Background
Inoperable aortic stenosis may be treated with either transcatheter aortic valve replacement (TAVR) or medical management (MM) with or without balloon aortic valvuloplasty (BAV). The aim of this study was to compare the long-term echocardiographic findings among TAVR, MM, and BAV in patients with severe, inoperable aortic stenosis.
Methods
A total of 358 inoperable patients in the Placement of Aortic Transcatheter Valves trial were randomized to MM or TAVR. Echocardiograms obtained at baseline, 30 days, and 1, 2, and 3 years were analyzed by a central core laboratory.
Results
At baseline, TAVR and MM were similar, with more frequent Society of Thoracic Surgeons score > 10 (51.7% vs 65.0%, P = .03) and larger end-systolic volumes (54.5 ± 29.3 vs 69.1 ± 48.0 mL, P = .03) in MM. By 30 days after TAVR, mean aortic valve gradient had decreased (from 43.8 ± 14.7 to 10.0 ± 4.3 mm Hg, P < .001), ejection fraction had increased (from 53.2 ± 12.4% to 56.7 ± 10.0%, P < .001), and left ventricular (LV) mass index had decreased (from 144.7 ± 36.1 to 140.0 ± 37.9 gm/m 2 , P < .05). After 1 year, aortic valve gradients and area were unchanged, while LV mass index had decreased by another 16 gm/m 2 (to 124 gm/m 2 ). By 30 days after BAV, mean aortic valve gradient had decreased from 43.4 ± 15.0 to 31.9 ± 11.1 mm Hg, while ejection fraction and LV mass index were unchanged; gradient reverted to baseline at 1 year. No changes in gradients or mass were seen in MM patients.
Conclusions
TAVR results in immediate and sustained relief in pressure overload and improved LV systolic function, with continued regression of hypertrophy over 3 years. Poor clinical results with BAV are explained by the modest and transient reductions in pressure overload with BAV, which were not accompanied by improved LV function or remodeling. TAVR is the preferred treatment in eligible inoperable patients (ClinicalTrials.gov identifier NCT00530894).
Highlights
- •
In inoperable patients with critical aortic stenosis, TAVR results in early and sustained improvements in aortic hemodynamics, ejection fraction, and hypertrophy.
- •
In contrast, balloon aortic valvuloplasty results in modest and transient improvements hemodynamics, with no accompanying increase in ejection fraction or hypertrophy regression.
- •
These findings provide the mechanistic explanation for the observed worse clinical outcomes in this group.
- •
TAVR is the preferred treatment in eligible patients with inoperable aortic stenosis.
Severe aortic stenosis is progressive and, once symptoms develop, carries a 5-year mortality >80% in both historical and contemporary populations. Multiple comorbidities may contribute to this grim prognosis and preclude surgical valve replacement. Transcatheter aortic valve replacement (TAVR) has become an acceptable alternative for patients with severe, symptomatic aortic stenosis with extreme surgical risk, on the basis of the results of the Placement of Aortic Transcatheter Valves (PARTNER) trial.
The medical management (MM) arm of the PARTNER trial allowed the use of medical therapies as well as balloon aortic valvuloplasty (BAV) in this extreme surgical risk cohort, thereby allowing an in-depth analysis of this procedure in contemporary practice. Although the initial improvements in valve gradients and area after TAVR in cohort B of the PARTNER trial have previously been reported, the current report compares valve function and associated longitudinal long-term compensation and remodeling in response to TAVR with those from contemporary BAV and MM as assessed by a central echocardiographic core laboratory with rigorous quality control. We hypothesized that this randomized comparison would provide mechanistic insights into the differing clinical outcomes associated with these two procedures.
Methods
The complete trial design, including inclusion and exclusion criteria and baseline characteristics, of cohort B of the PARTNER trial has been reported previously. The study randomly allocated 358 patients with severe, symptomatic aortic stenosis deemed to be “inoperable” by at least two cardiac surgeons and an interventional cardiologist to TAVR or MM. After randomization, the MM cohort could be treated with BAV at the discretion of the local heart team. Crossover to TAVR was permitted after 1 year. Patients were required to have both a site-determined aortic valve area of ≤0.8 cm 2 and a resting or dobutamine-induced peak velocity of ≥4 m/sec or a mean gradient of ≥40 mm Hg.
Transthoracic echocardiograms were obtained using a uniform image acquisition protocol at baseline, 30 days, 6 months, and 1 year and annually thereafter. All studies were analyzed by a central core laboratory with quality and measurement methodology previously reported. Ventricular size and function and valvular function were measured according to previously published guidelines. Both qualitative (visual) and quantitative (biplane Simpson’s method of disks) were used to report ejection fraction. Relative wall thickness was calculated as 2 × posterior wall thickness/left ventricular (LV) diameter. Core laboratory aortic annular measurements were performed in diastole. Site-reported systolic annular diameters were also recorded and used in computation of the cover index, defined as [1 − (measured diameter/nominal prosthesis diameter)] × 100%. Prosthesis-patient mismatch was defined as absent if the aortic valve area index was >0.85 cm 2 /m 2 , moderate if ≥0.65 to <0.85 cm 2 /m 2 , and severe if <0.65 cm 2 /m 2 . Native valve regurgitation and paravalvular regurgitation after TAVR were graded using a multiparametric approach in accordance with American Society of Echocardiography recommendations for native and prosthetic valves, except for the cumulative arc length of the jet(s) relative to the annular circumference, which used criteria of <10% for mild, 10% to 30% for moderate, and >30% for severe. Also in TAVR patients, LV outflow tract diameter and velocity measurements were made just proximal to the valve stent frame, to reduce the influence of any possible flow acceleration that could have altered hemodynamic data.
To avoid possible survival or cross-over bias in this high-risk population, longitudinal comparisons were performed using data from all patients with available paired echocardiograms for analysis at baseline and 30 days (171 TAVR and 181 MM patients) or baseline and 1 year (90 TAVR and 47 MM patients). Additional data at 2 and 3 years are shown for the paired 1-year cohort. Primary comparisons are performed between TAVR and MM; additionally, the MM patients were divided into those who underwent at least one BAV procedure before the 30 day echocardigoraphic study and those who did not. Patients undergoing initial BAV after 30 days were excluded from 1-year analyses. In addition, we explored the predictors of achieving a ≥50% reduction in mean aortic valve gradient after BAV compared with patients who did not.
Statistical Analysis
Continuous variables are summarized as mean ± SD or as medians and interquartile ranges, as appropriate, and were compared using the Student t test or Mann-Whitney rank sum test accordingly. Categorical variables were compared by using the χ 2 or Fisher exact test. Survival curves for time-to-event variables were constructed on the basis of all available follow-up data, with the use of Kaplan-Meier estimates, and comparisons were performed using the log-rank test. Data are based on an extract date of February 13, 2012. All statistical analyses were performed in SAS version 9.2 (SAS Institute Inc, Cary, NC).
Results
TAVR and MM Comparability
A total of 358 patients were randomized to receive either TAVR or MM ( Figure 1 ). Within the MM group, 117 patients (65%) underwent BAV within 30 days, while 60 did not. Demographics and selected baseline characteristics of these “as treated” populations are shown in Table 1 for the 264 subjects with paired baseline and 30 day echocardiographic data. Aside from a lower prevalence of Society of Thoracic Surgeons (STS) score > 10 with TAVR compared with MM (51.7% vs 65.0%, P = .03), there were no significant differences between the TAVR and MM groups. Within the MM group, the no-BAV cohort had lower STS scores and European System for Cardiac Operative Risk Evaluation scores and were more likely to have undergone prior BAV (40% vs 15%).
| Variable | TAVR ( n = 143) | MM ( n = 121) | P value, TAVR vs MM | MM no BAV ( n = 36) | MM BAV ( n = 85) | P value, no BAV vs BAV |
|---|---|---|---|---|---|---|
| Demographics and risk score | ||||||
| Age (y) | 83.19 ± 8.92 | 82.81 ± 9.03 | .73 | 82.09 ± 10.19 | 83.11 ± 8.54 | .57 |
| Men | 44.8% (64/143) | 47.9% (58/121) | .61 | 50.0% (18/36) | 47.1% (40/85) | .77 |
| Caucasian race | 93.7% (134/143) | 91.7% (111/121) | .54 | 86.1% (31/36) | 94.1% (80/85) | .14 |
| BMI (kg/m 2 ) | 27.02 ± 7.96 | 26.79 ± 6.08 | .79 | 26.71 ± 6.70 | 26.82 ± 5.84 | .94 |
| STS score > 10 | 51.7% (74/143) | 65.0% (78/120) | .03 | 51.4% (18/35) | 70.6% (60/85) | .05 |
| Logistic EuroSCORE | 27.65 ± 17.50 | 29.44 ± 18.61 | .42 | 23.05 ± 15.85 | 32.15 ± 19.11 | .01 |
| C risk factors | ||||||
| Diabetes | 31.5% (45/143) | 33.1% (40/121) | .78 | 44.4% (16/36) | 28.2% (24/85) | .08 |
| Hyperlipidemia | 73.4% (105/143) | 81.8% (99/121) | .10 | 80.6% (29/36) | 82.4% (70/85) | .81 |
| Smoking | 43.4% (62/143) | 49.6% (60/121) | .31 | 52.8% (19/36) | 48.2% (41/85) | .65 |
| Hypertension | 86.6% (123/142) | 86.8% (105/121) | .97 | 88.9% (32/36) | 85.9% (73/85) | .66 |
| Cardiovascular conditions | ||||||
| CHF | 98.6% (141/143) | 96.7% (117/121) | .42 | 97.2% (35/36) | 96.5% (82/85) | .83 |
| CAD | 69.9% (100/143) | 73.6% (89/121) | .52 | 72.2% (26/36) | 74.1% (63/85) | .83 |
| Cerebrovascular disease | 29.1% (41/141) | 28.7% (33/115) | .95 | 31.4% (11/35) | 27.5% (22/80) | .67 |
| Peripheral vascular disease | 28.2% (40/142) | 25.6% (31/121) | .64 | 27.8% (10/36) | 24.7% (21/85) | .72 |
| Prior BAV | 14.0% (20/143) | 22.3% (27/121) | .08 | 38.9% (14/36) | 15.3% (13/85) | .004 |
| 1 | 9.8% (14/143) | 20.7% (25/121) | .01 | 36.1% (13/36) | 14.1% (12/85) | .006 |
| 2 | 2.8% (4/143) | 0.8% (1/121) | .38 | 0.0% (0/36) | 1.2% (1/85) | .51 |
| ≥3 | 1.4% (2/143) | 0.8% (1/121) | 1.00 | 2.8% (1/36) | 0.0% (0/85) | .12 |
| Noncardiac conditions | ||||||
| Renal disease (CR ≥ 2 mg/dL) | 19.6% (28/143) | 19.2% (23/120) | .93 | 19.4% (7/36) | 19.0% (16/84) | .96 |
| COPD | 39.9% (57/143) | 50.4% (61/121) | .09 | 50.0% (18/36) | 50.6% (43/85) | .95 |
| Chest wall radiation | 10.5% (15/143) | 8.3% (10/121) | .54 | 11.1% (4/36) | 7.1% (6/85) | .46 |
At baseline, patients receiving TAVR compared with MM showed similar LV wall thickness, mass, and concentric hypertrophy ( Supplemental Table 1 ; available on www.onlinejase.com ). LV end-systolic volumes (69.1 ± 48.0 vs 54.5 ± 29.3 mL, P = .03) were higher, and ejection fraction tended to be slightly lower in the MM group, at 50 ± 14% versus 53 ± 12% ( P = .07). Within the MM group, there were no differences between BAV and no-BAV subjects. Similarly, there were no differences at baseline in aortic valve hemodynamics or regurgitation or mitral regurgitation between TAVR and MM. Within the MM group, the no-BAV group tended to have a lower mean gradient (38 vs 43 mm Hg, P = .06) and higher valve area (0.70 vs 0.63 cm 2 , P = .09) than the BAV group, although these differences did not reach significance.
Paired Baseline and 30-Day Data
Over the first 30 days, there were 16 deaths (11 TAVR, 5 MM) and four patients who converted from MM to surgical valve replacement ( Figure 1 ). Paired baseline and 30-day echocardiographic data were available in 264 patients (143 TAVR, 121 MM) for analysis of changes in aortic valve hemodynamics, regurgitation, and LV structure and function ( Supplemental Table 1 ).
Thirty days after TAVR, peak and mean aortic valve gradients fell markedly: from 73.3 ± 24.1 to 19.8 ± 8.4 mm Hg and from 43.8 ± 14.7 to 10.0 ± 4.3 mm Hg, respectively ( P < .001 for both), while aortic valve area increased from 0.64 ± 0.18 to 1.56 ± 0.43 cm 2 and aortic valve area index rose from 0.36 ± 0.11 to 0.89 ± 0.26 cm 2 /m 2 ( Figure 2 ). In the TAVR group, the mean cover index was 12.93 ± 5.84, and the prevalence of moderate or severe prosthesis-patient mismatch was 48.4% at 30 days. The prevalence of moderate or severe total aortic regurgitation (AR) was 21.2% at baseline and 13.5% at 30 days ( Figure 3 ) Moderate or severe AR was not predicted by any baseline echocardiographic parameter measured or by cover index. There were no echocardiographic predictors of moderate to severe AR, although such patients were more likely to have histories of chest wall irradiation ( P = .006) and less likely to have diabetes ( P = .007). The prevalence of mitral regurgitation was unchanged from baseline to 30 days ( P = .125). TAVR patients showed significant LV remodeling, with reduced LV mass and LV mass index (from 257.0 ± 76.4 to 251.3 ± 85.8 g and from 144.7 ± 36.1 to 140.0 ± 37.9 g/m 2 , P < .05 for both; Supplemental Table 1 ), while changes in wall thicknesses, cavity size, and relative wall thickness were not significant. LV function was significantly improved as measured by ejection fraction (53.2 ± 12.4% to 56.7 ± 10.0%, P < .001), fractional shortening (28.8 ± 11.8% to 30.9 ± 12.5%, P < .05), and midwall fractional shortening (13.5 ± 7.1% to 14.5 ± 6.7%, P < .01) ( Figure 4 ).
Among those undergoing BAV, there were significant reductions ( P < .001 for all) in peak and mean aortic valve gradients (from 73.4 ± 24.2 to 54.6 ± 78.45 mm Hg and from 43.4 ± 15.0 to 31.9 ± 11.1 mm Hg) and increases in aortic valve area (from 0.63 ± 0.21 to 0.79 ± 0.22 cm 2 ) and aortic valve area index (from 0.34 ± 0.11 to 0.44 ± 0.13 cm 2 /m 2 ), which, although significant, were of far less magnitude than those seen in the TAVR group. Despite hemodynamic improvement, the BAV subgroup showed no reduction in LV mass or improvement in systolic LV function ( Supplemental Table 1 ). The prevalence of mitral regurgitation was 25.3% at baseline and 30.1% at 30 days ( P = .1543). There were no changes in aortic valve hemodynamics; LV size, mass, or function; or valvular regurgitation in the no-BAV group.
In an effort to better understand the results of the procedure, we searched for baseline characteristics that differed between those with reductions in mean aortic gradient after BAV of ≥50% ( n = 10) and those with reductions of <50% ( n = 73). Unfortunately, other than lower post-BAV aortic gradients (peak, 39 vs 57 mm Hg [ P = .003]; mean, 23 vs 33 mm Hg [ P = .006]), no clinical or echocardiographic differences at baseline or 30 days were associated with differing degrees of hemodynamic success (data not shown). The prevalence of AR did not change from baseline to 30 days ( P = .44). Moderate or severe AR at 30 days was predicted by baseline echocardiographic parameters of larger LV end-systolic volume ( P = .02), lower ejection fraction ( P < .001), and lower stroke volume ( P = .01), as well as baseline AR ( P < .0001).
Paired Baseline and 1-Year Data
Between 30 days and 1 year, there were 81 deaths (35 TAVR, 46 MM), and six MM patients underwent surgical aortic valve replacement, while none underwent TAVR and were excluded from the analysis ( Figure 1 ). Paired baseline and 1-year echocardiographic data were available in 137 patients (90 TAVR, 47 MM, including 35 BAV and 12 no BAV).
One year after TAVR, patients showed stable peak and mean aortic valve gradients similar to those at 30 days, at 20.3 ± 9.4 and 10.7 ± 5.4 mm Hg ( P < .001 vs baseline for both), and aortic valve area was similarly stable at 1.61 ± 0.47 cm 2 , with an index of 0.9 ± 0.28 cm 2 /m 2 ( P < 0.001 vs baseline for both) ( Supplemental Table 2 ; available at www.onlinejase.com ). The prevalence of moderate or severe total AR was unchanged at 15.7%, and that of moderate to severe mitral regurgitation was 17.8%. There was no change in the severity of paravalvular, transvalvular, or total AR in the TAVR group from 30 days to 1 year. Patients showed further significant LV remodeling with reductions in LV mass and LV mass index (230.6 ± 75 g and 125.0 ± 33.1 g/m 2 , respectively; Supplemental Table 2 ). Reductions in cavity size and individual wall thicknesses were now significant, with LV diameter decreasing from 4.6 ± 0.7 cm at baseline to 4.4 ± 0.8 cm ( P < .05) and posterior wall thickness decreasing from 1.3 ± 0.2 to 1.2 ± 0.2 cm ( P < .01). Relative wall thickness was unchanged. LV function continued to improve, with fractional shortening increasing from 27.6 ± 11.18% to 34.8 ± 12.0% and midwall shortening increasing from 13.30 ± 6.35% to 16.9 ± 7.0% ( P < .0001 vs baseline for both).
As at 30 days, there were no significant changes from baseline to 1 year in LV diameter, wall thickness, hypertrophy, or function in either the MM or BAV groups ( Supplemental Table 2 ). Similarly, there were no changes in paravalvular, transvalvular, or total AR from 30 days to 1 year in either group. The modest improvements in aortic gradients and valve area seen at 30 days were no longer present, with these parameters being similar to paired baseline data as well as across the two MM groups ( Figures 2 and 3 ).
Long-Term Results
The group of TAVR patients with paired 1-year echocardiograms showed further significant LV remodeling, with LV mass and LV mass index at 3 years of 200.4 ± 62.6 g and 112.4 ± 33.4 g/m 2 , respectively ( n = 36; Figure 3 ). LV function remained unchanged from baseline in both groups at 2 years, while aortic valve hemodynamics were unchanged from immediately after TAVR in the TAVR group or baseline in the MM group, although <10 MM patients were alive with echocardiographic data at 3 years. The prevalence of moderate or severe total AR ( Figure 4 A) was 10.4% in the TAVR group at 2 years ( n = 67) and 8.9% at 3 years ( n = 45). At 2 years, the prevalence of moderate or severe total AR was 13.0% in the MM group at 2 years ( n = 23; Figure 4 B), 18% in the BAV group, and zero in the no-BAV group ( Figures 4 C and 4D). The prevalence of moderate to severe mitral regurgitation was 21.2% in the TAVR group at 2 years and 31% at 3 years. There was no moderate to severe mitral regurgitation reported at 2 years in the MM group.
Echocardiographic Variables Associated with Outcomes
Kaplan-Meier curves showing time to death in each of the four groups are shown in Figure 5 . Univariate analysis showed several baseline echocardiographic predictors of death at 2 years in the TAVR group, all associated with lower flow or lower gradient aortic stenosis, including lower stroke volume ( P = .001) and cardiac output ( P = .016) but not cardiac index. Also predictive of death at 2 years in the TAVR group were baseline LV outflow peak velocity ( P = .008) and flow integral ( P = .09), and lower peak ( P = .04) and mean ( P = .03) gradients, but not valve area ( Supplemental Table 3A ; available at www.onlinejase.com ) The only 30 day echocardiographic findings that were predictive of death at 2 years were lower peak and mean aortic valve gradients ( P = .02 and P = .01). Aortic valve area and AR were not predictive in this cohort.
Each of these baseline echocardiographic variables was also predictive of death at 2 years in the MM group. Additional baseline predictors included larger LV size, higher LV mass, worse systolic function, and more severe mitral regurgitation ( Supplemental Table 3B ). Values of these predictors at 30 days were also significantly associated with death; AR at 30 days was not.
Results
TAVR and MM Comparability
A total of 358 patients were randomized to receive either TAVR or MM ( Figure 1 ). Within the MM group, 117 patients (65%) underwent BAV within 30 days, while 60 did not. Demographics and selected baseline characteristics of these “as treated” populations are shown in Table 1 for the 264 subjects with paired baseline and 30 day echocardiographic data. Aside from a lower prevalence of Society of Thoracic Surgeons (STS) score > 10 with TAVR compared with MM (51.7% vs 65.0%, P = .03), there were no significant differences between the TAVR and MM groups. Within the MM group, the no-BAV cohort had lower STS scores and European System for Cardiac Operative Risk Evaluation scores and were more likely to have undergone prior BAV (40% vs 15%).
| Variable | TAVR ( n = 143) | MM ( n = 121) | P value, TAVR vs MM | MM no BAV ( n = 36) | MM BAV ( n = 85) | P value, no BAV vs BAV |
|---|---|---|---|---|---|---|
| Demographics and risk score | ||||||
| Age (y) | 83.19 ± 8.92 | 82.81 ± 9.03 | .73 | 82.09 ± 10.19 | 83.11 ± 8.54 | .57 |
| Men | 44.8% (64/143) | 47.9% (58/121) | .61 | 50.0% (18/36) | 47.1% (40/85) | .77 |
| Caucasian race | 93.7% (134/143) | 91.7% (111/121) | .54 | 86.1% (31/36) | 94.1% (80/85) | .14 |
| BMI (kg/m 2 ) | 27.02 ± 7.96 | 26.79 ± 6.08 | .79 | 26.71 ± 6.70 | 26.82 ± 5.84 | .94 |
| STS score > 10 | 51.7% (74/143) | 65.0% (78/120) | .03 | 51.4% (18/35) | 70.6% (60/85) | .05 |
| Logistic EuroSCORE | 27.65 ± 17.50 | 29.44 ± 18.61 | .42 | 23.05 ± 15.85 | 32.15 ± 19.11 | .01 |
| C risk factors | ||||||
| Diabetes | 31.5% (45/143) | 33.1% (40/121) | .78 | 44.4% (16/36) | 28.2% (24/85) | .08 |
| Hyperlipidemia | 73.4% (105/143) | 81.8% (99/121) | .10 | 80.6% (29/36) | 82.4% (70/85) | .81 |
| Smoking | 43.4% (62/143) | 49.6% (60/121) | .31 | 52.8% (19/36) | 48.2% (41/85) | .65 |
| Hypertension | 86.6% (123/142) | 86.8% (105/121) | .97 | 88.9% (32/36) | 85.9% (73/85) | .66 |
| Cardiovascular conditions | ||||||
| CHF | 98.6% (141/143) | 96.7% (117/121) | .42 | 97.2% (35/36) | 96.5% (82/85) | .83 |
| CAD | 69.9% (100/143) | 73.6% (89/121) | .52 | 72.2% (26/36) | 74.1% (63/85) | .83 |
| Cerebrovascular disease | 29.1% (41/141) | 28.7% (33/115) | .95 | 31.4% (11/35) | 27.5% (22/80) | .67 |
| Peripheral vascular disease | 28.2% (40/142) | 25.6% (31/121) | .64 | 27.8% (10/36) | 24.7% (21/85) | .72 |
| Prior BAV | 14.0% (20/143) | 22.3% (27/121) | .08 | 38.9% (14/36) | 15.3% (13/85) | .004 |
| 1 | 9.8% (14/143) | 20.7% (25/121) | .01 | 36.1% (13/36) | 14.1% (12/85) | .006 |
| 2 | 2.8% (4/143) | 0.8% (1/121) | .38 | 0.0% (0/36) | 1.2% (1/85) | .51 |
| ≥3 | 1.4% (2/143) | 0.8% (1/121) | 1.00 | 2.8% (1/36) | 0.0% (0/85) | .12 |
| Noncardiac conditions | ||||||
| Renal disease (CR ≥ 2 mg/dL) | 19.6% (28/143) | 19.2% (23/120) | .93 | 19.4% (7/36) | 19.0% (16/84) | .96 |
| COPD | 39.9% (57/143) | 50.4% (61/121) | .09 | 50.0% (18/36) | 50.6% (43/85) | .95 |
| Chest wall radiation | 10.5% (15/143) | 8.3% (10/121) | .54 | 11.1% (4/36) | 7.1% (6/85) | .46 |
At baseline, patients receiving TAVR compared with MM showed similar LV wall thickness, mass, and concentric hypertrophy ( Supplemental Table 1 ; available on www.onlinejase.com ). LV end-systolic volumes (69.1 ± 48.0 vs 54.5 ± 29.3 mL, P = .03) were higher, and ejection fraction tended to be slightly lower in the MM group, at 50 ± 14% versus 53 ± 12% ( P = .07). Within the MM group, there were no differences between BAV and no-BAV subjects. Similarly, there were no differences at baseline in aortic valve hemodynamics or regurgitation or mitral regurgitation between TAVR and MM. Within the MM group, the no-BAV group tended to have a lower mean gradient (38 vs 43 mm Hg, P = .06) and higher valve area (0.70 vs 0.63 cm 2 , P = .09) than the BAV group, although these differences did not reach significance.
Paired Baseline and 30-Day Data
Over the first 30 days, there were 16 deaths (11 TAVR, 5 MM) and four patients who converted from MM to surgical valve replacement ( Figure 1 ). Paired baseline and 30-day echocardiographic data were available in 264 patients (143 TAVR, 121 MM) for analysis of changes in aortic valve hemodynamics, regurgitation, and LV structure and function ( Supplemental Table 1 ).
Thirty days after TAVR, peak and mean aortic valve gradients fell markedly: from 73.3 ± 24.1 to 19.8 ± 8.4 mm Hg and from 43.8 ± 14.7 to 10.0 ± 4.3 mm Hg, respectively ( P < .001 for both), while aortic valve area increased from 0.64 ± 0.18 to 1.56 ± 0.43 cm 2 and aortic valve area index rose from 0.36 ± 0.11 to 0.89 ± 0.26 cm 2 /m 2 ( Figure 2 ). In the TAVR group, the mean cover index was 12.93 ± 5.84, and the prevalence of moderate or severe prosthesis-patient mismatch was 48.4% at 30 days. The prevalence of moderate or severe total aortic regurgitation (AR) was 21.2% at baseline and 13.5% at 30 days ( Figure 3 ) Moderate or severe AR was not predicted by any baseline echocardiographic parameter measured or by cover index. There were no echocardiographic predictors of moderate to severe AR, although such patients were more likely to have histories of chest wall irradiation ( P = .006) and less likely to have diabetes ( P = .007). The prevalence of mitral regurgitation was unchanged from baseline to 30 days ( P = .125). TAVR patients showed significant LV remodeling, with reduced LV mass and LV mass index (from 257.0 ± 76.4 to 251.3 ± 85.8 g and from 144.7 ± 36.1 to 140.0 ± 37.9 g/m 2 , P < .05 for both; Supplemental Table 1 ), while changes in wall thicknesses, cavity size, and relative wall thickness were not significant. LV function was significantly improved as measured by ejection fraction (53.2 ± 12.4% to 56.7 ± 10.0%, P < .001), fractional shortening (28.8 ± 11.8% to 30.9 ± 12.5%, P < .05), and midwall fractional shortening (13.5 ± 7.1% to 14.5 ± 6.7%, P < .01) ( Figure 4 ).
Among those undergoing BAV, there were significant reductions ( P < .001 for all) in peak and mean aortic valve gradients (from 73.4 ± 24.2 to 54.6 ± 78.45 mm Hg and from 43.4 ± 15.0 to 31.9 ± 11.1 mm Hg) and increases in aortic valve area (from 0.63 ± 0.21 to 0.79 ± 0.22 cm 2 ) and aortic valve area index (from 0.34 ± 0.11 to 0.44 ± 0.13 cm 2 /m 2 ), which, although significant, were of far less magnitude than those seen in the TAVR group. Despite hemodynamic improvement, the BAV subgroup showed no reduction in LV mass or improvement in systolic LV function ( Supplemental Table 1 ). The prevalence of mitral regurgitation was 25.3% at baseline and 30.1% at 30 days ( P = .1543). There were no changes in aortic valve hemodynamics; LV size, mass, or function; or valvular regurgitation in the no-BAV group.
In an effort to better understand the results of the procedure, we searched for baseline characteristics that differed between those with reductions in mean aortic gradient after BAV of ≥50% ( n = 10) and those with reductions of <50% ( n = 73). Unfortunately, other than lower post-BAV aortic gradients (peak, 39 vs 57 mm Hg [ P = .003]; mean, 23 vs 33 mm Hg [ P = .006]), no clinical or echocardiographic differences at baseline or 30 days were associated with differing degrees of hemodynamic success (data not shown). The prevalence of AR did not change from baseline to 30 days ( P = .44). Moderate or severe AR at 30 days was predicted by baseline echocardiographic parameters of larger LV end-systolic volume ( P = .02), lower ejection fraction ( P < .001), and lower stroke volume ( P = .01), as well as baseline AR ( P < .0001).
Paired Baseline and 1-Year Data
Between 30 days and 1 year, there were 81 deaths (35 TAVR, 46 MM), and six MM patients underwent surgical aortic valve replacement, while none underwent TAVR and were excluded from the analysis ( Figure 1 ). Paired baseline and 1-year echocardiographic data were available in 137 patients (90 TAVR, 47 MM, including 35 BAV and 12 no BAV).
One year after TAVR, patients showed stable peak and mean aortic valve gradients similar to those at 30 days, at 20.3 ± 9.4 and 10.7 ± 5.4 mm Hg ( P < .001 vs baseline for both), and aortic valve area was similarly stable at 1.61 ± 0.47 cm 2 , with an index of 0.9 ± 0.28 cm 2 /m 2 ( P < 0.001 vs baseline for both) ( Supplemental Table 2 ; available at www.onlinejase.com ). The prevalence of moderate or severe total AR was unchanged at 15.7%, and that of moderate to severe mitral regurgitation was 17.8%. There was no change in the severity of paravalvular, transvalvular, or total AR in the TAVR group from 30 days to 1 year. Patients showed further significant LV remodeling with reductions in LV mass and LV mass index (230.6 ± 75 g and 125.0 ± 33.1 g/m 2 , respectively; Supplemental Table 2 ). Reductions in cavity size and individual wall thicknesses were now significant, with LV diameter decreasing from 4.6 ± 0.7 cm at baseline to 4.4 ± 0.8 cm ( P < .05) and posterior wall thickness decreasing from 1.3 ± 0.2 to 1.2 ± 0.2 cm ( P < .01). Relative wall thickness was unchanged. LV function continued to improve, with fractional shortening increasing from 27.6 ± 11.18% to 34.8 ± 12.0% and midwall shortening increasing from 13.30 ± 6.35% to 16.9 ± 7.0% ( P < .0001 vs baseline for both).
As at 30 days, there were no significant changes from baseline to 1 year in LV diameter, wall thickness, hypertrophy, or function in either the MM or BAV groups ( Supplemental Table 2 ). Similarly, there were no changes in paravalvular, transvalvular, or total AR from 30 days to 1 year in either group. The modest improvements in aortic gradients and valve area seen at 30 days were no longer present, with these parameters being similar to paired baseline data as well as across the two MM groups ( Figures 2 and 3 ).
Long-Term Results
The group of TAVR patients with paired 1-year echocardiograms showed further significant LV remodeling, with LV mass and LV mass index at 3 years of 200.4 ± 62.6 g and 112.4 ± 33.4 g/m 2 , respectively ( n = 36; Figure 3 ). LV function remained unchanged from baseline in both groups at 2 years, while aortic valve hemodynamics were unchanged from immediately after TAVR in the TAVR group or baseline in the MM group, although <10 MM patients were alive with echocardiographic data at 3 years. The prevalence of moderate or severe total AR ( Figure 4 A) was 10.4% in the TAVR group at 2 years ( n = 67) and 8.9% at 3 years ( n = 45). At 2 years, the prevalence of moderate or severe total AR was 13.0% in the MM group at 2 years ( n = 23; Figure 4 B), 18% in the BAV group, and zero in the no-BAV group ( Figures 4 C and 4D). The prevalence of moderate to severe mitral regurgitation was 21.2% in the TAVR group at 2 years and 31% at 3 years. There was no moderate to severe mitral regurgitation reported at 2 years in the MM group.
Echocardiographic Variables Associated with Outcomes
Kaplan-Meier curves showing time to death in each of the four groups are shown in Figure 5 . Univariate analysis showed several baseline echocardiographic predictors of death at 2 years in the TAVR group, all associated with lower flow or lower gradient aortic stenosis, including lower stroke volume ( P = .001) and cardiac output ( P = .016) but not cardiac index. Also predictive of death at 2 years in the TAVR group were baseline LV outflow peak velocity ( P = .008) and flow integral ( P = .09), and lower peak ( P = .04) and mean ( P = .03) gradients, but not valve area ( Supplemental Table 3A ; available at www.onlinejase.com ) The only 30 day echocardiographic findings that were predictive of death at 2 years were lower peak and mean aortic valve gradients ( P = .02 and P = .01). Aortic valve area and AR were not predictive in this cohort.
Each of these baseline echocardiographic variables was also predictive of death at 2 years in the MM group. Additional baseline predictors included larger LV size, higher LV mass, worse systolic function, and more severe mitral regurgitation ( Supplemental Table 3B ). Values of these predictors at 30 days were also significantly associated with death; AR at 30 days was not.
Discussion
In this comprehensive, longitudinal echocardiographic study comparing outcomes in patients with inoperable symptomatic, severe aortic stenosis treated with TAVR or MM, we found the following: (1) TAVR was associated with an immediate and sustained reduction in transaortic pressure gradients, with an associated increase in aortic valve area; (2) the hemodynamic improvement in the TAVR group was accompanied by early improvement in LV systolic function and ongoing, gradual reduction in LV mass; (3) BAV was associated with a modest and short-term reduction in pressure overload, but no increase in AR, improvement in LV function, or LV mass reduction; and (3) MM without BAV is not associated with a change in aortic valve hemodynamics, LV function, or remodeling. These data provide mechanistic insights into the differing clinical outcomes associated with these two procedures.
The hemodynamic and remodeling changes accompanying TAVR have been previously reported in high-risk but operable patients but not in inoperable patients, whose higher risk status is well documented to affect clinical outcomes. The changes in LV mass and systolic function and aortic valve hemodynamics at 30 days, 1 year, and 2 years seen in the inoperable cohort are similar to those reported for the high-risk cohort. Given the similarity between operable and inoperable TAVR patients at baseline, this indicates that, even in very sick inoperable patients, the relief of pressure overload is accompanied by a variety of both immediate and longer term positive effects. Indeed, the continued regression of LV hypertrophy at 3 years in the TAVR group demonstrates ongoing plasticity of the ventricle occurring over years, which has not previously been appreciated.
AR is a significant concern in both TAVR (paravalvular leak) and BAV, albeit produced by very different mechanisms. The prevalence of moderate or severe AR in these inoperable TAVR patients is similar to that reported in operable patients but was not associated with an increased risk for death. This may be due to the greater burden of comorbidities and competing causes of mortality among these inoperable patients.
As expected, there were no immediate changes in the MM group without BAV. Of note, however, is that there was no further worsening of either hemodynamics or ventricular function or hypertrophy over ≥1 year of follow-up in patients with paired data. This paired analysis does not capture the changes in those who died before the 1-year time point and thus may represent a cohort of patients with better outcomes.
The BAV patients represent a special subset of the MM cohort, and the present study is an opportunity to explore the effects of this procedure in current practice. Compared with the MM patients, in whom investigators chose not to perform BAV, these patients had higher STS scores and logistic European System for Cardiac Operative Risk Evaluation scores, were less likely to have undergone prior BAV, and were more likely to have lower midwall fractional shortening. In addition, there were strong trends for BAV patients to have higher baseline peak and mean gradients with lower aortic valve area, possibly because of residual effects of prior BAV procedures in the MM group. It is important to note that a post-BAV reduction in mean gradient of ≥50 mm Hg was achieved in only a small minority of patients (12%), indicating the difficulty of producing a meaningful reduction in hemodynamic load with this procedure. Perhaps as a result, BAV was associated with overall modest reductions in aortic valve gradients, although the increase in valve area was at least as great as that reported in other contemporary series. However, in contrast to these studies, the improved hemodynamics in this PARTNER cohort after BAV failed to result in a significant improvement in ventricular function and likely reflects the “inoperability” of these patients and significant comorbidities, including coronary artery disease or intrinsic myocardial dysfunction, frailty, concomitant pulmonary disease, and renal disease. This possibility is supported by our findings of no significant change in regurgitation after BAV and the lack of association of moderate or severe AR at 30 days with death at 2 years in these patients. Although others have reported both an increase in AR and no change, these findings exclude an increase in AR secondary to BAV as a contributor to poor longer term outcome. Importantly, the time course of the echocardiographic findings closely matches that of survival and quality of life, with transient improvements in survival at 3 months compared with no-BAV patients and improved quality-of-life measures at 6 months compared with baseline, which disappeared by 6 and 12 months, respectively, as do the hemodynamic improvements. This finding has not been previously documented, although it is supported by those of Kefer et al ., who showed an association of early reduction in B-type natriuretic peptide postprocedurally, with improved LV function at 30 days, which may reflect a physiologically meaningful relief of afterload in this group.
Taken together, these data suggest that the hemodynamic benefits of BAV are modest and transient and that the very small hemodynamic benefit seen may be of limited use in most patients, even as a staged procedure or to assess a patient’s suitability for a more definitive procedure. BAV is, at best, likely a palliative tool rather than a long-term solution. We were, however, able to identify some features that were associated with a mean gradient of <20 mm Hg after BAV, which may assist in patient selection, including male sex, more LV hypertrophy, and a larger aortic root. A recent analysis of the PARTNER trial suggests that despite a higher STS score and greater prevalence of comorbidities, patients undergoing BAV as a bridge to intervention had similar 30-day outcomes afterollowing TAVR compared with those not undergoing BAV.
Limitations
Although we used paired echocardiographic data to evaluate changes over time, the high mortality rate reduced the size of our cohorts and may have introduced survivorship bias or a failure to detect important premortem changes occurring before the 1-year time point. Similarly, we are unable to describe with greater granularity the time course of increasing aortic valve gradients or the possible impact of moderate to severe AR after BAV between 30 days and 1 year. We excluded from the 30-day analysis the 4 patients in the BAV group who underwent subsequent surgical aortic valve replacement before the 30-day echocardiographic study and excluded from the 1-year analysis the six patients who underwent surgical aortic valve replacement between 30 days and 1 year; these patients may have had more benefit from BAV. Although we evaluated the echocardiographic changes and predictors of a post-BAV mean gradient of <20 mm Hg, there is no standard definition or cut point for a desirable outcome after BAV. The small size of the group receiving MM without BAV makes it difficult to draw conclusions. Finally, our findings are limited to those available in the PARTNER database. Thus, we are unable to assess diastolic function or strain, which may provide additional insights into the effects of BAV as well as TAVR.
Conclusions
This study is the first to use centrally analyzed echocardiographic data to directly compare the longitudinal structural and hemodynamic results of TAVR and MM with and without BAV in a randomized inoperable population of patients with symptomatic severe aortic stenosis. Good clinical outcomes with TAVR are explained by relief pressure overload and improvement of LV systolic function, with regression of hypertrophy continuing over a longer term. In BAV patients, this detailed longitudinal analysis of hemodynamics and LV remodeling delineates the mechanisms of previously observed poor clinical results, documenting reductions in pressure overload with BAV that are modest, transient, and not accompanied by other positive effects. Patients managed medically without BAV show few changes over time. TAVR is the preferred treatment in inoperable patients eligible for this procedure.
Supplementary data
| Variable | TAVR ( n = 143) | MM ( n = 121) | P value, TAVR vs MM | MM no BAV ( n = 36) | MM BAV ( n = 85) | P value, no BAV vs BAV |
|---|---|---|---|---|---|---|
| Cardiac measurements and function | ||||||
| IVSD dimension (cm) | ||||||
| Baseline | 1.54 ± 0.33 (128) | 1.61 ± 0.30 (108) | .12 | 1.69 ± 0.34 (32) | 1.57 ± 0.28 (76) | .07 |
| 30 d | 1.52 ± 0.34 (128) | 1.60 ± 0.35 (111) | .06 | 1.64 ± 0.38 (33) | 1.59 ± 0.33 (78) | .54 |
| LVED dimension (cm) | ||||||
| Baseline | 4.46 ± 0.76 (128) | 4.50 ± 0.81 (108) | .69 | 4.34 ± 0.90 (32) | 4.56 ± 0.77 (76) | .18 |
| 30 d | 4.40 ± 0.83 (128) | 4.46 ± 0.78 (111) | .57 | 4.46 ± 0.76 (33) | 4.45 ± 0.80 (78) | .95 |
| LVPWD dimension (cm) | ||||||
| Baseline | 1.32 ± 0.25 (128) | 1.35 ± 0.28 (108) | .33 | 1.41 ± 0.28 (32) | 1.33 ± 0.27 (76) | .15 |
| 30 d | 1.32 ± 0.27 (128) | 1.36 ± 0.27 (111) | .24 | 1.38 ± 0.29 (33) | 1.34 ± 0.27 (78) | .46 |
| LV relative wall thickness | ||||||
| Baseline | 0.61 ± 0.16 (128) | 0.63 ± 0.22 (108) | .41 | 0.68 ± 0.21 (32) | 0.61 ± 0.23 (76) | .09 |
| 30 d | 0.62 ± 0.20 (128) | 0.64 ± 0.22 (111) | .62 | 0.64 ± 0.20 (33) | 0.64 ± 0.22 (78) | .82 |
| LV mass (g) | ||||||
| Baseline | 256.95 ± 76.39 (128) | 273.55 ± 80.86 (108) | .11 | 280.82 ± 95.35 (32) | 270.48 ± 74.42 (76) | .53 |
| 30 d | 251.34 ± 85.82 (128) ∗ | 269.82 ± 76.91 (111) | .08 | 279.95 ± 86.48 (33) | 265.54 ± 72.67 (78) | .40 |
| LV mass index (g/m 2 ) | ||||||
| Baseline | 144.69 ± 36.14 (128) | 152.56 ± 46.89 (108) | .16 | 157.42 ± 56.50 (32) | 150.51 ± 42.46 (76) | .43 |
| 30 d | 140.00 ± 37.87 (122) ∗ | 153.62 ± 44.06 (99) | .01 | 157.76 ± 46.61 (31) | 151.73 ± 43.07 (68) | .49 |
| Ascending aorta (cm) | ||||||
| Baseline | 2.97 ± 0.39 (73) | 2.98 ± 0.46 (64) | .92 | 3.05 ± 0.49 (19) | 2.95 ± 0.45 (45) | .38 |
| 30 d | 3.01 ± 0.50 (43) | 3.08 ± 0.43 (46) † | .52 | 3.27 ± 0.51 (13) ∗ | 3.00 ± 0.38 (33) | .08 |
| LVED volume (mL) | ||||||
| Baseline | 114.90 ± 41.55 (92) | 126.43 ± 54.41 (64) | .16 | 117.16 ± 54.85 (20) | 130.65 ± 54.31 (44) | .29 |
| 30 d | 118.81 ± 49.49 (82) | 116.31 ± 51.53 (61) | .77 | 114.32 ± 46.15 (16) | 117.02 ± 53.78 (45) | .85 |
| LVES volume (mL) | ||||||
| Baseline | 54.50 ± 29.30 (92) | 69.11 ± 48.00 (64) | .03 | 64.32 ± 44.63 (20) | 71.29 ± 49.79 (44) | .50 |
| 30 d | 53.22 ± 34.03 (82) | 60.84 ± 41.71 (61) | .25 | 60.10 ± 40.49 (16) | 61.11 ± 42.59 (45) | .93 |
| Doppler stroke volume | ||||||
| Baseline | 62.04 ± 17.39 (135) | 59.87 ± 21.38 (115) | .38 | 64.09 ± 26.90 (33) | 58.17 ± 18.64 (82) | .23 |
| 30 d | 64.40 ± 19.86 (133) | 61.04 ± 17.49 (118) | .16 | 60.32 ± 21.76 (34) | 61.34 ± 15.58 (84) | .79 |
| Doppler cardiac output (L/min) | ||||||
| Baseline | 4.58 ± 1.36 (135) | 4.42 ± 1.49 (114) | .38 | 4.62 ± 1.65 (33) | 4.33 ± 1.42 (81) | .33 |
| 30 d | 4.76 ± 1.56 (133) | 4.58 ± 1.52 (117) | .36 | 4.39 ± 1.48 (34) | 4.66 ± 1.54 (83) | .40 |
| Doppler cardiac index (L/min/m 2 ) | ||||||
| Baseline | 2.61 ± 0.84 (135) | 2.46 ± 0.84 (114) | .17 | 2.54 ± 0.81 (33) | 2.43 ± 0.85 (81) | .51 |
| 30 d | 2.73 ± 0.92 (126) | 2.61 ± 0.89 (105) | .32 | 2.41 ± 0.70 (32) | 2.69 ± 0.95 (73) ∗ | .15 |
| Fractional shortening (%) | ||||||
| Baseline | 28.78 ± 11.77 (124) | 27.50 ± 11.90 (107) | .42 | 30.15 ± 11.37 (32) | 26.38 ± 12.02 (75) | .13 |
| 30 d | 30.88 ± 12.53 (118) ∗ | 27.24 ± 12.25 (109) | .03 | 26.58 ± 13.25 (31) | 27.50 ± 11.90 (78) | .73 |
| Midwall fractional shortening (%) | ||||||
| Baseline | 13.50 ± 7.14 (124) | 12.38 ± 6.98 (101) | .24 | 14.50 ± 7.39 (31) | 11.44 ± 6.63 (70) | .05 |
| 30 d | 14.51 ± 6.66 (116) † | 12.42 ± 7.81 (103) | .04 | 12.10 ± 8.62 (30) | 12.55 ± 7.51 (73) | .78 |
| LV ejection fraction (%) | ||||||
| Baseline | 54.04 ± 13.01 (142) | 50.44 ± 14.37 (119) | .03 | 51.13 ± 13.43 (34) | 50.16 ± 14.79 (85) | .73 |
| 30 d | 57.25 ± 10.19 (143) ‡ | 51.87 ± 14.01 (120) | .0006 | 52.24 ± 14.01 (36) | 51.71 ± 14.09 (84) | .83 |
| Aortic valve hemodynamics | ||||||
| AV peak gradient (mm Hg) | ||||||
| Baseline | 73.34 ± 24.13 (140) | 70.74 ± 23.08 (118) | .38 | 64.42 ± 18.95 (35) | 73.41 ± 24.23 (83) | .06 |
| 30 d | 19.85 ± 8.39 (139) ‡ | 57.33 ± 19.23 (119) ‡ | <.0001 | 64.28 ± 19.67 (34) | 54.55 ± 18.45 (85) ‡ | .002 |
| AV mean gradient (mm Hg) | ||||||
| Baseline | 43.61 ± 14.68 (140) | 41.71 ± 14.25 (118) | .30 | 37.70 ± 11.42 (35) | 43.40 ± 15.03 (83) | .053 |
| 30 d | 10.05 ± 4.30 (139) ‡ | 33.54 ± 11.82 (119) ‡ | <.0001 | 37.70 ± 12.66 (34) | 31.87 ± 11.12 (85) ‡ | .002 |
| AV area (EOA) (cm 2 ) | ||||||
| Baseline | 0.64 ± 0.18 (136) | 0.65 ± 0.22 (115) | .76 | 0.70 ± 0.24 (33) | 0.63 ± 0.21 (82) | .09 |
| 30 d | 1.56 ± 0.43 (133) ‡ | 0.76 ± 0.22 (118) ‡ | <.0001 | 0.68 ± 0.20 (34) | 0.79 ± 0.22 (84) ‡ | .12 |
| AV area index (cm 2 /m 2 ) | ||||||
| Baseline | 0.36 ± 0.11 (136) | 0.36 ± 0.12 (115) | .82 | 0.39 ± 0.12 (33) | 0.35 ± 0.11 (82) | .12 |
| 30 d | 0.89 ± 0.26 (126) ‡ | 0.42 ± 0.13 (106) ‡ | <.0001 | 0.37 ± 0.11 (32) | 0.44 ± 0.13 (74) ‡ | .12 |
| Total AR: baseline | .2954 | .1266 | ||||
| None/trace | 35.9% (51/142) | 39.3% (46/117) | 50.0% (17/34) | 34.9% (29/83) | ||
| Mild | 43.0% (61/142) | 47.0% (55/117) | 32.4% (11/34) | 53.0% (44/83) | ||
| Moderate/severe | 21.1% (30/142) | 13.7% (16/117) | 17.6% (6/34) | 12.0% (10/83) | ||
| 30 d | .1220 | .9702 | ||||
| None/trace | 25.5% (36/141) | 34.2% (41/120) | 33.3% (12/36) | 34.5% (29/84) | ||
| Mild | 61.0% (86/141) | 48.3% (58/120) | 50.0% (18/36) | 47.6% (40/84) | ||
| Moderate/severe | 13.5% (19/141) | 17.5% (21/120) | 16.7% (6/36) | 17.9% (15/84) | ||
| P vs baseline | .0039 | .3488 | .2959 | .4422 | ||
| Paravalvular AR: 30 d | NA | NA | ||||
| None/trace | 35.5% (50/141) | NA | NA | NA | ||
| Mild | 52.5% (74/141) | NA | NA | NA | ||
| Moderate/severe | 12.1% (17/141) | NA | NA | NA | ||
| Mitral regurgitation: baseline | .6287 | .3655 | ||||
| None/trace | 28.2% (40/142) | 29.9% (35/117) | 35.3% (12/34) | 27.7% (23/83) | ||
| Mild | 49.3% (70/142) | 43.6% (51/117) | 47.1% (16/34) | 42.2% (35/83) | ||
| Moderate/severe | 22.5% (32/142) | 26.5% (31/117) | 17.6% (6/34) | 30.1% (25/83) | ||
| 30 d | .6961 | .8340 | ||||
| None/trace | 38.3% (54/141) | 34.5% (41/119) | 33.3% (12/36) | 34.9% (29/83) | ||
| Mild | 39.0% (55/141) | 38.7% (46/119) | 36.1% (13/36) | 39.8% (33/83) | ||
| Moderate/severe | 22.7% (32/141) | 26.9% (32/119) | 30.6% (11/36) | 25.3% (21/83) | ||
| P vs baseline | .1250 | .4085 | .2136 | .1543 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
