8.2 ± 3.3 days after TAVI. Speckle-tracking echocardiography of the basal septal and lateral segments of the left atrium was performed to determine peak positive strain (R LA ), strain during early diastole (E LA ), and, if feasible, strain during atrial contraction (A LA ). Assessment of LV diastolic function included standard indices, the atrial fraction, and LA volumes.
Compared with baseline, the mean atrial reservoir (R LA ) (24.0 ± 11.2% vs 32.2 ± 14.0%, P < .001) and conduit function (R LA − E LA ) (13.9 ± 5.5% vs 20.8 ± 8.1%, P < .001) improved significantly after TAVI. There was a significant reduction in deceleration time (242 ± 56 vs 195 ± 65 msec, P < .001) and an improvement of pulsed-wave tissue Doppler–derived E′ (5.5 ± 1.8 vs 7.3 ± 2.3 cm/sec, P = .01). Regarding LA volumes, only the minimal LA volume index changed significantly. In contrast, there was no improvement in atrial contraction, that is, contractile function (E LA − A LA ) and atrial fraction. Moreover, the E/E′ ratio remained unchanged.
8.2 ± 3.3 days after TAVI, only the reservoir and conduit function of the left atrium improved, whereas LA contraction and LA volumes, except for the systolic volume index, remained unchanged. This was accompanied by improvement of early LV diastolic function, indicating acute recovery of LV relaxation and LA function.
Aortic stenosis (AS) is the most common valvular disease requiring surgical treatment in industrialized countries. It leads to chronically increased afterload, and the compensatory concentric left ventricular (LV) hypertrophy results in well-studied abnormal LV compliance, LV diastolic dysfunction, and increased LV end-diastolic pressure. These changes are pronounced in elderly patients. Modern echocardiographic methods such as strain imaging allow regional analysis of left atrial (LA) mechanics. In general, LA function comprises three separate components. First, during ventricular systole and the isovolumic relaxation time, the left atrium acts as a reservoir for the blood inflow from the pulmonary veins. Second, in early and mid LV diastole, blood flows passively through the left atrium into the left ventricle, similar to flow through a conduit. Finally, in late diastole, LA contraction completes LV filling. These three distinct phases can be analyzed separately using strain analysis. There is some evidence that LA function is impaired in severe AS, and improvement of reservoir function after surgical valve replacement has been described only in low-risk patients. Transcatheter aortic valve implantation (TAVI) is a rapidly evolving therapy for severe AS in high-risk patients that leads to an immediate reduction in LV afterload, without the confounding effects of major surgery. Until now, it has been unclear whether recovery of LA mechanics takes place in high-risk patients undergoing TAVI. The objective of our study was to describe the effects of TAVI on myocardial deformation of the left atrium using speckle-tracking echocardiographic strain analysis and LV diastolic function after TAVI.
We prospectively enrolled 46 consecutive patients with symptomatic severe AS with baseline transthoracic echocardiographic studies who underwent transfemoral TAVI at our center between November 2010 and July 2011. Indications for TAVI were the criteria described previously. The operative risks for all patients were calculated using the logistic European System for Cardiac Operative Risk Evaluation score. Written informed consent was obtained from each patient. The ethics committee of Charité Universitätsmedizin (Berlin, Germany) approved the study. Patients with significant coronary artery disease were completely revascularized by percutaneous coronary intervention before TAVI. We performed a pulmonary artery catheterization via the femoral vein in 22 patients (68.8%) before TAVI using standard techniques. Three patients were excluded because of mitral stenosis, and five patients were excluded because of reduced acoustic windows. Six patients were discharged before follow-up examination for logistic reasons. Taken together, echocardiographic data from 32 patients were analyzed. To exclude procedure-related confounders, the short-term effects of TAVI were analyzed by echocardiography as soon as the patients were considered clinically stable (i.e., discharged from the intensive care unit, fully mobilized, normalized blood pressures, and freedom from TAVI-associated arrhythmias). The median time from TAVI to the index echocardiographic examination was 8 ± 3.3 days, and 27 patients (84.4%) were examined within 1 week.
Echocardiographic and Doppler Measurements
Standard two-dimensional pulsed-wave Doppler and pulsed-wave Doppler tissue imaging (DTI) echocardiographic parameters were obtained from parasternal and apical acoustic windows according to the guidelines of the American Society of Echocardiography (ASE) using a Vivid 7 Dimension (GE Vingmed Ultrasound AS, Horton, Norway) with an M4S 1.5-MHz to 4.0-MHz transducer. Patients were imaged in the left lateral decubitus position. The LV ejection fraction (LVEF) and LA volumes were obtained on the basis of the recommendations of the ASE. LA volume indices (expansion index, passive emptying fraction, and active emptying fraction) were derived as described previously. LV mass was calculated using the Devereux formula and was indexed to body surface area calculated using the Mosteller formula. Continuous-wave Doppler examinations were analyzed for peak instantaneous velocity. Mean and maximum aortic valve pressure gradients were estimated for all patients by the modified Bernoulli equation using the flow velocity-time integral over the ejection period in continuous-wave Doppler recordings with a 100 mm/sec time scale on the x axis. Aortic valve area was determined using the continuity equation, following the recommendations of the ASE. Aortic and mitral regurgitation were assessed according to the recommendations of the European Association of Echocardiography.
Left ventricular diastolic function was assessed using pulsed-wave Doppler and pulsed-wave DTI recordings on the basis of the recommendations of the ASE. Transmitral flow was acquired to obtain peak early (E) and atrial (A) flow velocities. Atrial fraction was calculated on the basis of the ratio between the velocity-time integral of the A wave and the velocity-time integral of the diastolic transmitral flow. We used the average peak early diastolic (E′) velocity obtained from the septal and lateral sides of the mitral annulus in the four-chamber view with proper DTI settings. Systolic (S′) and late diastolic velocity (A′) as well as the isovolumic relaxation time were quantified using pulsed-wave DTI at the septal insertion sites of the mitral leaflets in the apical four-chamber view. The E/E′ ratio was calculated to estimate LV filling pressures. Tricuspid annular plane systolic excursion was assessed according to the recommendations of the ASE.
Two-Dimensional Speckle-Tracking Strain Analysis
Standard optimized two-dimensional ultrasound images from the apical four-chamber view with frame rates of 60 to 80 frames/sec were recorded and stored digitally for offline analysis (EchoPAC PC; GE Vingmed Ultrasound AS). Longitudinal strain of the LA septal basal and lateral segments was analyzed using two-dimensional speckle-tracking echocardiography. The timing of end ventricular systole was defined by aortic valve closure using the aortic valve click that can be seen in Doppler flow recordings in the LV outflow tract. After manual tracing of the endocardial borders of the atrial lateral wall, superior wall, and atrial septum, the software automatically traced the region of interest. To optimize tracking, the region of interest width was adjusted when necessary. Offline analysis determined peak positive strain (R LA ), strain during early diastole (E LA ) and, when feasible, strain during atrial contraction (A LA ). This allowed the calculation of LA conduit function (R LA − E LA ) and LA contractile function (E LA − A LA ) ( Figures 1 and 2 ).
Interobserver and Intraobserver Variability Analysis
Two echocardiographers, blinded to previously obtained data, separately measured longitudinal strain (R LA and LA conduit and contractile function) and data from 13 random patients for interobserver variability analysis. An experienced observer calculated strain values twice on 2 consecutive days for analysis of intraobserver variability. To determine interobserver and intraobserver variability, intraclass correlation coefficients were used.
All results are expressed as mean ± SD. Statistics were calculated using SPSS version 19.0 (IBM Corporation, Armonk, NY). The Mann-Whitney nonparametric test was used to compare echocardiographic data with baseline and follow-up values. The Kruskal-Wallis nonparametric H test was used to compare gender and regional differences. P values < .05 were considered statistically significant.
Of the 32 consecutive study patients, 14 (43.8%) were men. The mean age was 76.6 ± 9.6 years. The mean logistic European System for Cardiac Operative Risk Evaluation score was 19.2 ± 15.8%. At the baseline examination and after TAVI, 25 (78.1%) and 27 (84.4%) patients were in sinus rhythm, and seven (21.9%) and five (15.6%) patients had atrial fibrillation, respectively. Baseline data are given in detail in Table 1 and Appendix 1 . Invasive measurements are shown in Appendix 2 .
|Age (y)||76.6 ± 9.6|
|Aortic valve prosthesis|
|CoreValve 26 ∗||8 (25%)|
|CoreValve 29 ∗||15 (46.9%)|
|CoreValve 31 ∗||1 (3.1%)|
|Edwards SAPIEN 23 †||2 (6.3%)|
|Edwards SAPIEN 26 †||6 (18.7%)|
|Logistic EuroSCORE (%)||19.2 ± 15.8|
|BMI (kg/m 2 )||25.7 ± 5.1|
|BSA (m 2 )||1.84 ± 0.22|
|Time from baseline echocardiography to implantation (days)||46.0 ± 44.2|
|Time to echocardiography after implantation (days)||8.2 ± 3.3|
All standard echocardiographic findings are presented in Table 2 . There were no statistically significant differences between the baseline examination and the examination after TAVI regarding rhythm, heart rate, and mitral regurgitation. After valve implantation, peak transaortic velocity, mean transaortic systolic gradient, and effective orifice area improved significantly. No severe prosthesis-patient mismatch (aortic valve area < 0.65 cm 2 /m 2 ) could be detected. The LVEF and the LV mass index remained unchanged. In contrast, S′, representing LV longitudinal function, and tricuspid annular plane systolic excursion improved significantly.
|HR (beats/min)||70.4 ± 11.9||73.9 ± 8.4||NS|
|Sinus rhythm||25 (78.1%)||27 (84.4%)||NS|
|AF||7 (21.9%)||5 (15.6%)||NS|
|LVEF (%)||53.3 ± 15.0||54.5 ± 14.3||NS|
|LVM index (g/m 2 )||133.9 ± 32.4||137.7 ± 26.4||NS|
|TAPSE (mm)||19.5 ± 5.3||21.8 ± 5.7||.017|
|None||10 (31.3%)||7 (21.9%)||NS|
|Mild||14 (43.8%)||15 (46.9%)||NS|
|Moderate||7 (21.9%)||9 (28.1%)||NS|
|Severe||1 (3.1%)||1 (3.1%)||NS|
|Peak instantaneous velocity (m/s)||4.1 ± 0.8||2.1 ± 0.4||<.001|
|Mean systolic gradient (mm Hg)||41.7 ± 16.0||9.8 ± 4.2||<.001|
|EOA (cm 2 )||0.73 ± 0.19||1.9 ± 0.48||<.001|
The parameters reflecting early diastolic LV filling improved significantly: E and mean E′ increased ( Figure 3 A), and deceleration time decreased ( Figure 3 B) significantly. However, isovolumic relaxation time and the E/E′ ratio did not change significantly ( Figure 3 C). Parameters of atrial contraction—A′ ( Figure 3 D), the A-wave velocity-time integral, and atrial fraction ( Table 3 )—remained unchanged. Significantly fewer patients had no diastolic dysfunction after TAVI (0% vs 25%, P = .009), and in 15 patients (46.9%), diastolic dysfunction improved by at least one grade ( Figure 4 ). All standard diastolic parameters are presented in Table 3 and Appendix 3 .
|E (m/s)||0.98 ± 0.38||1.2 ± 0.39||.002|
|A (m/s)||0.86 ± 0.36||0.87 ± 0.4||NS|
|E/A ratio||1.36 ± 1.05||1.95 ± 2.3||NS|
|DT (msec)||241.9 ± 55.7||194.6 ± 64.9||<.0001|
|Mean E′ (cm/sec)||5.5 ± 1.8||7.3 ± 2.3||.001|
|Septal A′ (cm/sec)||6.9 ± 3.0||6.3 ± 2.6||NS|
|Lateral A (cm/sec)||6.1 ± 3.8||5.9 ± 3.2||NS|
|S′ (cm/sec)||4.5 ± 1.2||5.5 ± 1.6||.001|
|IVRT (msec)||108.8 ± 39.6||98.1 ± 37.0||NS|
|E/E′ ratio||18.7 ± 8.0||17.6 ± 7.3||NS|
|Atrial fraction||0.38 ± 0.14||0.34 ± 0.13||NS|
|Grade of diastolic dysfunction|
|None||0 (0%)||8 (25%)||.009|
|I||15 (48.4%)||10 (31.3%)||NS|
|II||13 (41.9%)||11 (34.3%)||NS|
|III||3 (9.7%)||3 (9.4%)||NS|
|Diastolic LA volume (mL)||78.1 ± 30.2||73.9 ± 30.6||NS|
|Diastolic LA volume index (mL/m 2 )||41.3 ± 14.5||39.6 ± 15.0||NS|
|Systolic LA volume (mL)||48.4 ± 25.1||40.9 ± 21.7||NS|
|Systolic LA volume index (mL/m 2 )||25.8 ± 13.0||21.8 ± 10.8||.049|
|Pre–A wave LA volume (mL)||53.2 ± 18.2||54.7 ± 22.1||NS|
|Pre–A wave LA volume index (mL/m 2 )||28.6 ± 8.2||29.1 ± 10.6||NS|
Reservoir function (R LA ) and conduit function (R LA − E LA ) obtained at the septal basal and lateral basal segments improved significantly after percutaneous valve implantation. In contrast, the contractile function of the left atrium (E LA − A LA ) did not change. The values of reservoir function and conduit function were significantly higher in the lateral basal segments compared with the septal basal segments before and after TAVI. Data are shown in detail in Table 4 and Figure 5 .
|Reservoir function (R LA ) (%)||24.0 ± 11.2||32.3 ± 14.0||<.001|
|E LA (%)||10.2 ± 8.4||11.5 ± 10.1||NS|
|Conduit function (R LA − E LA ) (%)||13.8 ± 5.5||20.8 ± 8.1||<.001|
|A LA (%)||−1.0 ± 2.0||−2.4 ± 2.3||.003|
|Contractile function (E LA − A LA ) (%)||13.4 ± 6.9||15.2 ± 8.9||NS|
|Reservoir function (R LA ) (%)||20.2 ± 13.0 ∗||25.7 ± 16.6 ∗||.03|
|E LA (%)||9.6 ± 9.7||11.4 ± 13.1||NS|
|Conduit function (R LA − E LA ) (%)||10.6 ± 6.4 ∗||14.4 ± 8.1 ∗||.029|
|A LA (%)||−1.4 ± 3.0||−3.5 ± 3.9 ∗||.042|
|Contractile function (E LA − A LA ) (%)||13.1 ± 8.9||15.6 ± 12.0||NS|
|Reservoir function (R LA ) (%)||28.2 ± 12.5 †||39.7 ± 16.4 †||.001|
|E LA (%)||10.9 ± 9.3||11.8 ± 10.5||NS|
|Conduit function (R LA − E LA ) (%)||17.3 ± 8.4 †||27.9 ± 12.3 †||<.001|
|A LA (%)||−0.6 ± 2.3||−1.2 ± 2.2 †||NS|
|Contractile function (E LA − A LA ) (%)||13.4 ± 7.8||15.6 ± 9.2||NS|
Heart rate and LVEF at baseline and after valve implantation did not differ significantly between men and women. Furthermore, all parameters regarding diastolic function and LA mechanics were equal at baseline in both groups.
In men, only average reservoir function improved significantly after TAVI. In woman, conduit function, E, and E′ increased and E/E′ decreased significantly, suggesting more pronounced improvements of LV relaxation and LV pressure. Nevertheless, deceleration time decreased significantly only in men (238.8 ± 47.9 vs 171.1 ± 57.4 msec, P = .001). Gender differences are presented in detail in Table 5 and Appendix 4 .
|Logistic EuroSCORE (%)|
|Men||21.5 ± 18.3|
|Women||17.4 ± 13.7|
|LVEF at baseline (%)|
|Men||49.0 ± 18.6|
|Women||56.7 ± 10.92|
|Men||0.91 ± 0.34||1.07 ± 0.27||NS|
|Women||1.04 ± 0.42||1.27 ± 0.45||.009|
|Men||5.96 ± 2.12||6.76 ± 2.60||NS|
|Women||5.21 ± 1.54||7.63 ± 1.95||.02|
|Men||16.6 ± 7.6||17.6 ± 7.5||NS|
|Women||20.46 ± 8.34||17.57 ± 7.32||.02|
|Reservoir function (R LA ) (%)|
|Men||23.2 ± 10.7||29.9 ± 16.3||.035|
|Women||24.7 ± 11.8||34.1 ± 12.1||.03|
|Conduit function (R LA − E LA ) (%)|
|Men||14.5 ± 6.8||17.9 ± 8.4||NS|
|Women||13.4 ± 4.5||23.0 ± 7.3||< .001|
|Contractile function (E LA − A LA ) (%)|
|Men||11.5 ± 5.4||13.8 ± 9.6||NS|
|Women||14.3 ± 8.7||16.5 ± 8.3||NS|