Abstract
Objectives:
Report the largest series of patients receiving a surgical reoperation after transcatheter aortic valve replacement (TAVR) using the Society of Thoracic Surgeons Adult Cardiac Surgery Database (STS ACSD).
Background:
TAVR has become an effective means of treating aortic stenosis. As TAVR is used in progressively lower risk cohorts, management of device failure will become increasingly important.
Methods:
The STS ACSD was queried for patients with a history of prior TAVR undergoing SAVR from 2011–2015. Observed to expected (O/E) mortality ratios were determined to facilitate comparison across re-operative indications and timing from index TAVR procedure.
Results:
123 patients met inclusion criteria (median age 77 years) with an STS predicted mortality rate (PROM) of 4%, 4–8%, and >8% in 17%, 24%, and 59%, respectively. Median time to re-operation was 2.5 months (IQR 0.7–13.0) and the operative mortality rate was 17.1%. Common indications for re-operation included early TAVR device failures such as paravalvular leak (15%), structural prosthetic deterioration (11%), failed repair (11%), sizing/position issues (11%), and prosthetic valve endocarditis (10%). All preoperative risk categories were associated with an increased O/E mortality ratio (PROM <4% - O/E 5.5; 4–8% - O/E 1.7; >8% - O/E 1.2).
Conclusions:
SAVR following acute to subacute failure of TAVR, while rare, is associated with worse than expected outcomes as compared to similar patients initially undergoing SAVR. Continued experience with this developing technology is needed to reduce the incidence of early TAVR failure and further define optimal treatment of failed TAVR prostheses.
Keywords: Transcatheter aortic valve replacement, TAVR, surgical aortic valve replacement, SAVR, device failure
CONDENSED ABSTRACT
Examination of the Society of Thoracic Surgeons Adult Cardiac Surgery Database revealed 123 patients who underwent SAVR following TAVR from 2011–2015. Analysis demonstrated that SAVR following acute to subacute failure of TAVR, while rare, is associated with worse than expected outcomes as compared to similar patients having SAVR as their initial form of valve replacement. Future STS PROM versions should consider incorporating failed TAVR since it clearly increases observed mortality. Continued experience with this developing technology is needed to reduce the incidence of early TAVR failure and further define optimal treatment of failed TAVR prostheses.
INTRODUCTION
Since its introduction in 2002, transcatheter aortic valve replacement (TAVR) has revolutionized the treatment of aortic valve stenosis.1–3 Although rare, device failures have been documented and may involve leak, migration, deterioration, thrombosis, and infection.4,5 Specific treatment of TAVR procedural and device failures depends on the mechanism of failure, but can include “valve-in-valve” replacement with a second TAVR or necessitate surgical aortic valve replacement (SAVR).5
Although initially only utilized in “high-risk” patients who were not candidates for open aortic valve replacement, TAVR is now being performed in younger and lower risk patients.1–3,6,7 Furthermore, the results of the two 2019 low surgical risk trials of TAVR compared with SAVR will undoubtedly lead to a substantial increase in TAVR volume over the coming years.8,9 It is plausible that low- and intermediate-risk TAVR patients may be candidates for SAVR should they develop problems with their TAVR. In addition, patients who were initially high-risk TAVR patients may experience cardiac, renal, and overall functional recovery following transcatheter valve replacement and subsequently become acceptable open surgical candidates if reoperation is needed. However, the outcomes of re-operation for TAVR failure are unknown. We performed a retrospective analysis of a large national clinical database to better understand this patient population and the outcomes of SAVR following TAVR.
METHODS
Data Source
The Society of Thoracic Surgeons Adult Cardiac Surgery Database (STS ACSD) was utilized for this analysis with permission from the STS Longitudinal Follow-Up and Linked Registries (LFLR) committee. The Institutional Review Board of Duke University approved this study prior to data analysis.
Study Cohort
The STS ACSD was queried for patients undergoing SAVR between July 2011 and March 2015 with either a history of prior TAVR or notation of an explanted TAVR device. Where possible, patients with their index TAVR procedure documented before 2006 and those that had greater than 100 months elapse between TAVR and subsequent SAVR procedures were excluded to prevent inclusion of early generation valve types and patients in pre-FDA pivotal trials. These data fields were removed from the database beginning in July 2014 and thus could not be used in patients undergoing SAVR after this time. A total of 205 patients were found to have had a previous TAVR within the STS ACSD during the study period (Figure 1). An additional 28 patients were not listed as having a previous TAVR but had a transcatheter aortic valve device explanted at the time of surgery. Of these 233 patients, 127 underwent a surgical aortic valve replacement. The remaining 106 patients underwent procedures other than surgical aortic valve replacement (n=103, Supplemental Table 1) or had missing procedure information (n=3) and were therefore excluded. An additional 4 patients were excluded due to their prior valve procedure occurring before 2006 or having more than 100 months elapse between their index valve procedure and reoperation, leading to a final study cohort of 123 patients. 30-day mortality was missing for 18 patients and was imputed to alive in these cases.
Figure 1:
Flow diagram illustrating study cohort.
Variable Definitions
Standard STS ACSD definitions were used in this study.10 Indications for operation were determined based on fields in the STS ACSD case reporting form and included structural prosthetic deterioration, prosthetic valve endocarditis, valve thrombosis, failed repair (e.g. operator unsuccessfully attempts to re-expand the TAVR device or achieve leak closure), paravalvular leak, valve entrapment, sizing/position issue, or other. Due to the small number of cases listed in each category, failed repair, paravalvular leak, and sizing/position issues were combined into one group as these were considered similar in mechanism and likely had some degree of overlap. Post-operative outcomes included stroke (neurologic deficit from disturbance of blood supply to brain lasting 24 hours or leading to death), prolonged ventilation (>24 hours), new renal failure (new dialysis requirement, increase in serum creatinine by 150%, decrease in glomerular filtration rate by 25%, or urine output <0.5 cc/kg/hr for 6 hours), new onset atrial fibrillation, new permanent pacemaker/ICD placement, operative mortality (in-hospital or within 30 days), and discharge location.
Statistical Analysis
Patients that met inclusion criteria were then stratified according to STS Predicted Risk of Mortality (STS PROM, 2008 isolated aortic valve replacement model) based on the calculated risk at the time of the SAVR reoperation, using previously established risk categories including those at low risk (STS PROM <4%), intermediate risk (STS PROM 4–8%), and high risk (STS PROM >8%).11 It is anticipated that this expected risk was in many cases lower than the risk at the time of the initial TAVR procedure, owing to some degree of post-TAVR recovery and allowing for inclusion of a modest cohort of low and intermediate risk patients during this analysis interval. Although the STS PROM models do not, by default, categorize patients with prior transcatheter valve replacement as a prior surgery, all patients in this analysis were assigned a risk category based upon undergoing a re-operation. It was anticipated that this assignment would elevate the ‘expected’ risk for these re-operative procedures and would therefore produce a conservative estimate of any additive risk associated with SAVR reoperation following TAVR. The development of this risk model and the variables included have previously been described.12
Baseline characteristics, operative characteristics, and outcomes were described overall and among each risk strata. The time to re-operation was only available for surgical cases performed between July 2011 and June 2014 due to changes in the case reporting form after this period. Outcomes were evaluated both by the indication for reoperation as well as the time between index TAVR and subsequent SAVR procedures.
To calculate the observed to expected (O/E) ratios, the expected mortality rate was based on risk assessment at the time of the re-operation using the STS ACSD predicted risk of operative mortality (PROM) model for isolated valve surgery patients undergoing a reoperation. The observed mortality was divided by this value, resulting in an observed to expected mortality ratio. The O/E ratios were calculated among strata of STS PROM, by re-operative indication, and across categories of the time between the index TAVR operation and current reoperation. Comparisons between O:E ratios were performed using a Wald test by treating the observed number of deaths as a binomial random variable and the expected number of deaths as fixed.13 All analyses were performed using SAS version 9.4 (SAS Institute, Cary NC).
RESULTS
Of the 123 patients who underwent SAVR following TAVR between July 2011 and March 2015, the median age was 77 years (IQR 67–84) and 38% (n=47) were female (Table 1). 21 patients (17%) were categorized as having a low pre-operative risk of mortality (STS PROM <4%), 30 (24%) had an intermediate pre-operative risk of mortality (STS PROM 4–8%), and 72 (59%) had a high pre-operative risk of mortality (STS PROM >8%). The majority of patients (76%, n=94) had symptoms of heart failure within two weeks prior to the operation and 81% (n=76) of these patients were classified as having NYHA class III or IV heart failure. While the majority of procedures among low and intermediate risk patients were performed on an elective basis (76% and 57%, respectively), the majority of procedures among high risk patients were urgent, emergent, or emergent salvage cases (76%).
Table 1:
Baseline and operative characteristics divided by pre-operative risk of mortality.
| Preoperative Risk of Mortality | ||||
|---|---|---|---|---|
| Low (<4%) (N=21) | Intermediate (4–8%) (N=30) | High (>8%) (N=72) | ||
| Baseline Characteristics | ||||
| Age (years) | 77 (67–84) | 64 (42–71) | 77 (72–82) | 80 (72–86) |
| Female Sex | 47 (38.2%) | 6 (28.6%) | 11 (36.7%) | 30 (41.7%) |
| Race | ||||
| Black | 7 (5.7%) | 2 (9.5%) | 1 (3.3%) | 4 (5.6%) |
| Diabetes | 44 (35.8%) | 0 | 9 (30.0%) | 35 (48.6%) |
| Hypertension | 107 (87.0%) | 15 (71.4%) | 25 (83.3%) | 67 (93.1%) |
| Last Creatinine (mg/dL) | 1.2 (0.8–1.5) | 1.0 (0.8–1.2) | 1.1 (0.9–1.3) | 1.3 (0.9–1.8) |
| Dialysis | 8 (6.5%) | 0 | 1 (3.3%) | 7 (9.7%) |
| History of Cerebrovascular Accident | 12 (9.8%) | 1 (4.8%) | 2 (6.8%) | 9 (12.5%) |
| Peripheral Vascular Disease | 39 (31.7%) | 3 (14.3%) | 7 (23.3%) | 29 (40.3%) |
| Body Mass Index (kg/m2) | 27.4 (23.7–32.7) | 29.2 (24.6–32.5) | 27.5 (24.1–32.0) | 26.8 (23.7–33.8) |
| Congestive Heart Failure within 2 weeks | 94 (76.4%) | 8 (38.1%) | 21 (70.0%) | 65 (90.3%) |
| NYHA Class III/IV | 76 (61.8%) | 5 (23.8%) | 14 (46.7%) | 57 (79.2%) |
| Atrial Fibrillation | 41 (33.3%) | 4(19.1%) | 9 (30.0%) | 28 (38.9%) |
| Previous PCI | 39 (31.7%) | 3 (14.3%) | 6 (20.0%) | 30 (41.7%) |
| Previous CABG | 35 (28.5%) | 5 (23.8%) | 9 (30.0%) | 21 (29.2%) |
| Operative Characteristics | ||||
| Status | ||||
| Urgent | 49 (39.8%) | 5 (23.8%) | 13 (43.3%) | 31 (43.1%) |
| Emergent Salvage | 4 (3.3%) | 0 | 0 | 4 (5.6%) |
| Operative Time (min) | 321 (253–412) | 357 (259–411) | 325(288–463) | 313(247–412) |
| CPB Time (min) | 146 (117–198) | 177 (126–237) | 144 (107–198) | 143 (114–185) |
| Cross Clamp Time | 102 (74–132) | 101 (78–176) | 96(70–123) | 102 (74–128) |
| Concomitant Procedures | ||||
| CABG | 7 (5.7%) | 0 | 2 (6.7%) | 5 (6.9%) |
Preoperative risk of mortality as determined by the Society of Thoracic Surgeons predicted risk of operative mortality risk model for isolated aortic valve procedures. Continuous variables are presented as median (interquartile range) while categorical variables are presented as frequency (percentage). CHF – Congestive Heart Failure. NYHA – New York Heart Association. PCI – Percutaneous Coronary Intervention. CABG – Coronary Artery Bypass Grafting. CPB – Cardiopulmonary Bypass.
The most common indications for re-operation (Table 2) were for TAVR valve-related issues including paravalvular leak (n=19, 15%), structural prosthetic deterioration (n=14 ,11%), failed repair (n=13, 11%), sizing or position issues (n=13, 11%), and prosthetic valve endocarditis (n=12, 10%). 3 patients (2%) had a simultaneous mitral valve replacement while 7 patients (6%) had a simultaneous coronary artery bypass graft. Median operative time was 321 min (IQR 253–412), median cardiopulmonary bypass time was 146 min (IQR 117–198), and median cross clamp time was 102 min (IQR 74–132). Among the 58 patients with available data for time between the two procedures (only available for patients from 2011–2014), the median time to the reoperation was 2.5 months (IQR 0.7–13.0, Figure 2).
Table 2:
Indications for surgical aortic valve replacement following transcatheter aortic valve replacement.
| Time to Reoperation* | |||||
|---|---|---|---|---|---|
| <1 Month (n=15) | 1–12 Months (n=24) | >12 Months (n=15) | |||
| Structural Prosthetic Deterioration | 14 (11.4%) | 10 (14.3%) | 0 | 4 (14.8%) | 6 (40.0%) |
| Prosthetic valve Endocarditis | 12 (9.8%) | 6 (10.3%) | 0 | 5 (18.5%) | 1 (6.7%) |
| Valve Thrombosis | 2 (1.6%) | 2 (3.5%) | 0 | 1 (3.7%) | 1 (6.7%) |
| Failed Repair | 13 (10.6%) | 6 (10.3%) | 2 (12.5%) | 3 (11.1%) | 1 (6.7%) |
| Paravalvular Leak | 19 (15.5%) | 10 (17.2%) | 5 (31.3%) | 4 (14.8%) | 1 (6.7%) |
| Valve Entrapment by Pannus, Suture, or Tissue | 2 (1.6%) | 2 (3.5%) | 0 | 1 (3.7%) | 1 (6.7%) |
| Sizing/Position Issue | 13 (10.6%) | 5 (8.6%) | 4 (25.0%) | 1 (3.7%) | 0 |
| Other | 26 (21.1%) | 15 (25.9%) | 4 (25.0%) | 7 (25.9%) | 4 (26.7%) |
| Unknown/Missing | 22 (17.9%) | 2 (3.5%) | 1 (6.3%) | 1 (3.7%) | 0 |
Only patients recorded between July 2011 and June 2014 had time to reoperation data available.
Figure 2:
Cumulative probability plot demonstrating variation in timing between TAVR and SAVR, among the 52 patients with this data available.
21 patients (17%) died within 30 days of the procedure or prior to discharge from the hospital (defined as operative mortality, Table 3 & Supplemental Table 2), including 14%, 10%, and 21% of patients with a low, intermediate, and high preoperative risk of mortality, respectively. Operative mortality was higher among patients requiring reoperation for endocarditis (25%) or a perivalvular leak, sizing or position problem, or failed repair (24%) as compared to patients with prosthetic deterioration (15%). 17 patients (14%) required another reoperation following the SAVR intervention within the same hospitalization and 50 (41%) required greater than 24 hours of ventilation. New renal failure occurred in 12 patients (10%), 4 (3%) patients experienced permanent stroke, and 18 (15%) required new permanent pacemaker/ICD placement.
Table 3:
Outcomes divided by pre-operative risk of mortality, reason for reoperation, and timing of repair.
| Preoperative Risk of Mortality | Indication for Reoperation | Timing to Reoperation* | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Low (<4%) (N=21) | Medium (4–8%) (N=30) | High (>8%) (N=72) | Prosthetic Deterioration (N=20) | Endocarditis (N=12) | Perivalvular Leak, Sizing/Position, or Failed Repair (N=45) | <1 Month (N=15) | 1–12 Months (N=24) | >12 Months (N=15) | ||
| Operative Mortality | 21 (17.1%) | 3 (14.3%) | 3 (10.0%) | 15 (20.8%) | 3 (15.0%) | 3 (25.0%) | 11 (24.4%) | 2 (12.5%) | 4(14.8%) | 3 (20.0%) |
| Discharge Location | ||||||||||
| Home | 44 (43.1%) | 16 (88.9%) | 13 (48.2%) | 15 (26.3%) | 11 (64.7%) | 3 (33.3%) | 13 (38.2%) | 6 (42.9%) | 11 (47.8%) | 9 (75.0%) |
| Extended Care/TCU | 46 (45.1%) | 2 (11.1%) | 11 (40.7%) | 33 (57.9%) | 5 (29.4%) | 3 (33.3%) | 17 (50.0%) | 7 (50.0%) | 9(39.1%) | 2 (16.7%) |
| Nursing Home | 4 (3.9%) | 0 | 2 (7.4%) | 2 (3.5%) | 0 | 1 (11.1%) | 2 (5.9%) | 0 | 1 (4.4%) | 0 |
| Other Hospital | 7 (6.9%) | 0 | 1 (3.7%) | 6(10.5%) | 1 (5.9%) | 2 (22.2%) | 1 (2.9%) | 1 (7.1%) | 2 (8.7%) | 1 (8.3%) |
| Stroke | 4 (3.3%) | 1 (4.8%) | 0 | 3 (4.2%) | 1 (5.0%) | 0 | 1 (2.2%) | 0 | 0 | 0 |
| Reoperation | 17 (13.8%) | 5 (23.8%) | 4 (13.3%) | 8(11.1%) | 2 (10.0%) | 1 (8.3%) | 10 (22.2%) | 2 (12.5%) | 2 (7.4%) | 0 |
| Prolonged Ventilation† | 50 (40.7%) | 4 (19.1%) | 8 (26.7%) | 38 (52.8%) | 7 (35.0%) | 6 (50%) | 21 (46.7%) | 8 (50.0%) | 9 (33.3%) | 6 (40.0%) |
| New Renal Failure | 12 (10.4%) | 1 (4.8%) | 3 (10.3%) | 8(12.3%) | 2 (11.1%) | 2 (22.2%) | 6(14.0%) | 0 | 2 (8.7%) | 2 (14.3%) |
| New Atrial Fibrillation | 29 (35.4%) | 5 (29.4%) | 8 (38.1%) | 16 (36.4%) | 5 (31.3%) | 1 (12.5%) | 11 (39.3%) | 3 (33.3%) | 6 (46.2%) | 2 (14.3%) |
| New Permanent Pacemaker/ICD | 18 (14.6%) | 5 (23.8%) | 4 (13.3%) | 9 (12.5%) | 4 (20.0%) | 3 (25.0%) | 4 (8.9%) | 3 (18.8%) | 5(18.5%) | 3 (20.0%) |
July 2011-June 2014 only
Prolonged Ventilation – Greater than 24 hours. TCU – Transitional Care Unit
Only 43% (n=44) of patients were discharged home following the procedure (Table 3). The vast majority of patients discharged to a location other than home were discharged to an extended care facility or transitional care unit (n=46, 45%),
The observed mortality was higher than expected across the spectrum of preoperative risk, including among those at low risk (O/E ratio: 5.48, 95% CI: 1.17–13.93), intermediate risk (O/E ratio: 1.66, 95% CI: 0.35–4.40), and high risk (O/E ratio: 1.16, 95% CI: 0.68–1.79, Table 4). Compared with the low risk group, O/E ratios were lower for the high risk group (p<0.01) and similar for the intermediate risk group (p=0.12).
Table 4:
Observed to expected ratios for mortality by preoperative risk, indications for operation, and timing between the transcatheter aortic valve replacement and the surgical aortic valve replacement.
| Group | N (total) | Observed Mortality | Expected Mortality | O/E Ratio (95% Cl) | P-Value |
|---|---|---|---|---|---|
| STS PROM Assuming All Patients Underwent Reoperation | |||||
| Low Risk (<4%) | 21 | 14.3% | 2.6% | 5.48 (1.17–13.93) | Ref |
| Intermediate Risk (4–8%) | 30 | 10.0% | 6.0% | 1.66 (0.35–4.40) | 0.119 |
| High Risk (>8%) | 72 | 20.8% | 17.9% | 1.16 (0.68–1.79) | 0.008 |
| Indication for Reoperation | |||||
| Prosthetic Valve Deterioration | 20 | 15.0% | 11.2% | 1.35 (0.29–3.40) | Ref |
| Endocarditis | 12 | 25.0% | 20.1% | 1.24 (0.27–2.84) | 0.915 |
| Perivalvular Leak, Sizing Issue, Failed Repair | 45 | 24.4% | 11.3% | 2.16(1.14–3.50) | 0.423 |
| Timing to Reoperation* | |||||
| Less 1 month | 15 | 12.5% | 12.2% | 1.03 (0.13–3.16) | Ref |
| 1–12 months | 24 | 14.8% | 13.5% | 1.10 (0.31–2.49) | 0.939 |
| Greater than 12 months | 15 | 20.0% | 11.1% | 1.80 (0.39–4.32) | 0.506 |
STS PROM – Society of Thoracic Surgeons Predicted Risk of Mortality. O/E – Observed to Expected Ratio. CI – Confidence Interval.
July 2011-June 2014 only
When investigating observed to expected mortality ratios by indication for reoperation (Table 4), patients undergoing reoperation for all indications had higher than expected mortality rates (O/E ratio > 1). However, there were no significant differences in O/E ratios when comparing indications for reoperation (p>0.05). Similarly, there were no significant differences in O/E ratios when examining differences by time between TAVR and the reoperation (1–12 months vs <1 month p=0.94; >12 months vs <1 month p=0.51), although there was a trend associated with increasing time to reoperation and increasing O/E ratios (<1 month O/E 1.03; 1–12 months O/E 1.10; >12 months O/E 1.80).
DISCUSSION
The introduction of TAVR in 2002 revolutionized the management of aortic valve stenosis.14 With the publication of results from the latest TAVR trials among low risk patients in 2019, TAVR has now been shown to be noninferior, and in some studies superior, to SAVR with respect to mortality or disabling stroke across the spectrum of preoperative risk.1,8,9,15–17 Furthermore, compared with SAVR, TAVR has been associated with an improved complication profile and a shorter length of hospital stay.18,19 However, mechanisms and outcomes of TAVR failure are incompletely understood. In this study, we present the US experience of SAVR following failed TAVR, as documented in the STS ACSD database. Paravalvular leak, failed repair, and structural prosthetic deterioration were among the most common reasons for surgical re-intervention, and operative mortality was greater than the expected rate across the spectrum of patient risk. Further research is needed to refine re-operative techniques and share best practices when explanting failed TAVR devices.
In this study, we identified 123 patients who underwent SAVR following TAVR, out of approximately 40,000 TAVR procedures performed in the US during the study period.20 This represents approximately 0.3% of TAVR patients, which is likely a slight overestimate given the inability to account for TAVRs performed as part of clinical trials. These patients represent TAVR failure as predominantly an early failure—either acute or subacute occurring immediately, months, or within a few years from implantation. While TAVR failure is relatively rare, it is expected that issues related to SAVR after failed TAVR will become more common for two reasons. First, TAVR was initially approved in the US in 2011 and not until 2015 did the number of patients receiving TAVR exceed 25,000 per year, and after 2017 TAVR volume has increased to over 50,000 patients a year.20 Most patients in this study did not have structural deterioration of the valve and those who did had early failure. Second, the management of structural deterioration of TAVR occurring after 5 or more years will be another future aspect of defining the role of SAVR versus TAVR valve-in-valve, especially in patients receiving TAVR at an age of less than 75 and those who are low risk for SAVR at the time of their initial TAVR.
Although the majority (59%) of patients included in this study were categorized as having a high preoperative risk of mortality (>8%), only patients deemed to be operative candidates would have undergone a SAVR reoperation. Therefore, this dataset is a unique mix of patients who were considered either intermediate or high-risk at the time of the initial TAVR procedure, but still low enough risk that they were candidates for a SAVR at the time of TAVR failure. Furthermore, patients who were initially considered high-risk at the time of their TAVR procedure likely experienced subsequent recovery of physiologic reserve, thus accounting for the cohort of patients who were considered low or intermediate risk at the time of their SAVR reoperation. As an increasing number of low risk patients receive a TAVR, the rate of patients requiring SAVR for a failed TAVR will increase, thus making the appropriate risk assessment of patients undergoing SAVR after TAVR failure all the more important.
The median operative time for these patients of 321 minutes is substantially longer than the median operative time around 200 minutes reported in published series of surgical aortic valve replacements.21 Although the reason for this additional 2 hours of operative time cannot be gleaned from the STS ACSD, it may be related to technical challenges associated with performing redo sternotomies, as approximately one quarter of patients had undergone prior CABG, or challenges associated with removal of the previous TAVR device. The median cardiopulmonary bypass time of 146 minutes in this study is still significantly longer than the 111 minute median bypass time demonstrated in a prior STS ACSD study of AVR procedures in patients with a previous CABG, however.22 Prior pathologic studies have demonstrated progressive neointimal growth occurring in TAVR valves over 90 days following implantation.23 While only approximately 7% of patients in this series required root replacement, this continuous neointimal growth in conjunction with difficulties associated with extracting the TAVR valve may have led to the longer operative times observed. Others have suggested that there is added complexity with extraction of the nitinol self-expanding cage of the CoreValve device.24 Unfortunately, the STS ACSD is not designed to reliably capture device model information, and these data were not available for this analysis. In a recently published single-center retrospective analysis of 15 SAVR after TAVR cases, Fukuhara and colleagues from the University of Michigan noted that neoendothelialization resulted in extremely challenging cases for transcatheter prostheses implanted greater than 1 year prior.25 Further, both balloon expandable and self-expanding prosthesis required extensive endarterectomy to safely remove the devices. Thus, future research optimizing surgical techniques will be vital, especially as we see a greater number of young patients experiencing long-term structural valvular degeneration.
Additionally, postoperative morbidity occurred more commonly in this cohort of patients than among patients undergoing conventional SAVR. Thourani and colleagues reported on the outcomes of 141,905 patients undergoing isolated SAVR, stratifying patients into a similar low, medium, and high PROM groups.15 The incidence of stroke, new renal failure, new atrial fibrillation, prolonged ventilation, and operative mortality were all substantially more common in this cohort than among conventional SAVR patients from their study, even when compared only to the high risk cohort (PROM >8%).15 This is likely secondary to the greater complexity associated with SAVR after TAVR procedures, including longer operative times, as well as variations in the patient baseline population not accounted for in current PROM models. The latter is supported by the finding that morbidity rates varied by indication for procedure as well as timing from TAVR. Patients receiving a SAVR due to endocarditis had a post-SAVR reoperation rate of 8% compared to 10% for prosthetic deterioration and 22% for perivalvular leak, sizing/position, or failed repair. Furthermore, new renal failure was twice as common in the endocarditis group, likely due to the infectious etiology and systemic inflammatory response present in these patients as well as the nephrotoxic effect of prolonged antibiotic usage. Operative mortality was similarly more common in the endocarditis group as well as the perivalvular leak, sizing/position, or failed repair group than it was in the prosthetic deterioration group.
The approximately 17% operative mortality rate observed in this study is higher than would be expected in a similar population of patients undergoing repeat SAVR. While not directly comparable due to differences in baseline risk, Kaneko and colleagues reported a 4.6% overall operative mortality rate for patients undergoing isolated redo SAVR from 2011–2013.26 Similarly, this operative mortality rate is much higher than would be expected in TAVR valve-in-valve (VIV) procedures.27 While the STS risk model used to determine O/E ratios has been demonstrated to have a c-index of 0.799 for operative mortality, we have shown that this model under-estimates the risk of patients who had previously undergone a TAVR.12 As an increasing number of low-risk patients undergo TAVR, and therefore will be candidates for a SAVR should their TAVR fail, it may be necessary to incorporate previous TAVR in future STS PROM models.
This is not the first study to investigate or describe valve deterioration issues following TAVR. Del Trigo and colleagues analyzed 1,521 patients in a multicenter registry and found a 4.5% rate (n=68) of valve deterioration defined as an absolute change in transvalvular gradient of 10 mm Hg or greater, 2.8% of which occurred within the first year.28 In the five year analysis of the PARTNER trial, Mack and colleagues compared valve hemodynamics between TAVR and SAVR and found no significant difference in mean aortic valve areas or mean valve gradients. Furthermore, there were no valve deteriorations requiring surgical replacement; however, in these trials with early generation TAVR prostheses, TAVR patients did have significantly higher rates of moderate or severe aortic regurgitation mainly secondary to paravalvular regurgitation, and this was associated with increased long-term mortality.29
There are several limitations to this study that should be taken into account when interpreting our findings. First, as discussed above, this study included a select set of patients who were high enough perioperative risk to undergo TAVR, but who were candidates for an open SAVR at a later date. We assigned risk categories based on operative risk estimated at the time of SAVR reoperation, which likely reflects some degree of post-TAVR recovery. As a result, nearly half of the study population were considered ‘low’ or ‘intermediate’ risk prior to SAVR and thus these findings are applicable to future lower risk TAVR cohorts. We have shown that previous TAVR is not fully adjusted for in current STS PROM models. This finding has important implications when assessing preoperative risk and planning timing of reoperation following TAVR, including early intervention for perivalvular leak. Furthermore, these data have implications for conversations between providers and their patients, including obtaining appropriate informed consent through shared decision making, especially since current models underestimate the true risk of SAVR following TAVR. Second, valve-in-valve (VIV) implantation is a lower risk method for treating TAVR deterioration, and may be a better alternative than open SAVR for TAVR valve deterioration.30 Unfortunately this is not a valid option for some indications for reoperation such as endocarditis or more severe perivalvular leak. Third, this study was limited by a relatively small sample size, particularly with regard to the analysis of operative timing, and was thus underpowered to detect a difference in observed to expected mortality ratios between various reoperation time points. However, a trend toward increasing risk with longer time from TAVR implantation does suggest increased complexity of these cases. Furthermore, our analysis was restricted to patients who underwent TAVR procedures in the 5 years following FDA approval and as such, may not be completely generalizable as surgeons gain increased experience with SAVR following TAVR. Fourth, data regarding explanted TAVR valve type was not of high enough fidelity in the database to include in the analysis, preventing examination of the association between valve type and reoperation. Finally, as a deidentified database, the granularity that would be present in a single center study is not available, and therefore detailed descriptions of all factors influencing decision making regarding operative timing are not obtainable. Similarly, inaccuracies in coding may have prevented some patients who underwent SAVR after TAVR from being included in our cohort.
In conclusion, although TAVR has been demonstrated to be a suitable option for many patients requiring aortic valve replacement, both early and late prosthesis-related issues may require surgical reoperation. In this study, which is the largest series of patients requiring SAVR following TAVR reported, we have demonstrated that SAVR following TAVR is associated with worse than expected outcomes – possibly related to the added complexity of removing a well incorporated TAVR prosthesis. Continued experience with this developing technology will help to further define optimal treatment options in the setting of failed TAVR prostheses.
Supplementary Material
Central illustration:
Flow diagram illustrating study cohort (A) and Cumulative probability plot demonstrating variation in timing between TAVR and SAVR, among the 52 patients with this data available (B)
PERSPECTIVES.
WHAT IS KNOWN?
The use of TAVR is expected to continue to increase dramatically over the coming few years. As TAVR is used in progressively lower risk cohorts, management of device failure will become increasingly important.
WHAT IS NEW?
Here, we examine the largest series to date of patients requiring SAVR following TAVR and demonstrate worse than expected outcomes among this cohort. These findings may be the result of examining a population of patients with high baseline risk, considering they all underwent TAVR initially before the low risk TAVR trials, as well as possibly the result of added complexity associated with removing a well incorporated TAVR prosthesis.
WHAT IS NEXT?
Continued experience with this developing technology will help to further define optimal treatment options in the setting of failed TAVR prostheses. Further, studies examining management of failed TAVR prostheses among young patients experiencing long-term structural valvular degeneration will be vital.
ACKNOWLEDGEMENT
The Society of Thoracic Surgeons (STS) National Database provided the data for this research. Analysis funding was provided by the DCRI and an FDA U01 grant (PI: Brennan). Dr. Jawitz received funding from National Institutes of Health grant 5T32HL069749.
Funding:
The Society of Thoracic Surgeons (STS) National Database provided the data for this research. Dr. Jawitz received funding provided by NIH T-32 grant 5T32HL069749
ABBREVIATIONS AND ACRONYMS
- SAVR
surgical aortic valve replacement
- TAVR
transcatheter aortic valve replacement
- STS
Society for Thoracic Surgeons
- ACSD
Adult Cardiac Surgery Database
- FDA
Food & Drug Administration
- CABG
coronary artery bypass grafting
- PROM
Predicted Risk of Mortality
- O/E
observed to expected ratio
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosures:
none
Tweet/handle: @ojawitzMD ; In the largest series to date of patients requiring SAVR following TAVR, worse than expected outcomes were demonstrated
REFERENCES
- 1.Thourani VH, Kodali S, Makkar RR, et al. Transcatheter aortic valve replacement versus surgical valve replacement in intermediate-risk patients: a propensity score analysis. Lancet. 2016;387(10034):2218–2225. [DOI] [PubMed] [Google Scholar]
- 2.Leon MB, Smith CR, Mack M, et al. Transcatheter Aortic-Valve Implantation for Aortic Stenosis in Patients Who Cannot Undergo Surgery. New England Journal of Medicine. 2010;363(17):1597–1607. [DOI] [PubMed] [Google Scholar]
- 3.Leon MB, Smith CR, Mack MJ, et al. Transcatheter or Surgical Aortic-Valve Replacement in Intermediate-Risk Patients. New England Journal of Medicine. 2016;374(17):1609–1620. [DOI] [PubMed] [Google Scholar]
- 4.Latib A, Naim C, De Bonis M, et al. TAVR-associated prosthetic valve infective endocarditis: results of a large, multicenter registry. Journal of the American College of Cardiology. 2014;64(20):2176–2178. [DOI] [PubMed] [Google Scholar]
- 5.Mylotte D, Piazza N. Transcatheter aortic valve replacement failure: deja vu ou jamais vu? Circulation Cardiovascular interventions. 2015;8(4). [DOI] [PubMed] [Google Scholar]
- 6.Gleason TG, Reardon MJ, Popma JJ, et al. 5-Year Outcomes of Self-Expanding Transcatheter Versus Surgical Aortic Valve Replacement in High-Risk Patients. Journal of the American College of Cardiology. 2018;72(22):2687–2696. [DOI] [PubMed] [Google Scholar]
- 7.Reardon MJ, Van Mieghem NM, Popma JJ, et al. Surgical or Transcatheter Aortic-Valve Replacement in Intermediate-Risk Patients. The New England journal of medicine. 2017;376(14):1321–1331. [DOI] [PubMed] [Google Scholar]
- 8.Mack MJ, Leon MB, Thourani VH, et al. Transcatheter Aortic-Valve Replacement with a Balloon-Expandable Valve in Low-Risk Patients. N Engl J Med. 2019;380(18):1695–1705. [DOI] [PubMed] [Google Scholar]
- 9.Popma JJ, Deeb GM, Yakubov SJ, et al. Transcatheter Aortic-Valve Replacement with a Self-Expanding Valve in Low-Risk Patients. N Engl J Med. 2019;380(18):1706–1715. [DOI] [PubMed] [Google Scholar]
- 10.Adult Cardiac Surgery Database Data Collection Forms. https://www.sts.org/registries-research-center/sts-national-database/adult-cardiac-surgery-database/data-collection. Accessed on February 10, 2019.
- 11.Brennan JM, Holmes DR, Sherwood MW, et al. The association of transcatheter aortic valve replacement availability and hospital aortic valve replacement volume and mortality in the United States. The Annals of thoracic surgery. 2014;98(6):2016–2022; discussion 2022. [DOI] [PubMed] [Google Scholar]
- 12.O’Brien SM, Shahian DM, Filardo G, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2--isolated valve surgery. The Annals of thoracic surgery. 2009;88(1 Suppl):S23–42. [DOI] [PubMed] [Google Scholar]
- 13.Winston CA, Hill AN, Chen MP, Shang N, Becerra JE. Confidence intervals and statistical testing for ratio measures of percent change. Stat Med. 2012;31(27):3295–3298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cribier A, Eltchaninoff H, Bash A, et al. Percutaneous transcatheter implantation of an aortic valve prosthesis for calcific aortic stenosis: first human case description. Circulation. 2002;106(24):3006–3008. [DOI] [PubMed] [Google Scholar]
- 15.Thourani VH, Suri RM, Gunter RL, et al. Contemporary real-world outcomes of surgical aortic valve replacement in 141,905 low-risk, intermediate-risk, and high-risk patients. The Annals of thoracic surgery. 2015;99(1):55–61. [DOI] [PubMed] [Google Scholar]
- 16.Thyregod HGH, Steinbrüchel DA, Ihlemann N, et al. Transcatheter versus surgical aortic valve replacement in patients with severe aortic valve stenosis: 1-year results from the all-comers NOTION randomized clinical trial. Journal of the American College of Cardiology. 2015;65(20):2184–2194. [DOI] [PubMed] [Google Scholar]
- 17.Smith CR, Leon MB, Mack MJ, et al. Transcatheter versus surgical aortic-valve replacement in high-risk patients. N Engl J Med. 2011;364(23):2187–2198. [DOI] [PubMed] [Google Scholar]
- 18.Arora S, Strassle PD, Kolte D, et al. Length of Stay and Discharge Disposition After Transcatheter Versus Surgical Aortic Valve Replacement in the United States. Circ Cardiovasc Interv. 2018;11(9):e006929. [DOI] [PubMed] [Google Scholar]
- 19.Siemieniuk RA, Agoritsas T, Manja V, et al. Transcatheter versus surgical aortic valve replacement in patients with severe aortic stenosis at low and intermediate risk: systematic review and meta-analysis. BMJ. 2016;354:i5130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Carroll J Personal comunication of data presented at the STS-ACC TVT Registry Stakeholders meeting on March 4, 2019 in Washington DC. In. [Google Scholar]
- 21.Dogan S, Dzemali O, Wimmer-Greinecker G, et al. Minimally invasive versus conventional aortic valve replacement: a prospective randomized trial. JOURNAL OF HEART VALVE DISEASE. 2003;12(1):76–80. [PubMed] [Google Scholar]
- 22.Vassileva CM, Aranki S, Brennan JM, et al. Evaluation of The Society of Thoracic Surgeons Online Risk Calculator for Assessment of Risk in Patients Presenting for Aortic Valve Replacement After Prior Coronary Artery Bypass Graft: An Analysis Using the STS Adult Cardiac Surgery Database. Ann Thorac Surg. 2015;100(6):2109–2115; discussion 2115–2106. [DOI] [PubMed] [Google Scholar]
- 23.Yahagi K, Torii S, Ladich E, et al. Pathology of self-expanding transcatheter aortic valves: Findings from the CoreValve US pivotal trials. Catheter Cardiovasc Interv. 2018;91(5):947–955. [DOI] [PubMed] [Google Scholar]
- 24.Mangi AA, Ramchandani M, Reardon M. Surgical Removal and Replacement of Chronically Implanted Transcatheter Aortic Prostheses: How I Teach It. Ann Thorac Surg. 2018;105(1):12–14. [DOI] [PubMed] [Google Scholar]
- 25.Fukuhara S, Brescia AA, Shiomi S, et al. Surgical explantation of transcatheter aortic bioprostheses: Results and clinical implications. J Thorac Cardiovasc Surg. 2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kaneko T, Vassileva CM, Englum B, et al. Contemporary Outcomes of Repeat Aortic Valve Replacement: A Benchmark for Transcatheter Valve-in-Valve Procedures. Ann Thorac Surg. 2015;100(4):1298–1304; discussion 1304. [DOI] [PubMed] [Google Scholar]
- 27.Dvir D, Webb JG, Bleiziffer S, et al. Transcatheter aortic valve implantation in failed bioprosthetic surgical valves. JAMA. 2014;312(2):162–170. [DOI] [PubMed] [Google Scholar]
- 28.Del Trigo M, Munoz-Garcia AJ, Wijeysundera HC, et al. Incidence, Timing, and Predictors of Valve Hemodynamic Deterioration After Transcatheter Aortic Valve Replacement: Multicenter Registry. Journal of the American College of Cardiology. 2016;67(6):644–655. [DOI] [PubMed] [Google Scholar]
- 29.Mack MJ, Leon MB, Smith CR, et al. 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. Lancet. 2015;385(9986):2477–2484. [DOI] [PubMed] [Google Scholar]
- 30.Dvir D, Webb J, Brecker S, et al. Transcatheter aortic valve replacement for degenerative bioprosthetic surgical valves: results from the global valve-in-valve registry. Circulation. 2012:CIRCULATIONAHA 112.104505. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.