Methods

This is a retrospective analysis of a TAVR database at a single academic institution. A database of consecutive patients who underwent TAVR at the Bluhm Cardiovascular Institute between January 2012 and June 2014 was reviewed. The Institutional Review Board approved analysis of all data. Data were obtained using several sources, including available clinical information and laboratory values in the electronic medical record and echocardiographic data record (Synapse Cardiovascular, Fujifilm USA, Stamford, CT). Outcomes were obtained by periodic monitoring through our valve clinic via telephone call and/or follow-up visit. Baseline clinical characteristics, vital signs, laboratory findings, echocardiography, and cardiac catheterization were routinely collected per institutional protocol. In-hospital laboratory, medication, transfusion, complications, and pre- and post-TAVR echocardiographic data were also obtained.

Analysis of pre- and post-TAVR echocardiograms and clinical data was performed. The pre-TAVR transthoracic echocardiogram chosen for analysis was the most immediate preprocedural echocardiogram available in the medical record. The post-TAVR transthoracic echocardiogram performed at 1-month follow-up was compared with baseline. Two-dimensional and Doppler echocardiographic measurements (mitral inflow Doppler, mitral annular tissue Doppler, left atrial [LA] volume, LV dimensions, and LV volumes) were performed in accordance with the guidelines of the American Society of Echocardiography (ASE). Specific measurements of mitral inflow Doppler, mitral annular tissue Doppler, and LA volume were systematically measured by one investigator (J.L.F.), and LV volumes as measured by the apical biplane method of discs were measured by another investigator (K.B.).

A subset of 20 randomly selected patients was used for interobserver correlation and agreement in measurement of important diastolic parameters by another investigator (J.E.A.B.), and another subset of 10 randomly selected patients were used for intraobserver correlation and agreement with initial measurements. In the apical four-chamber view, the pulsed-wave Doppler sample volume was targeted at the tips of the mitral valve to assess peak early (E) and late (A) diastolic transmitral filling velocities and E-wave deceleration time on pulsed-wave spectral Doppler. In the same view, peak early diastolic mitral annular tissue velocities (e′) were obtained by placing the pulsed-wave tissue Doppler sample volume on both the lateral and septal mitral annulus and recorded using pulsed-wave Doppler. LA volume was measured using the biplane area-length (ellipsoid) method, LV mass was estimated using the truncated ellipse equation, aortic valve area was calculated using the continuity equation, and peak and mean systolic transaortic gradients were calculated using the simplified Bernoulli equation. Parameters reported for patients with atrial fibrillation during either echocardiogram involved the average value of 3–4 nonconsecutive beats with cycle lengths within 20% of the average cycle length where applicable. Patients were excluded from analyses when they had echocardiographic findings that rendered their diastolic analysis impossible or incomplete, including severe mitral regurgitation (MR), any degree of mitral stenosis (MS), mitral valve repair, or mitral valve replacement. Severity of valvular regurgitation was determined by the reading echocardiographer using standard qualitative and quantitative measurements. Mitral annular calcification (MAC) was graded using qualitative measurements: mild, less than 1/3 annular circumference; moderate, between 1/3 and 2/3 annular circumference; and severe, greater than 1/3 annular circumference. Right ventricular dysfunction was determined qualitatively by the reading echocardiographer using standard measures and graded on a scale of 0–3 (none, mild, moderate, or severe).

Diastolic dysfunction grade was determined using two grading systems: (1) the Kuwaki et al. classification scheme, which modified the Redfield et al. scheme with the addition of an intermediate grade, notated as 1a, where grade 0 (0.75 < E/A < 1.5, deceleration time [DT] > 140 ms, E/e′<10), grade 1 (E/A ≤ 0.75, DT > 140 ms, E/e′ < 10), grade 1a (E/A ≤ 0.75, DT > 140 ms, E/e′ ≥ 10), grade 2 (0.75 < E/A < 1.5, DT > 140 ms, E/e′ ≥ 10), and grade 3 (E/A ≥ 1.5, DT ≤ 140 ms, E/e′ ≥ 10) have progressively worsening parameters as grade increases; and (2) the ASE and the European Society of Cardiovascular Imaging (ASE/EACVI) diastolic dysfunction grading system, which incorporates several diastolic parameters to assign a grade 0–3. To calculate E/e′, lateral e′ was used for the Kuwaki grading system, while the average of septal and lateral e′ was used for the ASE/EACVI grading system, as per methodologies of the respective grading systems. A significant change in diastolic dysfunction was defined as an increase or decrease in diastolic dysfunction grade. We determined that the standard deviation between two LA volume index (LAVI) measurements on the same patient by the same investigator in this data set was 2.18 mL/m ² , and we therefore defined a significant change in LAVI as an increase or decrease in LAVI by > 3 mL/m ² . Patients with atrial fibrillation, in whom grading of diastolic dysfunction is not possible, were treated as a separate group and compared with patients whose diastolic function was graded.

Differences in baseline demographics, medical history, physical findings, laboratory findings, and echocardiographic findings were compared between patients with and without grade 2 or greater diastolic dysfunction at baseline. Comparison between groups was assessed using the Kruskal-Wallis test for continuous variables and the Pearson χ ² test for categorical variables. For changes in diastolic dysfunction grade, the paired nonparametric Wilcoxon signed-rank test was used to compare preoperative and postoperative continuous diastolic dysfunction grade data to avoid any incorrect assumptions associated with respect to our data and Gaussian distribution; Fisher’s exact test was used for comparison of categorical variables. P < .05 was considered statistically significant. Inter- and intraobserver correlation and agreement for important diastolic parameters were calculated using Pearson’s linear regression and Bland-Altman analysis, respectively. Kaplan-Meier survival analysis was performed using 1-year mortality data with respect to diastolic dysfunction grade. Specifically, survival analysis was stratified for baseline diastolic dysfunction, follow-up post-TAVR diastolic dysfunction grade, change in diastolic dysfunction grade, and change in LAVI.

Predictors of 1-year mortality and combined 1-year mortality and cardiovascular (CV) hospitalization were determined using a multivariable regression model. The model was determined using a process of backward selection, starting with univariate demographic, medical history, physical and laboratory findings, and echocardiographic parameters that predicted 1-year mortality or mortality/CV hospitalization with a significance level of P < .05. Echocardiographic parameters included both pre-TAVR and dynamic parameters between the pre-TAVR and post-TAVR echocardiograms. The least significant variables were removed until all remaining variables had a significance level of P < .05. All data were analyzed using Stata MP, version 14.0 (StataCorp, College Station, TX).

Results

A total of 120 patients with severe AS who underwent TAVR were identified, all of whom had baseline echocardiograms. There were 30 patients excluded from the analysis for the following reasons: severe MR ( n = 9), MS ( n = 15), mitral valve replacement ( n = 4), and mitral valve repair ( n = 2), leaving 90 patients for the pre-TAVR analysis. Of these patients, 17 did not have a complete 1-month post-TAVR echocardiogram available for review at our institution, either due to death ( n = 6), follow-up outside the institution ( n = 10), or a non-Doppler echocardiogram at our institution ( n = 1). There was one patient not included in the post-TAVR analysis whose MR worsened to severe post-TAVR and was not included in the analysis. Therefore, there were a total of 72 echocardiograms without significant mitral valve disease available for analysis on follow-up. Interobserver correlation and mean bias were all within the acceptable range for E-wave velocity, E-wave deceleration time, A-wave velocity, lateral e′ velocity, and LAVI ( R ² = 0.977, 0.959, 0.984, 0.937, 0.934; mean ± SD bias 0.11 ± 2.95 cm/sec, −5.47 ± 8.03 msec, −0.33 ± 1.67 cm/sec, 0.005 ± 0.377 cm/sec, −1.02 ± 3.07 mL/m ² , respectively). Intraobserver correlation and mean bias were similarly acceptable ( R ² = 0.982, 0.977, 0.979, 0.978, 0.958; mean ± SD bias −1.40 ± 2.24 cm/sec, −2.47 ± 8.16 msec, −1.66 ± 2.26 cm/sec, 0.17 ± 0.26 cm/sec, −0.79 ± 2.18 mL/m ² , respectively).

Baseline characteristics for patients are shown in Table 1 . Patients were predominantly octogenarians with normal LV ejection fraction (LVEF) who underwent transfemoral TAVR with balloon-expandable transcatheter prosthetic valve. There were no significant differences in baseline clinical characteristics between those with and without grade 2 or greater diastolic dysfunction. Comparisons of echocardiographic parameters are shown in Table 2 . Pre-TAVR, patients with higher diastolic dysfunction grades had lower LVEF, E-wave deceleration time, A-wave velocity, and e′ velocities and higher LV end-systolic volume, E-wave velocity, E/A ratio, A-wave velocity, LAVI, and pulmonary artery systolic pressure. Baseline aortic valve area, aortic valve gradients, and New York Heart Association (NYHA) functional class were not significantly different between diastolic dysfunction grades pre-TAVR. Lateral e′ velocity and aortic valve area were higher post-TAVR compared with baseline, while E/e′ lateral, and aortic gradients were lower post-TAVR compared with baseline. There were 77 (85.6%) patients with at least mild MAC and no difference between the degrees of MAC across categories of diastolic dysfunction. NYHA functional class was significantly improved post-TAVR compared with pre-TAVR. There were no differences pre-TAVR versus post-TAVR in degree of MR (data not shown).

Table 2

Echocardiographic variables and symptom class, pre- and post-TAVR

Variable	Pre-TAVR						Post-TAVR						P overall
	Grade 0–1a ( n = 22)	Grade 2 ( n = 31)	Grade 3 ( n = 15)	AF ( n = 16)	Uncategorized (n = 6)	P value	Grade 0–1a ( n = 15)	Grade 2 ( n = 30)	Grade 3 ( n = 4)	AF ( n = 16)	Uncategorized ( n = 7)	P value
LVEF (%), all patients	60.5 ± 12.9	59.4 ± 10.4	46.5 ± 17.5	57.2 ± 11.3	43.3 ± 10.5	.007	56.6 ± 13.0	59.0 ± 13.5	48.0 ± 16.8	55.3 ± 14.8	56.6 ± 8.4	.61	.86
LV end-diastolic volume (mL)	79.5 ± 43.0	88.5 ± 37.6	100.3 ± 46.3	83.6 ± 38.6	83.5 ± 23.4	.55	84.4 ± 32.9	92.8 ± 39.4	84.7 ± 10.3	86.8 ± 32.6	81.4 ± 34.9	.84	.77
LV end-systolic volume (mL)	34.2 ± 28.9	37.9 ± 23.4	56.5 ± 37.0	38.1 ± 25.3	47.6 ± 18.0	.068	36.8 ± 19.4	40.5 ± 27.3	43.5 ± 13.1	40.8 ± 21.1	35.9 ± 17.5	.92	.91
Interventricular septal thickness (cm)	1.26 ± 0.19	1.26 ± 0.31	1.31 ± 0.19	1.22 ± 0.22	1.32 ± 0.19	.83	1.31 ± 0.25	1.30 ± 0.29	1.24 ± 0.32	1.25 ± 0.21	1.16 ± 0.27	.67	.79
LV posterior wall thickness (cm)	1.21 ± 0.17	1.22 ± 0.28	1.13 ± 0.21	1.27 ± 0.19	1.36 ± 0.18	.65	1.17 ± 0.19	1.21 ± 0.22	1.15 ± 0.20	1.24 ± 0.15	1.10 ± 0.18	.51	.46
LV mass index (g/m 2 )	109 ± 35	114 ± 35	116 ± 34	117 ± 23	157 ± 28	.23	108 ± 36	114 ± 42	119 ± 11	121 ± 32	95 ± 26	.62	.64
MAC						.62						.22	>.99
None	3 (13.6%)	7 (22.6%)	1 (6.7%)	2 (12.5%)	0		3 (20%)	3 (10%)	0	3 (18.8%)	0
Mild	12 (54.5%)	8 (25.8%)	5 (33.3%)	5 (31.3%)	3 (50.0%)		8 (53.3%)	10 (33.3%)	1 (25.0%)	5 (31.3%)	3 (42.9%)
Moderate	6 (27.3%)	12 (38.7%)	9 (60.0%)	7 (43.8%)	2 (33.3%)		3 (20.0%)	12 (40.0%)	2 (50.0%)	8 (50.0%)	4 (57.1%)
Severe	1 (4.5%)	4 (12.9%)	0	2 (12.5%)	1 (16.7%)		1 (6.7%)	5 (16.7%)	1 (25.0%)	0	0
E-wave (cm/sec)	67 ± 12	110 ± 32	123 ± 22	115 ± 23	128 ± 17	<.001	76 ± 17	117 ± 24	133 ± 24	125 ± 30	122 ± 27	<.001	.114
E-wave deceleration time (msec)	269 ± 81	241 ± 60	119 ± 14	168 ± 23	170 ± 38	<.001	282 ± 71	236 ± 56	131 ± 10	177 ± 40	161 ± 42	<.001	.52
A-wave (cm/sec)	109 ± 22	109 ± 29	55 ± 15		65 ± 32	<.001	104 ± 30	113 ± 28	52 ± 22		79 ± 46	.0087	.20
E/A ratio	0.64 ± 0.13	1.03 ± 0.22	2.45 ± 0.97		2.68 ± 1.15	<.001	0.78 ± 0.24	1.07 ± 0.21	2.90 ± 1.14		2.15 ± 1.43	<.001	.57
Lateral e′ (cm/sec)	5.49 ± 2.09	5.93 ± 1.80	6.29 ± 2.59	8.24 ± 2.57	5.55 ± 2.53	.019	7.87 ± 2.02	6.77 ± 1.72	8.08 ± 2.06	10.22 ± 3.71 ( n = 13)	10.10 ± 3.81	.083	<.001
Septal e′ (cm/sec)	4.71 ± 1.85	4.52 ± 1.21	4.96 ± 2.16	5.99 ± 1.69	5.02 ± 2.11	.047	4.64 ± 1.21	5.14 ± 1.51	5.68 ± 1.09	6.81 ± 2.19 ( n = 14)	5.39 ± 1.53	.019	.093
E/e′ lateral	13.9 ± 5.6	19.9 ± 7.3	23.3 ± 12.2	15.9 ± 8.2	28.7 ± 15.8	.0026	10.6 ± 4.9	18.7 ± 8.0	17.6 ± 6.5	13.4 ± 6.2 ( n = 13)	12.4 ± 6.6	<.001	.018
E/e′ septal	15.7 ± 4.9	26.6 ± 12.4	28.9 ± 12.4	20.7 ± 7.8	29.4 ± 12.7	<.001	17.2 ± 4.9	24.5 ± 7.8	24.4 ± 7.5	20.0 ± 10.9 ( n = 14)	22.5 ± 13.1	.026	.38
LAVI (mL/m 2 )	38.2 ± 9.6 ( n = 21)	46.8 ± 18.9	46.9 ± 13.0	59.6 ± 20.0 ( n = 15)	57.8 ± 18.5	.0032	35.1 ± 10.9	41.2 ± 15.1	52.7 ± 13.2	49.6 ± 18.9	38.2 ± 18.6	.10	.055
PA systolic pressure (mmHg)	30.0 ± 12.5 ( n = 8)	38.8 ± 12.1 ( n = 25)	51.2 ± 15.1 ( n = 13)	47.2 ± 13.8 ( n = 11)	47.7 ± 19.6 ( n = 6)	.0098	32.5 ± 4.9 ( n = 8)	38.7 ± 14.9 ( n = 22)	43.0 ± 22.3 ( n = 4)	43.3 ± 9.6 ( n = 14)	37.0 ± 15.3 ( n = 5)	.28	.21
Aortic valve area (cm 2 )	0.69 ± 0.19	0.75 ± 0.21	0.64 ± 0.22	0.67 ± 0.19	0.68 ± 0.30	.42	1.86 ± 0.84	1.70 ± 0.70	1.74 ± 0.42	1.89 ± 0.74	2.01 ± 0.86	.88	<.001
Aortic peak gradient (mmHg)	75.8 ± 23.7	77.1 ± 22.1	68.0 ± 17.2	67.3 ± 17.5	72.9 ± 12.0	.53	22.5 ± 9.5	22.9 ± 10.5	22.1 ± 11.3	17.9 ± 9.3	17.7 ± 16.3	.16	<.001
Aortic mean gradient (mmHg)	44.7 ± 14.9	45.8 ± 13.8	40.6 ± 11.6	38.8 ± 11.0	41.7 ± 5.9	.59	11.5 ± 4.6	12.2 ± 5.8	12.1 ± 7.0	9.1 ± 4.8	9.3 ± 8.9	.099	<.001
More than mild paravalvular leak							0	1	1	2	1
NYHA functional class						.26						.080	<.001
I	0	0	0	0	0		1 (7.1%)	4 (16.7%)	0	1 (8.3%)	0
II	3 (14.3%)	3 (10.3%)	0	1 (7.1%)	1 (16.7%)		10 (71.4%)	10 (41.7%)	3 (75.0%)	2 (16.7%)	2 (40.0%)
III	14 (66.7%)	21 (72.4%)	8 (57.1%)	9 (64.3%)	4 (66.7%)		3 (21.4%)	9 (37.5%)	1 (25.0%)	6 (50.0%)	2 (40.0%)
IV	4 (19.0%)	5 (23.8%)	6 (42.9%)	4 (28.6%)	1 (16.7%)		0	1 (4.2%)	0	3 (25.0%)	1 (20.0%)

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Differential Impact of Net Atrioventricular Compliance on Clinical Outcomes in Patients with Mitral Stenosis According to Cardiac Rhythm
2. Pulmonary Flow as an Improved Method for Determining Cardiac Output in Mice after Myocardial Infarction
3. Author Index
4. Subject Index

Stay updated, free articles. Join our Telegram channel

Join