If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Despite the rapid adoption of transcatheter aortic valve replacement (TAVR), there are scant data regarding aortic valve reintervention after initial TAVR.
Methods
Between 2011 and 2019, 1487 patients underwent a TAVR at the University of Michigan. Among these, 24 (1.6%) patients required an aortic valve reintervention. Additionally, 4 patients who received a TAVR at another institution underwent a valve reintervention at our institution. We retrospectively reviewed these 28 patients.
Results
The median age was 72 years, 36% were female and 86% of implanted TAVR devices were self-expandable. The leading indications for reintervention were structural valve degeneration (39%) and paravalvular leak (36%). The cumulative incidence of aortic valve reintervention was 4.6% at 8 years. Most (71%) were deemed unsuitable for repeat TAVR because of the need for concurrent cardiac procedures (50%), unfavorable anatomy (45%), or endocarditis (10%). TAVR valve explant was associated with frequent concurrent procedures, consisting of aortic repair (35%), mitral repair/replacement (35%), tricuspid repair (25%), and coronary artery bypass graft (20%). Seventy-one percent of aortic procedures were unplanned but proved necessary because of severe adhesion of the devices to the contacting tissue. There were 3 (15%) in-hospital mortalities in the TAVR valve explant group, whereas there was no mortality in the repeat TAVR group.
Conclusions
Repeat TAVR procedure was frequently not feasible because of unfavorable anatomy and/or the need for concurrent cardiac procedures. Careful assessment of TAVR procedure repeatability should be weighed at the initial TAVR workup especially in younger patients who are expected to require a valve reintervention.
The high proportion of TAVR explant among patients with failing TAVR illustrates that judicious clinical judgement is crucial for appropriately selecting TAVR candidates as an initial valve strategy.
Only 29% of patients with a failed transcatheter aortic valve replacement (TAVR) received a repeat TAVR primarily because of unfavorable anatomy and/or the need for concurrent cardiac procedures. This higher-than-expected proportion of TAVR explant as a valve reintervention illustrates that judicious clinical judgement is crucial when appropriately selecting TAVR candidates as an initial valve strategy.
See Commentaries on pages 1333 and 1335.
Transcatheter aortic valve replacement (TAVR) is an established alternative to surgical aortic valve replacement (SAVR) for patients with severe aortic stenosis and appropriate anatomy, irrespective of the Society of Thoracic Surgeons (STS) Predicted Risk of Mortality score for SAVR, because of a growing body of evidence showing procedural safety and valve performance for currently available TAVR devices.
Outcomes of redo transcatheter aortic valve replacement for the treatment of postprocedural and late occurrence of paravalvular regurgitation and transcatheter valve failure.
With the growth in TAVR usage, however, little is known about the fate of patients with a failed TAVR device. The details for valve reinterventions were not reported in the previous randomized controlled trials, despite the consistently higher rates of reintervention in the TAVR cohort compared with SAVR at 5 years.
5-year outcomes of transcatheter aortic valve replacement or surgical aortic valve replacement for high surgical risk patients with aortic stenosis (PARTNER 1): a randomised controlled trial.
Of particular concern is the lack of details regarding clinical scenarios that required surgical TAVR explantation (TAVR-explant).
We believe when considering an aortic valve replacement approach in patients who are suitable TAVR or SAVR candidates, who will likely outlive the longevity of their TAVR device, exceedingly careful thought is required to determine which patients will be best suited for TAVR or SAVR as the initial procedure to ensure best lifetime management of their aortic valve disease. This study was undertaken to review our experience of aortic valve reintervention after the index TAVR procedure to gain insight to ultimately help refine patient selection for TAVR versus SAVR as the initial valve procedure.
Methods
The University of Michigan institutional review board approved all aspects of the study (HUM00190884; approved on August 6, 2020). The approval included a waiver of informed consent.
Patients and Study Design
We retrospectively reviewed data of 1487 consecutive patients who underwent a TAVR at our institution between April 28, 2011 and August 30, 2019. Of these, 4 patients with intraoperative death and 4 patients with intraoperative conversion from TAVR to SAVR were excluded. This resulted in the final study cohort consisting of 1479 patients. Among these, 24 (1.6%) patients required an aortic valve reintervention. Additionally, 4 patients who received a TAVR procedure at another institution underwent a reintervention at our institution. We retrospectively reviewed the clinical details of these 28 patients. The flow diagram of the patient cohort is summarized in Figure E1. Abstracted data included the following: patient demographic, clinical, and treatment variables, pre-TAVR and pre-reintervention computed tomography angiography (CTA), perioperative and follow-up echocardiographic variables, adverse events, and survival. In-hospital mortality was defined as mortality during the admission after a valve reintervention procedure. Investigators used the National Death Index database, medical record review, and a telephone survey to obtain long-term survival data. Follow-up was complete in 95% of the entire group of TAVR recipients and 100% in patients with valve reintervention as of September 30, 2020.
Evaluation of Repeat TAVR Candidacy
Data of all patients were reviewed by our multidisciplinary structural heart team. Image analysis of CTA with standard TAVR protocol was performed in 19 (68%) patients with TAVR failure. Patients with mandatory TAVR-explant (ie, mechanical structural issues with the TAVR device, aortic aneurysm, endocarditis) or unplanned TAVR-explant during nonaortic valve procedures did not undergo an image analysis. The Vitrea Workstation TAVR planning application (Vital Images, Minnetonka, Minn) and/or 3mensio (Medical Imaging BV, Bilthoven, The Netherlands) were used for the image analysis, including: aortic annulus (circumference, maximal, and minimal diameters), sinus of Valsalva widths and heights, the diameter of the sinotubular junction, and coronary heights (from the annulus to each coronary ostium). The image-based feasibility of repeat TAVR was determined on the basis of aortic root anatomy including sinotubular junction diameter and sinus height relative to the TAVR valve diameter and height. Valve-to-coronary distance
was assessed to determine the risk of coronary obstruction. Valve-to-coronary distance <4 mm was considered high risk for coronary obstruction during repeat TAVR. Aortic roots with unfavorable features for coronary obstruction were classified as follows: (1) sequestered sinus of Valsalva,
defined as an aortic root with equal or lower and narrower sinotubular junction than the height and the diameter of the implanted TAVR device (Figure 1); (2) deficient sinus of Valsalva,
defined as an aortic root with effacement of the sinotubular junction with insufficient valve-to-coronary distance (Figure 2). An image example with favorable anatomy for repeat TAVR is shown in Figure E2 for comparison.
Figure 1Computed tomography angiography (CTA) image of high-risk features (sequestered sinus of Valsalva [SOV]) for coronary obstruction and correlation with intraoperative photographs. A, CTA image showing the left coronary artery (LCA) height, left SOV height, and commissural height within a sequestered SOV, defined as an aortic root with equal or lower and narrower sinotubular junction (STJ) than the height and the diameter of the implanted transcatheter aortic bioprosthesis device. Having an adequate coronary height does not guarantee suitable anatomy for repeat transcatheter aortic valve replacement (repeat-TAVR). B, Intraoperative simulation of the left coronary obstruction if a repeat-TAVR was performed in this patient with a failed self-expandable device within the sequestered root. C, CTA image showing the right coronary height, right SOV height, and device height. D, Intraoperative simulation of right coronary obstruction upon repeat-TAVR in the presence of the sequestered SOV. The Evolut R is from Medtronic Inc (Minneapolis, Minn). RCA, Right coronary artery.
Figure 2Computed tomography angiography (CTA) image of high-risk features for coronary obstruction (deficient sinus of Valsalva [SOV]) and correlation with intraoperative photographs. A, Axial CTA image showing a short valve-to-coronary (VTC) distance. B, The left coronary artery (LCA; arrow) height, left SOV height, and device height within a deficient SOV, defined as an aortic root with effacement of the sinotubular junction (STJ) with insufficient VTC distance. C, Intraoperative photograph of the left coronary ostium (arrow) and a failed balloon-expandable transcatheter aortic bioprosthesis with severe stenosis within a deficient aortic root. SAPIEN valves are from Edwards Lifesciences (Irvine, Calif).
For self-expandable devices, the aortotomy was made at the level of the distal edge of the stent frame. For balloon-expandable devices, a standard aortotomy level was used. Dense endothelial growth of the aorta (neoendothelialization) between the device and the contacting tissue was typically present (Figure E3), requiring careful separation of the aorta, anterior mitral leaflet, and/or membranous septum. The device always came out easily when this was achieved circumferentially while the stent cage of the device was deformed by the 2 Kocher clamps (the double Kocher clamp technique
The procedure was performed in a standardized fashion in a hybrid operating room. After hemodynamic assessment, completion transthoracic echocardiography and aortography were performed to assess the valve function, valve position, and paravalvular leak (PVL). Post deployment balloon valvuloplasty was performed afterward as needed. The grade of PVL was evaluated according to the Valve Academic Research Consortium-2 criteria as none, trace, mild, moderate, and severe.
Continuous variables are expressed as mean ± standard deviation for normally distributed variables and median with interquartile range (IQR) for non-normally distributed variables. Categorical variables are presented as proportion and absolute number. Differences between groups were detected using the χ2 test or Fisher exact test for categorical variables and Student t test or Mann-Whitney U test for continuous variables.
Survival data are depicted using the Kaplan-Meier method and the log rank test with corresponding 95% confidence interval. Because of the significant number of competing events (death) during follow-up and resultant event overestimation rate, the cumulative incidence of valve reintervention was estimated using a competing risks regression using the method of Fine and Gray.
This method uses a semiparametric regression for the longitudinal data in the presence of competing risks, positing a model for the subhazard function of a failure event of primary interest. All P values were the result of 2-tailed tests. The statistical analyses were performed using SPSS 27.0 (IBM Corp, Armonk, NY) and Stata 14.2 (StataCorp, College Station, Tex).
Results
Patient Demographic and Clinical Characteristics
Patient demographic and clinical characteristics are shown in Table 1. Time to reintervention was 1.2 years (IQR, 0.2-3.2), 1.3 years (IQR, 0.2-3.2), and 0.5 years (IQR 0.1-5.4) in the entire, TAVR-explant, and repeat TAVR cohorts, respectively. Fourteen (50%) had previous sternotomies and 36% already had a permanent pacemaker. Eighty-six percent of devices were self-expandable. Among the 28 index TAVRs, 10 (36%) were TAVRs within a surgical bioprosthesis (TAVR-in-SAVR). Most of patients in the TAVR-explant group were in New York Heart Association class IV heart failure at time of surgery.
Defined as (estimated) glomerular filtration rate <60 mL/min/1.73 m2 for 3 months or more.18
16 (57)
11 (55)
5 (63)
Dialysis
2 (7)
2 (10)
0
COPD
8 (29)
4 (20)
4 (50)
Porcelain aorta
4 (14)
3 (15)
1 (13)
History of stroke
5 (18)
3 (15)
2 (25)
Atrial fibrillation
13 (46)
9 (45)
4 (50)
Pulmonary hypertension
9 (32)
7 (35)
2 (25)
Left ventricular ejection fraction, %
35 (26-55)
33 (26-55)
53 (27-66)
Permanent pacemaker
10 (36)
6 (30)
4 (50)
Frailty
14 (50)
8 (40)
6 (75)
Original valve pathology
Degenerative aortic stenosis
22 (79)
15 (75)
7 (88)
Rheumatic
1 (4)
1 (5)
0
Aortic insufficiency
2 (7)
1 (5)
1 (12)
Radiation-associated
2 (7)
2 (10)
0
Congenital subvalvular stenosis
1 (4)
1 (5)
0
Previous sternotomy
14 (50)
10 (50)
4 (50)
TAVR-in-SAVR
10 (36)
8 (40)
2 (25)
Recipient surgical valve
Freestyle
5 (50)
3 (38)
2 (100)
PERIMOUNT
2 (20)
2 (25)
0
Mosaic
2 (20)
2 (25)
0
Homograft
1 (10)
1 (13)
0
Implanted TAVR valve
CoreValve
14 (50)
8 (40)
6 (75)
Evolut R
10 (36)
9 (45)
1 (13)
Sapien
1 (4)
1 (5)
0
Sapien XT
1 (4)
0
1 (13)
Sapien 3
2 (7)
2 (10)
0
Valve size, mm
29 (24-31)
29 (23-34)
29 (26-31)
Risk classification at original TAVR
Extreme risk
3 (11)
2 (10)
1 (13)
High risk
14 (50)
9 (45)
5 (63)
Moderate risk
10 (36)
8 (40)
2 (25)
Low risk
1 (4)
1 (5)
0
STS-PROM at original TAVR
4.4 (2.7-7.4)
3.4 (2.4-6.0)
7.6 (3.9-10.6)
STS-PROM at reintervention
8.5 (5.8-15.3)
9.2 (5.6-15.3)
7.9 (5.9-22.4)
NYHA functional classification at original TAVR
1
0
0
0
2
3 (11)
1 (5)
2 (25)
3
23 (82)
17 (85)
6 (75)
4
2 (7)
2 (10)
0
NYHA Functional Classification at reintervention
1
1 (4)
1 (5)
0
2
0
0
0
3
8 (29)
3 (15)
5 (63)
4
19 (68)
16 (80)
3 (38)
Data are expressed as n (%) or median (IQR), as appropriate. The Freestyle and Mosaic bioprostheses are from Medtronic Inc (Minneapolis, Minn), and the PERIMOUNT is from Edwards Lifesciences (Irvine, Calif). The CoreValve Evolut R system is from Medtronic Inc. SAPIEN valves are from Edwards Lifesciences. TAVR-explant, Surgical transcatheter bioprosthesis explantation; TAVR, transcatheter aortic valve replacement; COPD, chronic obstructive pulmonary disease; TAVR-in-SAVR, TAVR within surgical bioprosthesis; STS-PROM, Society of Thoracic Surgeons Predicted Risk of Mortality; NYHA, New York Heart Association; IQR, interquartile range.
∗ Defined as (estimated) glomerular filtration rate <60 mL/min/1.73 m2 for 3 months or more.
As for TAVR CTA measurements from the original TAVR procedure (Table E1), patients with repeat TAVR had higher right coronary and right sinus of Valsalva heights than patients with TAVR-explant.
Indications for Aortic Valve Reintervention
The clinical indications for aortic valve reintervention are summarized in Table 2. Some patients had more than 1 reason for reoperation. Structural valve degeneration in 11 patients and severe symptomatic PVL in 10 patients were most common. One patient developed a ventricular septal defect after the index TAVR procedure and another had intermittent left main obstruction and coronary ischemia due to device migration. A 72-year-old man with history of homograft aortic root replacement who developed severe aortic insufficiency and simultaneous stroke in the presence of a 6-cm ascending aortic aneurysm underwent TAVR as a bridge to definitive surgical intervention.
Table 2Clinical indications for the aortic valve reintervention procedure
All patients were reviewed by our multidisciplinary structural heart team and 20 (71%) patients were deemed not suitable for a repeat TAVR procedure. Many patients had more than 1 reason for repeat TAVR exclusion including the need for concurrent cardiac procedures and unfavorable anatomy (Figure 3). The details for concurrently performed procedures during TAVR-explant are described in the Operative Data section.
Figure 3Summary of repeat transcatheter aortic valve replacement (TAVR) candidacy exclusion. Many patients had more than 1 reason for redo TAVR exclusion. Therefore, the sum of all percentages is greater than 100%. A, Clinical reasons for the exclusion from repeat TAVR candidacy included the need for concurrent cardiac procedures (50%) and unfavorable anatomy (45%). B, Breakdown of patients with unfavorable redo TAVR anatomy. The risk of coronary obstruction (67%) consisted of 44% sequestered sinus of Valsalva and 22% deficient sinus of Valsalva with insufficient valve-to-coronary distance. Other factors included insufficient annulus size with risk of device constraint (33%) and risk of device migration (33%).
Unfavorable anatomy was attributable to the risk of coronary obstruction (6/9; 67%), specifically from sequestered sinus of Valsalva (Figure 1) and deficient sinus of Valsalva (Figure 2) with insufficient valve-to-coronary distance. Additionally, 3 (33%) patients, who had a TAVR-in-SAVR procedure with a 23-mm self-expandable device, had an inadequate inner annulus size with potential resultant device constraint and/or patient-prosthesis mismatch. With respect to device migration (3/9; 33%), all cases were related to PVL. All index TAVR valves were relatively low-positioned because of large native aortic valve perimeter or TAVR-in-SAVR within a large-size stentless bioprosthesis and resultant retrograde migration.
Operative Data
TAVR-explant
The operative data are summarized in Table 3. Explanted devices comprised: 8 (40%) CoreValve and 9 (45%) Evolut R (both from Medtronic Inc, Minneapolis, Minn); 1 (5%) Sapien and 2 (10%) Sapien 3 (both from Edwards Lifesciences, Irvine, Calif). Implanted surgical valve types were stented/stentless bioprostheses in 80% and mechanical prosthesis in 20% with the median valve size of 25 mm (IQR, 23-27). Of note, 2 (10%) were unplanned TAVR-explant during nonaortic valve procedures (mitral replacement and coronary artery bypass graft [CABG]). Intraoperative transesophageal echocardiography showed a much greater degree of aortic insufficiency/PVL than the findings on the previous surveillance transthoracic echocardiography.
Table 3Operative data in patients with transcatheter aortic bioprosthesis explant (n = 20)
Variable
Value
Redo sternotomy
10 (50)
First time redo
8 (40)
Second time redo
1 (5)
Fifth time redo
1 (5)
Cardiopulmonary bypass time, minutes
202 (152-250)
Aortic cross-clamp time (minutes)
168 (120-199)
Circulatory arrest
1 (5)
Implanted valve
Bioprosthesis
16 (80)
Mechanical prosthesis
4 (20)
Valve size (mm)
25 (23-27)
Concomitant procedure (s)
13 (65)
Root enlargement
2 (10)
Mitral
7 (35)
Tricuspid
5 (25)
Mitral and tricuspid
3 (15)
Aortic repair
7 (35)
Ascending aortic replacement
2 (10)
Partial arch replacement
1 (5)
Full or partial root replacement
6 (30)
CABG
4 (20)
VSD repair
1 (5)
ECMO
2 (10)
IABP
2 (10)
Variables are expressed as n (%) or median (IQR), as appropriate. CABG, Coronary artery bypass graft; VSD, ventricular septal defect; ECMO, extracorporeal membrane oxygenation; IABP, intra-aortic balloon pumping; IQR, interquartile range.
Concurrent procedures were frequent (13/20; 65%), consisting of aortic repair (7/20; 35%), mitral repair or replacement (7/20; 35%), tricuspid repair (5/20; 25%), and CABG (4/20; 20%). Of patients requiring aortic repair, 71% (5 of 7) were unplanned because of tissue trauma during the device explant (Figure E3). These unplanned aortic repair cases included 1 ascending replacement (self-expandable device) and 4 partial or full aortic root replacement (2 self-expandable, 2 balloon-expandable devices). The median age of these devices were 3.4 (IQR, 2.0-5.6) years. As for concurrent mitral procedures, 29% (2/7) were because of mitral impingement by the implanted TAVR device and the rest were severe degenerative pathologies that were originally borderline lesions at the time of original TAVR. There was no tissue damage in the left ventricular outflow tract or anterior mitral leaflet during explantation. Two (10%) patients with preoperative severe pulmonary hypertension required extracorporeal membrane oxygenation support after surgery and 1 died postoperatively.
Repeat TAVR
The procedural data are summarized in Table E2. Implanted devices were mostly self-expandable (7/8; 88%). The mean pressure gradient across the TAVR was 9.8 ± 4.1 mm Hg. Post deployment PVL was common and 75% had at least mild PVL. Three (38%) were moderate or greater. Five (63%) underwent balloon aortic valvuloplasty for PVL (80%) and constrained device (20%). In 1 patient, despite 3 repeat balloon valvuloplasties, significant PVL remained due to extensive annulus calcification.
Postprocedural Outcomes and Survival
The early postprocedural outcomes are summarized in Table E3. There were 3 (15%) in-hospital mortalities in the TAVR-explant group with no mortality in the repeat TAVR group. The cause of death was postoperative end-organ failure in 2 patients. One patient had preprocedural severe biventricular failure with multivessel and left main coronary artery disease in the setting of 2 moderate PVLs. The other patient had severe pulmonary hypertension in the setting of severe TAVR valve stenosis and mixed severe mitral stenosis and regurgitation. The third mortality was due to sudden cardiac arrest in the intensive care unit with unknown cause. There were no strokes in this series. New or worsening renal failure was common in both groups. Among those without a preoperative pacemaker, 1 in each group required permanent pacemaker implantation.
The mean follow-up period after reintervention was 2.1 ± 1.9 years. The estimated 2-year survival was 55 ± 14% and 71 ± 17% (P = .49), respectively, for the TAVR-explant and the repeat TAVR group (Figure E4). There were no patients who required another aortic valve re-reintervention.
Cumulative Incidence of Valve Reintervention
The cumulative incidence of aortic valve reintervention among patients who received a TAVR procedure at our institution was examined. Among the 1479 TAVR recipients, comprising 208 (14%) balloon-expandable and 1271 (86%) self-expandable devices including 161 (11%) TAVR-in-SAVR procedures (within 105 stentless and 56 stented bioprostheses), 24 (1.6%) patients required an aortic valve reintervention. The mean follow-up after index TAVR was 3.0 ± 1.9 years. There have been a total of 639 (43%) deaths which were treated as competing events. The overall cumulative incidence of aortic valve reintervention was 4.6% at 8 years (Figure 4, A).
Figure 4Cumulative incidence of aortic valve reintervention after transcatheter aortic valve replacement (TAVR). The cumulative incidence was estimated using a competing risks regression using the method of Fine and Gray. Deaths were counted as competing events. A, Entire cohort consisting of 1479 TAVR recipients. B, Comparison of TAVR within surgical bioprosthesis (TAVR-in-SAVR; n = 161) versus native valve TAVR (n = 1318). SHR, Subdistribution hazard ratio; CI, confidence interval.
Furthermore, TAVR-in-SAVR procedures resulted in a higher incidence of aortic valve reintervention compared with the native valve TAVR cohort (13.2% vs 3.8%, subdistribution hazard ratio, 3.7; 95% confidence interval, 1.6-8.8; P = .003; Figure 4, B). Among patients with TAVR-in-SAVR failure, 50% were TAVR procedures within a stentless bioprosthesis and PVL or valvular insufficiency was commonly seen (71%). Additionally, 29% demonstrated an associated mitral regurgitation due to low device implantation related to migration.
Discussion
In this report we present our experience with reintervention for TAVR failure since our TAVR program began in 2011. Unlike previous investigations in which the focus was only repeat TAVR feasibility within preselected patients with favorable anatomy after omitting TAVR-explant cases,
Outcomes of redo transcatheter aortic valve replacement for the treatment of postprocedural and late occurrence of paravalvular regurgitation and transcatheter valve failure.
in this study we analyzed all patients with a valve reintervention.
The primary findings in this study were: (1) the cumulative incidence of valve reintervention after index TAVR (14% balloon-expandable and 86% self-expandable devices) was 4.6% over the 8 years; (2) TAVR-in-SAVR was associated with more frequent aortic valve reintervention; (3) the leading indication for reintervention was structural valve degeneration (39%) and PVL (36%); (4) 71% required a TAVR-explant, rather than a repeat TAVR due to the necessity of other concurrent procedure (50%) and/or unfavorable anatomy (45%); (5) 65% required concurrent procedures including 25% unplanned aortic repair because of tissue trauma during TAVR device explantation; (6) greater than mild PVL was very frequent during repeat TAVR procedures and most required balloon aortic valvuloplasty.
Despite the recent global paradigm shift in the aortic valve practice, there are scant data regarding valve reoperation after the initial TAVR procedure. One of the reasons for the paucity of reports is the insufficient number of TAVR recipients who outlived the durability of their TAVR valve. Beyond 5 years, many patients are not alive among high-risk TAVR recipients.
5-year outcomes of transcatheter aortic valve replacement or surgical aortic valve replacement for high surgical risk patients with aortic stenosis (PARTNER 1): a randomised controlled trial.
Furthermore, it is speculated that TAVR failure has been under-reported because most patients without favorable anatomy for repeat TAVR were deemed inoperable. In fact, the small number of post-TAVR patients lost to follow-up in the present study were mostly high-/prohibitive-risk surgical candidates. Because a device failure in “operable” patients remains an extremely rare event at each TAVR center, the only form of reports with respect to post-TAVR reintervention has been large database studies, which are lacking granular information.
The causes of aortic valve reintervention appear distinctly different between TAVR and SAVR. In the Nordic Aortic Valve Intervention (NOTION) trial,
which compared TAVR versus SAVR in 280 low-risk patients with aortic stenosis, the incidence of structural valve degeneration at 5 years in patients treated with TAVR was significantly lower than in patients treated surgically (3.9% vs 26.1%; P < .0001). Despite the striking difference in structural valve degeneration, bioprosthetic valve failure, defined as valve-related death, reintervention or structural valve degeneration, was similar between SAVR and TAVR. Additionally, 21.6% of the TAVR cohort had significant PVL, which was commonly seen in the present study. The frequent PVL in the repeat TAVR group was likely the reason for the relatively short time interval (0.5 years) between the index and repeat TAVR. The 5-year outcomes of The Placement of Aortic Transcatheter Valves (PARTNER) trial for intermediate-risk patients showed a higher reintervention rate in the TAVR cohort (3.2% vs 0.8%) with lack of details for this subset of patients.
The 5-year outcomes of the CoreValve US Pivotal High Risk trial also showed a higher rate of valve reintervention in the TAVR group (3.0% vs 1.1%: P = .04).
Similarly, the details of reintervention procedure and outcomes were lacking in their report. Importantly, these reported reintervention rates were calculated without considering competing risk events (ie, death) during the follow-up. Therefore, the true reintervention rates would have likely been much higher than the reported number if competing events were taken into account. Because of the lack of previous data, the implication of our overall reintervention rate (4.6%) and TAVR-explant rate (71%) remain unclear. We believe the valve reintervention rates are also dictated by site experience. Our program is regarded as a high-volume center (>2500 bioprosthetic SAVR during the same study period) along with the high TAVR volume.
The Clinical Effect of TAVR-Explant
Unlike those clinical trial data, real-world application of TAVR without strict exclusion criteria involves patients with various degrees of synchronous cardiac lesions. Patients with complex coronary disease or multivalvular disease who undergo a TAVR could sacrifice more comprehensive management for the sake of convenience.
Of note, 56% required other concurrent procedures, including aortic repair (26%), mitral (21%), CABG (16%), and/or tricuspid procedures (6%). The 30-day mortality for TAVR-explant in patients with isolated SAVR (n = 345) versus SAVR with concomitant procedure (n = 437) was 14% versus 24% (P = .001), respectively. The present study results were in line with the STS study with respect to the frequency of concurrent procedures. The long aortic cross-clamp time alone (median, 168 minutes) is obviously reflective of the procedure complexity. Additionally, these TAVR-explants were performed by 483 surgeons (median, 1.0 case per surgeon [IQR, 1.0-2.0]) from 313 centers (median, 1.0 case per center [IQR, 1.0-3.0]).
This higher-than-expected mortality rate might be multifactorial. A number of patients in the present study had multiple medical comorbidities and most presented in class III-IV heart failure. On the basis of our experience, there is clearly a technical learning curve for TAVR-explant, whereas the median procedure per surgeon was only 1.0 in the STS series. Thus, the dismal outcomes might reflect the combination of the complexity of the procedure, surgeon's inexperience, and the sickness of these patients.
Appropriateness of TAVR as the First Valve Strategy
Previous TAVR investigations provided neither any insights into the feasibility of repeat TAVR procedures nor the potential clinical scenario of TAVR-explant. In the present study, nearly half of patients with TAVR valve failure had unfavorable anatomic features for repeat TAVR, resulting in a 71% TAVR-explant rate, mostly because of the risk of coronary obstruction. Presently, a high valve implantation during TAVR has been recommended and universally practiced to minimize postprocedural pacemaker implantation.
However, applying the optimized implantation approach during the initial TAVR might prevent patients the opportunity to receive a redo TAVR or retain coronary access caused by a relatively high valve implantation. Balloon-assisted bioprosthetic aortic scallop intentional laceration to prevent coronary artery obstruction might be a viable strategy to increase eligibility for repeat TAVR
Given the results of our study, a major concern is raised as to whether TAVR as the first valve strategy is appropriate in younger patients who inevitably require valve reinterventions at a much older age in the future. More importantly, these low-risk young patients choosing initial TAVR therapy should be informed of the future procedural risks of TAVR-explant which frequently requires morbid device explantation and possible unplanned concurrent procedures. For these reasons, careful assessment of aortic root anatomy and the feasibility of a repeat TAVR procedure should be part of the “initial” TAVR workup for young patients choosing to undergo a TAVR procedure. Otherwise, the mechanical prosthesis option should be considered, considering the cumulative reoperation risk and long-term mortality benefit in this younger population.
Furthermore, although TAVR in previous SAVR bioprosthesis has been a generally accepted safe practice, this present study shows a 3 times higher risk of TAVR failure in the TAVR-in-SAVR versus the TAVR in native valve group (Figure 4, B). This might be explained by an institutional bias, because we perform TAVR-in-SAVR procedures within stentless bioprostheses relatively frequently.
Study Limitation
This study has several inherent limitations including its retrospective nature with a small sample size. In addition to the frequent stentless bioprostheses usage for SAVR, dominant self-expandable device usage in our TAVR practice represents another selection bias. In view of the large number of TAVR devices implanted worldwide, further investigation involving other institutions using standardized methodology is highly warranted.
Conclusions
This study highlighted an important but underappreciated and undescribed consequence of TAVR failure and valve reintervention. The paucity of reports in aortic valve reintervention might be because of the rarity of this clinical scenario, lack of TAVR long-term follow-up, TAVR recipients not outliving the longevity of the TAVR device, under-reporting of these procedures, or a combination of these factors. The anatomic feasibility of repeat TAVR appears limited and it largely remains a theoretical strategy without supporting data. As the role of TAVR for treatment of aortic stenosis has expanded into the lower-risk patient population, a notable increment of TAVR-explant procedures is expected in the next decade, as shown in this study. In this context, more judicious clinical judgement and candidate selection is extremely crucial and the multidisciplinary TAVR team should be mindful of “lifetime management” when planning valve type for the initial aortic valve procedure. Figure 5 shows a graphical summary of the present investigation. Video 1, summarizing the present study, is available.
Figure 5Summary of the present study. Among patients who received a transcatheter aortic valve replacement (TAVR) at University of Michigan between 2011 and 2019, the cumulative incidence of aortic valve reintervention was 4.6% at 8 years. Seventy-one percent required a surgical TAVR valve explant (TAVR-explant) and 29% received a repeat TAVR, resulting in the repeat TAVR to TAVR-explant ratio of 0.4. The TAVR-explant procedure frequently involved complex concurrent procedures and the in-hospital mortality was 15%. Considering the lower-than-expected feasibility of the repeat TAVR procedure and complexity of the TAVR-explant clinical scenario, careful assessment of TAVR procedure repeatability and extremely thoughtful patient selection should be weighed at the initial TAVR workup. CABG, Coronary artery bypass graft.
Dr Deeb was supported by Medtronic Inc as site principal investigator for the Pivotal, Extreme, High, SURTAVI, and Low Risk TAVR trials. Money went to University of Michigan; no personal remuneration was received. Dr Patel serves as a consultant for Medtronic Inc. All other authors reported no conflicts of interest.
The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest.
Figure E2Computed tomography angiography image of favorable anatomy for repeat transcatheter aortic valve replacement (TAVR). A, Computed tomography angiography image of the sinus of Valsalva (SOV) in relation to the left coronary artery (LCA) in the presence of a balloon-expandable TAVR device. There are adequate space and valve-to-coronary distance in the SOV to preserve coronary artery flow for repeat TAVR. B, The SOV in relation to the right coronary artery (RCA) with adequate space for repeat TAVR.
Figure E3An intraoperative photograph of an old self-expandable transcatheter bioprosthesis with denuded aortic intima remnants attached to the stent cage, representing the unique technical difficulty of late transcatheter aortic bioprosthesis explantation.
Figure E4Kaplan-Meier curves for 2-year survival after aortic valve reintervention stratified according to the performed procedure. TAVR-explant, Surgical transcatheter aortic bioprosthesis explantation; TAVR, transcatheter aortic valve replacement; CI, confidence interval.
Table E1Computed tomography angiography measurements of the aortic annulus and aortic root at the time of the original transcatheter aortic valve replacement procedure
Variable
TAVR-explant (n = 20)
Repeat TAVR (n = 8)
P value
Annulus diameter
25.4 ± 4.6
24.1 ± 2.6
.50
Annulus perimeter
83.3 ± 13.5
78.2 ± 8.3
.37
Coronary height: right
14.5 ± 4.3
19.9 ± 6.2
.014
Coronary height: left
13.4 ± 4.0
16.4 ± 5.0
.11
Sinus of Valsalva height: right
20.5 ± 3.3
24.2 ± 4.7
.028
Sinus of Valsalva height: left
20.5 ± 5.3
21.7 ± 3.6
.59
Sinus of Valsalva height: noncoronary
23.0 ± 5.7
23.9 ± 4.1
.70
Sinus of Valsalva width: right
33.7 ± 4.6
32.7 ± 4.6
.68
Sinus of Valsalva width: left
34.8 ± 5.3
33.5 ± 4.9
.59
Sinus of Valsalva width: noncoronary
35.7 ± 6.0
32.7 ± 4.5
.25
Sinotubular junction diameter
33.3 ± 6.3
35.9 ± 4.6
.31
Data are expressed as mean ± standard deviation, except where otherwise noted. Bold indicates P < .05. TAVR-explant, Surgical transcatheter bioprosthesis explantation; TAVR, transcatheter aortic valve replacement.
Table E2Operative data in patients with repeat TAVR (n = 8)
Variable
Value
Implanted valve
Evolut R
3 (38)
CoreValve
3 (38)
Evolut Pro Plus
1 (13)
Sapien 3
1 (13)
Valve size, mm
28 (26-31)
Immediate post deployment paravalvular leak
Trace
2 (26)
Mild
2 (26)
Mild-to-moderate
1 (13)
Moderate
2 (26)
Severe
1 (13)
Post deployment balloon valvuloplasty
5 (63)
Constrained device
1 (13)
Paravalvular leak at completion
None
1 (13)
Trace
2 (26)
Mild
4 (50)
Moderate
1 (13)
Mean transvalvular gradient, mm Hg
9.8 ± 4.1
Data are expressed as n (%), median (interquartile range), or mean ± standard deviation, as appropriate. The Evolut R, CoreValve, and Evolut Pro Plus are from Medtronic Inc (Minneapolis, Minn). The Sapien 3 is from Edwards Lifesciences (Irvine, Calif). TAVR, Transcatheter aortic valve replacement.
Among patients without end-stage renal disease who were receiving dialysis (n = 26). Definition adopted from the Society of Thoracic Surgeons Data Specifications.25
Among patients without end-stage renal disease who were receiving dialysis (n = 26). Definition adopted from the Society of Thoracic Surgeons Data Specifications.25
Among patients without end-stage renal disease who were receiving dialysis (n = 26). Definition adopted from the Society of Thoracic Surgeons Data Specifications.25
Among patients without pacemaker (n = 18) at time of transcatheter valve explant.
Data are expressed as n (%) or median (interquartile range), as appropriate. TAVR-explant, Surgical transcatheter bioprosthesis explantation; TAVR, transcatheter aortic valve replacement; N/A, not applicable; ICU, intensive care unit.
∗ Among patients without end-stage renal disease who were receiving dialysis (n = 26). Definition adopted from the Society of Thoracic Surgeons Data Specifications.
Outcomes of redo transcatheter aortic valve replacement for the treatment of postprocedural and late occurrence of paravalvular regurgitation and transcatheter valve failure.
5-year outcomes of transcatheter aortic valve replacement or surgical aortic valve replacement for high surgical risk patients with aortic stenosis (PARTNER 1): a randomised controlled trial.
With the continued growth of transcatheter aortic valve replacement (TAVR) procedure volume, so too there is an increasing focus on the still-uncertain and evolving treatment algorithm for severe aortic stenosis, especially in younger, lower-risk patients likely to require multiple interventions and devices over their lifetime.1,2 For better or worse, much of the decision making is currently driven by patient choice. Recent reports describing the technical challenges and generally suboptimal outcomes of explanting failed TAVR devices from either native or surgical valves have served to validate growing concerns.
01112-0&id=fx1.jpg){kind=link}
01112-0&id=fx2.jpg){kind=link}
01112-0&id=gr1.jpg){kind=link}
01112-0&id=gr2.jpg){kind=link}
01112-0&id=gr3.jpg){kind=link}
01112-0&id=gr4.jpg){kind=link}
01112-0&id=gr5.jpg){kind=link}
01112-0&id=fx4.jpg){kind=link}
01112-0&id=fx5.jpg){kind=link}
01112-0&id=fx6.jpg){kind=link}
01112-0&id=fx7.jpg){kind=link}
01112-0&id=fx8.jpg){kind=link}