High Residual Gradient Following a Self-Expandable Transcatheter Aortic Valve-in-Valve Implantation — Risk Factor Analysis, Outcomes, and Survival

Introduction Transcatheter aortic valve-in-valve implantation (TAVI-ViV) can be associated with unfavorable hemodynamic outcomes. This study aimed to estimate the prevalence, identify the risk factors, and evaluate the outcomes and survival of patients with high residual gradients after TAVI-ViV. Methods A total of 85 patients were included in the study. The cohort was divided into group A, with postprocedural mean pressure gradient (PG) ≥ 20 mmHg, and group B, with mean PG < 20 mmHg. Results Postprocedural PG ≥ 20 mmHg was observed in 24.7% of the patients. In a univariate analysis, preoperative gradient, pre-existing patient-prosthesis mismatch (PPM), deep valve implantation, small degenerated valves, and an older generation of transcatheter aortic valves were found to be risk factors for high residual gradient. Multivariate analysis showed that preexisting maxPG > 60 mmHg, implantation level of 4 mm below neo-annulus, and degenerated valve size ≤ 23 mm were independent predictors of high residual gradient. There were no differences in early morbidity (myocardial infarction, pacemaker implantation, stroke, acute renal insufficiency) between groups. Kaplan-Meier estimated that the survival rate was comparable at one and five years regardless of postoperative gradient. Survivors with high residual mean gradient were significantly affected by a high New York Heart Association (NYHA) class. Conclusion High residual transvalvular gradient after TAVI-ViV is not rare, but it does not significantly affect mortality. High residual mean gradient has a negative impact on NYHA functional class improvement after the procedure. High preoperative gradient, implantation level, and small failed bioprosthesis may predispose to increased residual gradient.


INTRODUCTION
Transcatheter treatment of failed aortic bioprosthesis has emerged as a safe and effective therapy and is a less invasive approach than chest reopening surgery, especially in patients with prohibitive or high surgical reoperative risk [1][2][3][4] . Due to the old valve being left behind, transcatheter aortic valve-in-valve implantation (TAVI-ViV) can be associated with unfavorable hemodynamic outcomes, such as increased transprosthetic gradients [5][6][7] . The updated Valve Academic Research Consortium (or VARC-2) recognizes postprocedural transvalvular mean gradient > 20 mmHg as prosthetic valve dysfunction [8] . However, the influence of high residual gradient following TAVI-ViV on post-implantation outcomes remains controversial and long-term consequences of higher gradient seem to be unclear [9][10][11] . This study aimed to identify risk factors and evaluate the outcomes and survival of patients with high residual gradients after TAVI-ViV.

Design and Population
We retrospectively analyzed 86 consecutive high-risk patients with degenerated aortic bioprostheses who underwent TAVI-ViV between March 2010 and July 2019 at Sana Heart Center in Cottbus, Germany. All patients were discussed by the Heart Team consisting of a cardiac surgeon, an interventional cardiologist, and cardiac anesthesiologists and they were disqualified from a conventional repeat surgery due to a high-risk profile. Individuals with acute endocarditis, requiring concomitant cardiac procedures, with previously implanted mechanical or transcatheter valves were excluded. One patient was excluded due to the intraoperative death caused by a perforation of the left ventricle by the guidewire. Finally, 85 patients met the eligibility criteria and were included in the study. The cohort

Statistical Analysis
Continuous variables were expressed as means ± standard deviation, while categorical variables were expressed as number and percentages. For continuous data, the Student's t-test or Mann-Whitney's U-test were used for between groups comparisons, while categorical variables were compared with Pearson χ 2 test. To identify the independent predictors of high residual gradient (meanPG ≥ 20 mmHg) after TAVI-ViV, we built a multivariate logistic regression model for the whole cohort by using all preoperative variables presented in Table 1 in addition to intraoperative indices such as implantation height or valve type (CoreValve™ or Evolut™ R). A multivariate logistic regression analysis was performed using a stepwise backward regression including only factors identified during the univariate analysis with a P-value ≤ 1. Survival curves were calculated using the Kaplan-Meier estimator and the comparison between both groups was made using the log-rank test (Mantel-Cox test). Statistical significance was assumed at P<0.05. The statistical analysis was computed with STATISTICA ver. 13 for Windows software (TIBCO StatSoft, Inc., Tulsa, Oklahoma, USA). Routine admission for operation f Intervention or surgery is performed on the current admission for medical reasons and these patients cannot be sent home without a definitive procedure g Operation before the beginning of the next working day after decision to operate

Baseline Characteristics
A total of 85 patients were included in the final analysis (45.9% male, mean age 79.8 ± 5.7 years, European System for Cardiac Operative Risk Evaluation II 11.8 ± 6.0%). All demographics and preoperative clinical data of the study population were summarized in Table 1. During the postprocedural echocardiography on days 4-13 after TAVI-ViV, high residual gradient was observed in 21 patients (24.7%). In patients with higher postoperative meanPG, a history of patient-prosthesis mismatch (PPM) after initial surgery, higher gradients pre TAVI-ViV, and smaller and stented degenerated valves were significantly more prevalent.

Operative Data
The procedure was performed mainly using the femoral approach (97.6%) and conscious sedation with local anesthesia (91.8%). Pre-dilatation was a standard manner and postdilatation was required in three patients to get the fully expanded valve. No neurological protection, coronary protection system, or valve fracture were used. All implanted valves were self-expandable, CoreValve™ (38.8%) or CoreValve™ Evolut™ R (61.2%). Repositioning of the valve was possible only in the new generation valve (CoreValve™ Evolut™ R) and was needed approximately in half of these patients. Procedures resulting in higher postprocedural meanPG were performed more often with the older valve type, lasted significantly longer, and the new valve was implanted deeper ( Table 2).

In-hospital Outcomes
Four patients (4.7%) died during the first 30 days after the procedure, one patient in group A and three patients in group B (P=0.989). There were no differences in the ICU stay, in-hospital stay, and complication rates between the groups. Kidney function improved after the procedure, regardless of postprocedural mean gradient (Groups A +18.4% vs. +14.4% Group B, P=0.243). Twenty-six patients (30.6%) had postoperative paravalvular leak, four in group A and 19 in group B. Moderate paravalvular leak occurred in three patients in group B. There were no cases of postprocedural severe paravalvular leak. Early complications (< 30 days post TAVI-ViV) are summarized in Table 4. respectively. Their results are comparable to these presented in the Placement of Aortic Transcatheter Valves 2 valve-in-valve registry. Webb et al. [9] observed the rate of 34.3% patients with high postprocedural residual gradient. They observed a significant higher mortality at one year in patients with postoperative elevated mean gradient (≥ 20 mmHg) (16.7% vs. 7.7%, P=0.01); however, high gradient was no more a risk factor of all-cause mortality during a three-year follow-up (P=0.15) [1] . Opposite to these findings, there is substantial evidence suggesting no effect of hemodynamic results on mortality after TAVI-ViV. Akodad et al. [11] compared patients who underwent TAVI-ViV with patients after native transcatheter aortic valve implantation and the repeat procedure was associated with higher postoperative gradient at one month (18.3 vs. 11.6 mmHg, P=0.0004) and one year (18.1 vs. 11.4 mmHg, P<0.0001). TAVI-ViV was also associated with a higher rate of patients with mean aortic gradient ≥ 20 mmHg (37.5% vs. 8.4%, P=0.0002); however, the hemodynamical outcomes do not affect the one-year mortality from a cardiac cause after both procedures (TAVI-ViV 2.1% vs. nonTAVI-ViV 2.4%, P=0.9). Guimarães et al. [10] analyzed a six-year experience of nine heart centers with TAVI-ViV. Nearly 40% of the patients presented postoperatively high residual gradient > 20 mmHg and they did not observe any significant differences in longterm survival after TAVI-ViV. Authors emphasize the complexity of these group of patients suggesting clinical characteristics and extension of comorbidities as more important in mortality than echocardiographic outcomes. The need of aortic valve reintervention due to failed bioprosthesis concerns mostly aged patients with a number of comorbidities; therefore, the risk of a repeat open-chest surgery is always significantly increased. Over

Follow-up Period
Twenty-seven patients (31.8%) died during the study period over nine years. Kaplan-Meier estimated that the survival rate after one, two, and five years was 87.7%, 81.5%, and 61.7%, respectively. The postoperative transaortic gradient had no impact on survival -group A vs.  (Figure 3). At the end of the follow-up period, 50 out of 58 survivors (86.2%) were found in NYHA I or II functional classes (group A 69.2% vs. group B 91.1%, P=0.044). One survivor with discharge meanPG 30 mmHg was found in NYHA Class IV seven years after TAVI-ViV and required re-TAVI-ViV.

DISCUSSION
Increased transprosthetic gradients following TAVI-ViV are frequent and range from 17% to 44%, what is similar to our results [9][10][11][12][13][14] . There is some evidence that unfavorable hemodynamic outcomes may impact survival. Pibarot et al. [15] found that patients with pre-existing severe PPM presented more often elevated mean gradient following TAVI-ViV (47.5% vs. 29.6%, P=0.001). Meanwhile, severe PPM was associated with 2.4-and 1.8-fold higher rates of 30-day mortality and one-year mortality, Investigators confirmed no association between high residual gradient and short-term survival. In our study, we did not find any significant differences in one-, two-, and five-year cumulative and cardiac-related mortality rates between the groups. Our results support the opinion that high residual gradient following TAVI-ViV has no impact on mortality in these high-risk profile patients. In our cohort, postoperative gradient did not significantly affect the postoperative survival, which amounted to 87.7% at one year and half of our patients were octogenarians and five patients were over 90 years old. These findings are also confirmed in a paper prepared by Wernly et al. [16] . They described the results of 223 patients operated in six German heart centers and observed high postinterventional mean gradient in every fourth patient. Residual stenosis did not affect one-year mortality. Bleiziffer et al. [17] analyzed data from the Valve-in-Valve International Data (VIVID) Registry and observed elevated residual gradients in 27.9% of TAVI-ViV patients. was similar to the results presented by other authors (Dvir et al. [13] [83.2%], Wernly et al. [16] [76%], Ihlberg et al. [14] [88.1%], and Webb et al. [9] [87.6%]). Notwithstanding, we observed a significantly higher improvement of NYHA functional classes in patients with lower postprocedural mean gradient (P=0.044).
Another important goal of our study was to identify risk factors for high residual mean gradient following TAVI-ViV. High preoperative gradient, pre-existing PPM, deep valve implantation, small degenerated valves, and older generation of valves were found to be risk factors for high residual gradient in a univariate analysis. During a multivariate analysis, preexisting peak gradient > 60 mmHg, implantation level of 4 mm below neo-annulus, and the degenerated valve size ≤ 23 mm were identified as independent predictors of elevated mean gradient after TAVI-ViV.
Sá et al. [18] performed a meta-analysis of seventeen studies comprising 71,106 patients (PPM n=25,846 patients and non-PPM n=45,260 patients) and found that more than one third of the patients leave the operation room after native transcatheter aortic valve implantation with significant PPM. Severe PPM was a risk factor of one-year mortality. The same group led by Sá [19] prepared a largest meta-analysis of seventy studies (n=108,182 patients) and found that more than half of the patients after conventional aortic valve replacement present significant PPM direct after the procedure. They observed association between severity of PPM and mortality. Pibarot et al. [15] assessed preexisting severe PPM as a risk factor of high post TAVI-ViV gradients (47.6% vs. 29.5%, P=0.001). They observed an occurrence of pre-existing PPM in 7.6% of the patients, which is a similar frequency to ours. In our cohort, seven patients (8.2%) presented pre-existing PPM, one patient died eight months after TAVI-ViV, one patient required re-TAVI-ViV, and two survivors were found in NYHA class III. Due to the small number of patients with severe preexisting PPM, it was difficult to perform a meaningful analysis. Bleiziffer et al. [17] also confirmed a negative impact of PPM on postoperative high gradient after TAVI-ViV.
Scholz et al. [20] focused their study on results after selfexpandable TAVI-ViV depending on degenerated valve size. Small valves (< 23 mm) were associated with significant higher postoperative gradient than bigger valves (22.8 ± 9.4 mmHg vs. 15.1 ± 7.1 mmHg, P=0.013). Simonato et al. [21] used the individuals from the VIVID Registry to study the effect of deep valve implantation on hemodynamical outcomes. They found the strong correlation between deep implantation and better hemodynamics after TAVI-ViV. They defined the optimal implantation level at 0-5 mm below the neo-annulus, which was also noticed in our study. The authors discussed the possibility of aggressive dilatation or even valve fracture, as a novel technique which could ensure the beneficial effect on hemodynamical outcomes. Valve fracture is a new promising technique, which allows receiving the maximum achievable effective orifice area of the new transcatheter valve; however, the currently available evidence is not strong enough to recommend routine valve fracture in TAVI-ViV patients. A large multicenter prospective study evaluating the role of valve fracture on higher postoperative gradients following TAVI-ViV is urgently needed [21][22][23][24][25] .

Limitations
This analysis has several limitations. It is a single-center study with a relatively small sample size. The main limitations of this study are the lack of randomization of the treated groups and the retrospective nature of this investigation. Another limitation is the absence of PPM evaluation following TAVI-ViV and echocardiographic results during the follow-up period. Larger studies will be needed to confirm these results, especially a randomized controlled trial to evaluate the results of intermediate risk patients with degenerated bioprostheses.

CONCLUSION
Although mean transvalvular aortic gradient ≥ 20 mmHg is not rare after TAVI-ViV procedures, it does not affect significantly either early or late mortality. However, elevated mean gradient reduces the improvement of NYHA functional class after TAVI-VIV. Preoperative gradient > 60 mmHg, deep valve implantation, and small size of the surgically implanted bioprosthesis may predispose to increased residual gradient.