Introduction

This state-of-the–transcatheter aortic valve replacement (TAVR) in the United States report presents the data submitted to the ACC-STS TVT Registry (Society of Thoracic Surgeons [STS]–American College of Cardiology [ACC] Transcatheter Valve Therapy Registry), from 276,316 patients who have undergone TAVR in the United States from 2011 to 2019. The Registry has been the national repository of data on new transcatheter valve therapies, including a TAVR module, to capture all procedures that use U.S. Food and Drug Administration (FDA)-approved TAVR devices (1,2). The first TAVR device approval occurred in late 2011 for patients with symptomatic, severe aortic stenosis deemed at extreme risk for surgical aortic valve replacement (SAVR). Subsequently, label expansion and new device approvals have extended access to TAVR to patients deemed to be at high risk (2012), intermediate risk (2016), and low risk (2019) for SAVR, as well as to patients with degenerated surgically implanted aortic tissue valves (3). The TAVR National Coverage Determination (NCD) from the Centers for Medicare & Medicaid Services (CMS), first released in 2012 and updated in 2019, links reimbursement to submission of patient-level data on all patients receiving commercially approved TAVR to a qualified national registry with the goal of gathering additional evidence on this transformative therapy (4). The Registry has been approved for this purpose ( NCT01737528) and has been operational since 2011 (5). The timelines for FDA approvals of TAVR, CMS release of TAVR NCDs, and relevant Registry events are presented in Table 1.

Methods

Data source

Using the Registry database, we report the year-by-year data from hospital sites performing TAVR, including procedure volumes, patient characteristics, procedure characteristics, and outcomes (Table 2). The data collection form (DCF) and data definitions for the TAVR module are available at the Registry’s website (6). Deadlines for submission of data from each calendar quarter are communicated to sites, but sites occasionally modify and submit data after deadlines, and this results in small changes in the subsequent summary statistics. The coronavirus disease 2019 (COVID-19) pandemic has impaired submission of data from some sites in 2020.

Patient-level data for the baseline and in-hospital time point are available for all patients treated through the end of 2019. Thirty-day outcome data are available for patients treated on/before September 30, 2019. All 1-year outcomes are for patients treated on/before September 30, 2018. One-year follow-up also uses data links to the CMS database with CMS performing the linkage using direct patient identifiers. Complete 1-year post-TAVR data using the CMS linkage are available for Medicare fee-for-service patients treated through calendar year 2017. Partial 2018 1-year post-TAVR data using the CMS linkage are available for patients treated on/before March 31, 2018. More recent CMS-linked data are currently not available.

The Registry has approval from a central institutional review board (IRB, Advarra, Columbia, Maryland), and sites have received a waiver of informed consent under the Common Rule based on the IRB finding that the Registry constitutes a minimal risk to the patient. Because the societies, as sponsors of the Registry, have IRB approval and a waiver of informed consent, and because the data are already routinely collected, individual sites participating in the Registry do not need to obtain local IRB approval before enrolling in the Registry.

Data quality

Data quality was assessed at 3 stages. First, confidential Data Quality Reports are sent to all sites after their quarterly data submission. The Data Quality Reports inform sites whether their data pass rigorous electronic quality checks on completeness. Second, all data are subsequently transferred to Duke Clinical Research Institute (DCRI) where additional quality checks occur. Clinicians at DCRI independently adjudicate stroke, transient ischemic attack, and repeat aortic valve intervention using source documentation. Third, yearly audits are performed in conjunction with a Quality Innovation Network Quality Improvement Organization contractor (7). Every year, approximately 10% of randomly selected sites are audited by nurse auditors trained by the ACC. The data included in the results section, tables, and figures of this report have been reviewed and finalized for analysis by the DCRI.

Data presentation

The tables and figures contain data from all TAVR procedures performed at sites active through 2019. Data elements collected during the pre-procedure evaluation, procedure-related hospitalization, and follow-up at 30 days and 1 year are reported from all patients with the time of the most recent results as clarified in the preceding text. Data results presented in the text are median values with interquartile ranges (IQR) in parentheses.

Trends for many data elements are presented graphically with year-to-year comparisons. Other trends are presented as changes from early TAVR experience (defined as patients treated from late 2011 up until the end of 2013) compared with current TAVR experience (defined as patients treated in 2019).

Analysis by risk groups

Understanding the factors affecting trends is influenced by the expansion of approved indications over time to include increasingly lower-risk patients. Unlike prior Registry reports, when the patient population was more homogenous with only patients with high to extreme risk for SAVR receiving TAVR, the patient population now receiving TAVR encompasses a broad spectrum of surgical risk profiles, that is, from low to extreme risk (8,9). Although surgical risk category is no longer a determinant of candidacy for TAVR in the United States, national trends, as well as a program’s performance, can be better understood by evaluating outcomes within different patient risk categories.

Therefore, data from all patients were categorized into 3 subgroups on the basis of traditional SAVR risk: high or extreme risk, intermediate risk, and low risk. We employed the treatment team’s assessment of surgical risk, which reflects important patient characteristics, such as frailty, that are not captured in the STS calculator for isolated SAVR risk (10).

Not unexpectedly, some patients receiving TAVR were classified as intermediate and low risk before the FDA-approved expansion of indications to these risk groups. The small number of patients reported in the intermittent-risk category treated before FDA approval in mid-2016 and in the low-risk category treated before FDA approval in mid-2019, therefore, represent off-label cases.

Analysis of site volumes

Yearly site volumes were calculated based on site-reported volumes for each year and the number of months a site was performing TAVR in the first year of the site’s activation. Because new sites were being opened each year and may have started performing TAVRs after the first month of the year, there needed to be a calculation of annualized volumes for new sites. See the legend of Table 3 for details.

Analysis of Valve-in-Valve TAVR

Aortic valve-in-valve (V-in-V) procedures represented a form of TAVR that is captured in the Registry. There are several forms of V-in-V performed, and different relevant data elements in the DCF are defined in the data dictionary (6). A planned or pre-procedure indication for V-in-V may be for either degeneration of a surgically implanted tissue valve (TAVR-in-SAVR) or dysfunction of a TAVR valve (TAVR-in-TAVR). Separately, a data element captures an urgent or intraprocedure indication for TAVR-in-TAVR that may occur with acute TAVR dysfunction or deployment of the initial TAVR valve too low or high with resultant aortic regurgitation (AR) and/or unstable anchoring. This is captured in the data element concerning Valve-in-Valve Procedure Status with the option being “Immediate Intraprocedure.”

TAVR system used

Data are collected on the TAVR technology used during the procedure, including the type of TAVR valve implanted. As shown in Table 1, there have been different TAVR valve models from 3 different manufacturers that have received FDA approval from 2011 through 2019. The TAVR valves currently approved have 3 different modes of deployment: balloon-expandable, self-expanding, and mechanically expanding.

Missing data

Missing data for reported data elements are quantified in Table 2 and Supplemental Tables 1 to 3. The results of 30-day and 1-year outcomes for only those patients with submitted data for that element are denoted in the tables as the nonmissing dataset. The nonmissing dataset is used in the Results and in the figures.

Results

From 2011 through 2019, 276,316 patients underwent TAVR with data submitted to the Registry. Data are from 49 sites (a site in Wyoming opened in 2020), as well as cases from the 2 sites in the District of Columbia and 2 sites in Puerto Rico, a U.S. territory. Table 2 presents demographics, clinical characteristics, procedure performance, and in-hospital, 30-day, and 1-year outcomes. These data were compiled by a data run on June 17, 2020, at DCRI. In the Supplemental Appendix are 3 tables of similar data elements but categorized by SAVR risk as assigned by the local heart team, including high or extreme risk (n = 179,397) (Supplemental Table 1), intermediate risk (n = 84,108) (Supplemental Table 2), and low risk (n = 11,534) (Supplemental Table 3).

Overview of TAVR hospital sites and case volume

TAVR sites

The number of U.S. TAVR sites at the end of August 2020 was 715. The year-by-year trend in the number of U.S. hospital sites performing TAVR is shown as part of the Central Illustration. With the opening of a site in Wyoming in March 2020, TAVR programs exist in all 50 U.S. states. The geographic distribution of U.S. TAVR sites is shown in Figure 1A, and Figure 1B show the number of TAVR sites per state. The only TAVR sites in the United States not included in the Registry are those in a few military hospitals and the Veterans Authority (VA) medical system: as of mid-2019 there were 8 VA TAVR programs.

TAVR volume versus SAVR volume

The annual volume of TAVR have increased every year since 2011, and in 2019, TAVR volume (n = 72,991) exceeded all forms of surgical AVR (n = 57,626). Figure 2 displays the annual number of TAVR, isolated SAVR, and other operations involving SAVR, using data from the STS National Database. Annual TAVR volume exceeded isolated SAVR volume in 2016 after TAVR was approved for intermediate surgical risk patients. Annual TAVR volume exceeded all forms of SAVR volume in 2019, coinciding with FDA approval of TAVR for low-risk patients. For the first time, SAVR volume has clearly declined in 2019 as quantified in the legend of Figure 2.

TAVR volume:V-in-V

Elective or planned V-in-V TAVR has increased from 305 cases between 2011 and 2013 to 4,508 in 2019. Immediate or urgent V-in-V during TAVR has decreased from 322 cases between 2011 and 2013 to 208 in 2019. Figure 3 demonstrates the trends in the annual procedural volume of these 2 categories of aortic V-in-V.

Elective V-in-V is predominantly TAVR-in-SAVR with only a small number of elective TAVR-in-TAVR. As shown in Table 2, the pre-procedure indication of TAVR-in-TAVR was infrequent with a total of only 404 patients having been treated in all years. The annual number of planned TAVR-in-TAVR has slowly been increasing, with 110 treated in 2019. This compares with 15,382 patients having planned TAVR-in-SAVR, with 4,480 treated in 2019 alone. Thus, all forms of elective V-in-V were performed in 15,898 in the period 2011 to 2019.

Separately, immediate intraprocedure TAVR-in-TAVR was performed in 1,899 cases over the period 2011 to 2019.

There is an additional relevant data element called Procedure Indication that lists the spectrum of stenosis, insufficiency, and mixed hemodynamic abnormalities but also includes the option of Failed Bioprosthetic Valve. The volumes for the later are significantly less than the volumes for both TAVR-in-SAVR and TAVR-in TAVR categories, probably due to data entry personnel choosing the option that describes the hemodynamic abnormality of the SAVR or TAVR valve rather than the Failed Bioprosthetic Valve option.

Procedure volume by site

The number of TAVR procedures performed per site varies markedly. Table 3 shows the annual trends in mean, median, range, and first and third quartiles of annual TAVR volume per site and the number of sites performing fewer than 50 TAVRs in a given year. The 50-case annual threshold is aligned with the Joint Report of the American Association for Thoracic Surgery, ACC, Society for Cardiovascular Angiography and Interventions, and STS expert consensus document published in 2019 (11). The number of sites performing TAVR has steadily increased, the total annual volume of TAVRs has increased, and the mean annual procedural volume per site has now increased in 2019 to 110 with a median of 84 (IQR: 50 to 137). In 2019, 161 sites performed <50 cases.

Patient characteristics

Demographics

In 2019, the median age of individuals undergoing TAVR was 80 years (IQR: 73 to 85 years), compared with 84 in the years immediately following the initial FDA approval (Table 2). The median age in high/extreme risk patients in 2019 was 81 years (IQR: 74 to 87 years), for the intermediate-risk patients 80 years (IQR: 74 to 84 years), and for the low-risk cohort 75 years (IQR: 70 to 81 years) (Supplemental Tables 1 to 3). There has been a small shift from equal male/female distribution of 48.8%/51.1% in the early TAVR period to a predominance of males in 2019 of 55.8%/44.2%. For all years, patients undergoing TAVR were predominantly of white race. In the first half of 2019, 4.0% were Black/African American, and 5.2% were of Hispanic or Latino ethnicity.

Patient risk categorization

There has been a substantial evolution in the clinical characteristics of patients undergoing TAVR driven by the expansion of approved indications for TAVR. The median 30-day STS PROM (Predicted Risk of Mortality) has steadily fallen from 6.9% (IQR: 4.6% to 10.7%) in 2013 to 4.4% (IQR: 2.7% to 7.2%) in 2019 (Figure 4).

The year-by-year results of the local heart team’s assessment of SAVR risk in patients undergoing TAVR are shown in the Central Illustration. The annual number of patients deemed high/extreme risk initially increased but then stabilized, but remains substantial; this group accounts for approximately 31,000 to 33,000 procedures in each of the last 4 years. The volume of intermediate-risk patients steadily has increased to 32,697 procedures in 2019. The number of patients deemed low risk is starting to increase, with 8,395 patients treated in 2019 representing 11.5% of all TAVR patients.

Comorbid conditions

Patients undergoing TAVR often have other comorbid conditions. The burden of comorbidity has declined with expansion of TAVR into lower-risk populations (Table 2). Gait speed was still abnormal in 69.8% of TAVR patients in 2019. Hostile chest and porcelain aorta, although substantially less common than in the early TAVR experience, were noted in 5.2% and 2.1% of patients, respectively, in 2019. Those using supplemental oxygen has declined from 13.8% to 7.2%.

Functional class

The proportion of patients with New York Heart Association (NYHA) functional class IV symptoms before TAVR has declined from 22.2% to 10.0% from early TAVR experience to 2019. The majority of patients have NYHA functional class III symptoms, in a proportion that has been relatively stable from early TAVR (59.6%) to 2019 (56.75%). There has been a substantial increase in the proportions of patients with NYHA functional class II symptoms, from 13.7% during the early TAVR period to 28.2% in the first half of 2019.

Patient-reported health status at baseline

Patient-reported health status at baseline is shown in Table 4. The Kansas City Cardiomyopathy Questionnaire (KCCQ) provides a measure of the patient’s perception of their health status, including symptoms, impact on physical and social function, and their quality of life. During the early TAVR experience the median score was 44 (IQR: 26 to 64) and, over subsequent years, has slowly increased (Table 2). In 2019, baseline KCCQ scores were available in 67,783 patients representing 93% of patients undergoing TAVR. For all TAVR patients in 2019, the median baseline score was 47 (IQR: 28 to 67). In 2019, the median baseline KCCQ summary score for those patients classified by the heart team as high/extreme risk was 41 (IQR: 23 to 60), intermediate risk was 49 (IQR: 32 to 69), and low risk was 58 IQR: 39 to 78).

Procedure indication

The dominant indication for TAVR was severe native valve aortic stenosis. Since the early period of TAVR to 2019, this has minimally changed from 96.2% to 91.9%. Other indications in 2019 include failed bioprosthetic valve (4.25%), mixed aortic stenosis/AR (3.0%), and primary AR (0.7%). There are no FDA-approved TAVR technologies for AR, and these cases, therefore, represent off-label use of TAVR valves.

Procedure status

Procedures in 2019 were predominantly classified as elective (91.4%); procedures were considered urgent in 8.0%, emergent in 0.3%, and salvage in 0.07%.

Location

The hybrid operating room was used for 54.7% of TAVR procedures in 2019. The use of either a hybrid or regular catheterization laboratory increased slightly from 38.9% in early TAVR experience to 44.7% in 2019.

Anesthesia

A recent report from the Registry documented an increase in the use of conscious sedation from 33% in 2016 to 64% in early 2019, with a large variability in the use of general anesthesia with a few centers using only general anesthesia (12). Furthermore, Registry data demonstrated that the use of conscious sedation compared with general anesthesia was associated with a small, but statistically significantly, lower risk of in-hospital mortality (adjusted risk difference [aRD] 0.2%; p = 0.010), 30-day mortality (aRD 0.5%; p < 0.001), hospital length of stay (LOS, adjusted difference 0.8 days; p < 0.001), and more frequent discharge to home (aRD 2.8%; p < 0.001) (12).

Access site

The vascular access site for TAVR has evolved substantially, as shown in the Central Illustration and Figure 5. First, there has been a steady increase in the use of femoral access, from 57.1% in the early TAVR period to 95.3% in 2019 (Table 2). In 2019, the high/extreme-risk cohort had femoral access in 93.65% compared with 96.2% in the intermediate-risk cohort and 97.8% in low-risk patients. In 2013, there was a transient drop in the percentage of patients having femoral access because of the FDA expansion of indications to include alternative access.

A second trend has been the change in the types of alternative access used (Figure 5). In the early TAVR experience, transapical and direct aortic approaches were predominantly employed, but in 2019, only 213 patients (0.3%) had transapical access and 341 (0.5%) had direct aortic access. In 2019, axillary-subclavian was the most commonly employed alternative approach, that is, 1,816 cases (2.5%). Carotid access has increased to 662 cases (0.9%). Transcaval access was added as an option in April of 2019, and 121 cases were subsequently reported. Previously transcaval was included in the “other’ category.

Use of bypass and conversion to open heart surgery

The use of cardiopulmonary bypass (CPB) has steadily decreased from 4.1% to 0.4%. The use of CPB is currently surprisingly similar in the high/extreme-risk cohort (0.5%), intermediate-risk cohort (0.35%), and low-risk cohort (0.4%). Conversion to open heart surgery (OHS) has declined from 1.4% to 0.4% since TAVR was initially approved. Conversion to OHS is currently similar in the high/extreme-risk cohort (0.4%), intermediate-risk cohort (0.4%), and low-risk cohort (0.5%). These findings suggest that CPB and OHS are most likely required due to unexpected complications that occur at similar rates across the spectrum of patient risk, although it is not known to what extent patients in each risk category have expressed their decision to not be converted to OHS if problems arise during their TAVR.

Second valve and percutaneous coronary intervention

The need for immediate V-in-V procedure has declined (Figure 3). Percutaneous coronary intervention is infrequently performed during TAVR procedures (1.8% in 2019).

Use of adjunctive techniques

Three important adjunctive techniques and technologies have become part of clinical practice and were added as data elements to the DCF in 2018 and 2019 (Table 1). Volume data captured by the Registry regarding these novel approaches to prevent TAVR-related complications are presented in Table 5. First, the Sentinel device (currently Boston Scientific, Marlborough, Massachusetts) was FDA-approved in December 2017 for cerebral protection. From January 2018 until the end of 2019, Sentinel has been used in 11,877 patients. Second, bioprosthetic or native aortic scallop intentional laceration to prevent iatrogenic coronary artery obstruction (BASILICA) was performed in 41 patients undergoing native valve TAVR and 125 patients undergoing V-in-V TAVR in 2018 to 2019. Finally, intentional fracture of the sewing ring of surgically implanted bioprosthetic valves using a high-pressure, noncompliant balloon was performed in 332 patients during V-in-V TAVR in 2018 to 2019. Dedicated reports on the safety and effectiveness of these 3 adjunctive techniques will be forthcoming from the Registry.

TAVR system used in 2019

For native valve TAVR, 72.3% were balloon-expandable valves, 26.7% were self-expanding valves, and 1.0% were the recently approved mechanically expanded valves. For V-in-V TAVR, 53.3% were self-expanding valves, 46.5% were balloon-expandable valves, and 0.2% were mechanically expanded valves. Sites did not report complete valve delivery systems and valve type data in 2.6% patients submitted to the Registry, including patients who had aborted procedures.

Hospital LOS

LOS has declined from a median of 7 days (IQR: 4 to 10 days) to 2 days (IQR: 1 to 3 days) for all patients (Figure 6). In 2019, patients deemed by the heart team to be high/extreme risk had a median LOS of 2 days (IQR: 1 to 5 days), intermediate-risk patients also had LOS of 2 days (IQR: 1 to 3 days), whereas those in the low-risk cohort had a median LOS of only 1 day (IQR: 1 to 2 days).

Patient disposition

In 2019, the majority of patients (90.3%) are discharged home, 6.6% to rehabilitation or extended care facility, and 2.45% to a nursing home. Patients during the early period of TAVR were often discharged to another facility (Figure 7).

Outcomes

Mortality

The year-by-year decline in mortality from the early TAVR experience to 2019 has been steady and dramatic, with in-hospital mortality falling from 5.4% to 1.3%, and 30-day mortality decreasing from 7.2% to 2.5%. These trends are shown in the Central Illustration .

Shown in Supplemental Table 1, the 30-day mortality for those deemed high/extreme risk has declined from 7.15% to 3.8%, and 1-year mortality has decreased from 24.3% to 16.6% from the early TAVR experience to 2018.

In-hospital, 30-day, and 1-year mortality rates were assessed in patients treated in 2018 and broken down by risk cohort (Supplemental Tables 1 to 3). All patients treated in 2018 (n = 59,168) had an in-hospital mortality rate of 1.5%, 30-day rate of 2.6%, and 1-year rate of 12.55% (for the 35,654 patients with 1-year data). Comparing mortality rates for those classified in 2018 as high/extreme (n = 30,993) and intermediate (n = 26,566) showed the following trends: in-hospital mortality rates were 2.1% and 0.8%, 30-day mortality rates were 3.65% and 1.4%, and 1-year mortality rates were 16.6% and 8.3%, respectively.

Stroke

In-hospital stroke rates (Figure 8) have decreased slightly from early TAVR (2.1%) to more recent experience in 2019 (1.6%), as well as 30-day stroke rates (2.75% vs. 2.3%). In-hospital and 30-day stroke rates in 2019 for the high/extreme-risk cohort were 1.9% and 2.7%, intermediate-risk cohort were 1.3% and 1.9%, and the low-risk cohort were 1.3% and 1.9%. The 30-day stroke rate for those classified as high-extreme risk has not declined from the early TAVR experience (2.8%) to 2019 (2.7%).

Permanent pacemaker

Figure 9 demonstrates that permanent pacemaker implantation during the index TAVR hospitalization peaked in 2015 at 13.2% and by 2019 was 8.3%. However, this decrease occurred in the context of shorter hospital LOS (Figure 6). The 30-day pacemaker rate in the early TAVR experience was 10.9%, peaked in 2015 at 15.1%, and then slowly declined, but in 2019 was still substantial at 10.8% and not different from the early TAVR experience. A substantial proportion of pacemaker insertions occurred between hospital discharge and 30 days for all patients and over all years of data collection. For example, in 2019, the in-hospital versus 30-day pacemaker rate was 9.5% versus 11.8% in the high/extreme-risk cohort, 7.9% versus 10.3% in the intermediate-risk cohort, and 6.15% versus 8.2% in the low-risk cohort.

Dialysis

The need for in-hospital dialysis as a result of TAVR has declined from 1.97% during the early TAVR experience to 0.4% in 2019. The need for in-hospital dialysis in 2019 for the high/extreme-risk cohort was 0.7%, the intermediate-risk cohort was 0.3%, and the low-risk cohort was 0.1%.

Bleeding

The use of blood transfusion has declined from 18.2% during the early TAVR experience to 5.8%. Rates of life-threatening/disabling bleeding during index hospitalization declined from 6.3% in the early TAVR experience to 1.8% in 2019. In 2019, the high/extreme-risk cohort had an in-hospital life-threatening/disabling bleeding rate of 2.3%, whereas the intermediate-risk cohort rate was 1.45%, and the low-risk cohort rate was 1.2%. The in-hospital life-threatening/disabling bleeding rate for those at high/extreme risk has declined from 6.3% during the early TAVR experience to 2.3% for 2019.

Vascular complications

Thirty-day major vascular access site complications have declined from 1.6% to 1.3% (Table 2). The 2019 rate was highest in the high/extreme-risk cohort at 1.5%, than in the intermediate-risk cohort at 1.1%, and least in the low-risk cohort at 0.7%. The 30-day major vascular access site complications for those at high/extreme risk has decreased slightly from the early TAVR experience (1.9%) to 2019 (1.8%). Vascular complication rates need to be interpreted in the context of the major shift to predominantly femoral access over the years.

Aortic regurgitation (includes paravalvular leaks)

Moderate/severe AR 30-days post-TAVR was present in 8.0% of patients in the early TAVR experience and has fallen to 1.6% in 2019. In 2019, the rate in the high/extreme-risk cohort was 1.7%, in the intermediate-risk cohort 1.4%, and in the low-risk cohort 1.4%. The 30-day rate of moderate/severe AR for those classified as high-extreme risk has decreased from the early TAVR experience (8.1%) to 2019 (1.75%).

High aortic mean gradient early post-TAVR

The proportion of patients with mean gradients ≥20 mm Hg, either in-hospital or at 30 days, has increased (5.3% before and during 2013 vs. 6.2% in 2019) (Table 2). In 2019, the proportion of patients with the 20 mm Hg or greater gradient in the high/extreme-risk cohort was 6.1%, in the intermediate-risk cohort 6.0%, and in the low-risk cohort 8.2%.

Higher aortic mean gradient at 1 year

An increase in aortic valve mean gradient from post-procedure to 1 year of 10 mm Hg or greater occurred in 3.45% for year 2013 and was detected in 4.7% of patients treated in 2018, in those with 1-year data available. The rate for patients treated in 2018 for the high/extreme-risk cohort was 4.8%, 4.55% in the intermediate-risk cohort, and 4.1% in the low-risk cohort.

1 year alive and well

Table 4 shows the baseline, 30-day, and 1-year KCCQ data of patients treated in 2018. Substantial improvements in KCCQ scores are demonstrated following TAVR. The percentage of patients with complete data for this metric is 71%.

The percentage of patients treated in 2018 achieving a favorable outcome at 1 year, defined as being alive, with a 1-year KCCQ score of ≥60 and no decline of ≥10 points versus baseline score, are presented in Table 4. This stringent definition of a favorable outcome was achieved in 80.7% of all patients, including 77.7% of high/extreme-risk, 83.6% of intermediate-risk, and 85.8% of low-risk patients.

A first look at real-world results in low-risk patients

This report provides an opportunity to perform a preliminary assessment of patients having TAVR who were classified by the local heart team as being low risk for SAVR. As noted in Table 1, FDA approved TAVR in low-risk patients in August of 2019. The 2 published trials of TAVR in low-risk patients included 496 and 725 patients, respectively, who were randomized to TAVR and received the treatment, that is, in the as-treated cohort (13,14). In 2019, there were 8,395 patients entered into the Registry who were classified as low risk with 1,294 treated in the first half of 2019 and 7,101 treated in the second half. The demographics of low-risk patients in the Registry were similar to those enrolled in the trial with a median age of 75 years (IQR: 70 to 81 years) as compared with mean ages of 73 and 74 years in the 2 trials. In the Registry, 65% were male versus 67.5% and 66% in the 2 trials, and 93% were white versus 92% in the 1 trial reporting race. The Registry patients had a median STS PROM score of 2.3 (IQR: 1.6 to 3.45) versus 1.9 and 1.9 in the 2 trials. In the Registry, 48.9% were NYHA functional class III or IV compared with 31.2% and 25.1%. Registry patients had femoral access in 97.8%; by study design, all patients in the 2 low-risk trials had femoral access. Median LOS for Registry patients was 1 day (IQR: 1 to 2 days) compared with a mean LOS of 3 days in 1 of the trials, with the caveat that the 2 trials treated patients in 2016 to 2018. In-hospital mortality for the Registry patients was 0.5% compared with 0.4% in 1 trial. Thirty-day mortality was 0.4% and 0.5% in the 2 trials. At the time of this report, comparisons between low-risk patients in the Registry and those in the 2 trials are limited because 30-day follow-up is incomplete for the majority of Registry patients treated in the second half of 2019, and in the 2 low-risk trials, 30-day outcomes are reported more comprehensively than in-hospital outcomes.

Discussion

Multiple temporal trends including some of historic proportions are identified from these data and comprehensively describe the current state of TAVR in the United States, up until the onset of the COVID-19 pandemic. The rapid growth in the number of sites and case volume, the broader spectrum of patients treated, TAVR becoming the dominant form of AVR, and the lower procedure burden with fewer complications are all well-documented in these data. The dramatic decrease in LOS, the high percentage of patients being discharged directly to home, and 8 of 10 patients, including high-extreme risk patients, achieving at 1 year the “alive and well” patient-reported outcome are all testaments to the on-going reduced burden of TAVR treatment and achievement of benefits to patients including those who are elderly with a heavy dose of comorbid conditions.

This report from the Registry, using a huge volume of observational data, documents these changes and other substantial advances in TAVR. The logical next question is what accounts for these changes. The field has many dynamic elements that may be considered, including improvements in TAVR technology and techniques, adjunctive technologies, the level of heart team experience, and changes in the patient risk profiles.

In Table 6, 10 highlights from this report are listed along with associated next steps in using the Registry to perform a deeper dive to identify associations and areas for further trials and improvements in care. As noted, the impact of COVID-19 and the extent and timing of a subsequent recovery will depend on a myriad of factors including geographical differences in COVID-19 penetration, local policies having an impact on the resumption of nonemergent procedures, and patients’ willingness to access the medical system out of fears of COVID-19 infectivity (12).

The indications for TAVR in the United States are outlined in FDA approvals. Beginning in 2021, the Registry will start to collect and report data on appropriate use criteria as defined by a professional society consensus document (15). The consensus document was written before TAVR being approved for low-risk patients and, therefore, will need to be updated. In general, patients receiving TAVR have native aortic stenosis in 91.5% and failed prosthetic aortic valves in 4.5% in 2019. Only 3.55% were reported to be NYHA functional class I, that is, asymptomatic.

The increasing incidence of TAVR V-in-V procedures (6.7% of all TAVRs in 2019) must be considered when interpreting the overall, that is, in all TAVR, Registry results. TAVR V-in-V procedures have a high rate of initial success (16) but increase the risk several-fold for severe prosthesis–patient mismatch, which in turn has been associated with an increase in short-term mortality and heart failure rehospitalization (17), as well as an associated with higher gradients and mortality, particularly in small surgical prostheses (18). These procedures may also pose different risks for other outcomes, including stroke, pacemaker, and paravalvular regurgitation compared with TAVR performed in a native annulus.

Patient-reported outcomes: KCCQ assessment

The Registry uniquely incorporates patient-reported outcomes and has used the KCCQ tool administered to patients at baseline, 30-days post-TAVR, and 1-year post-TAVR. The results as displayed in Table 4 show the substantial impact of aortic valve disease on patients pre-TAVR with a median baseline KCCQ of 45 in 2018. The increase to a median value of 81 took only 30 days post-TAVR to be manifest, and by 1-year, the median KCCQ was 86, albeit this reflects the results only among survivors.

The percentage of patients currently undergoing TAVR having a favorable 1-year outcome of being both alive and well has increased to approximately 8 of 10 patients from a previous report during the early TAVR experience when only 6 of 10 achieved this metric of success (7). Even patients deemed high/extreme risk have now achieved this key 1-year metric in 78%, which is also noteworthy because of the large number of patients who continue to be treated in this risk category.

Outcomes by risk categories

By analyzing outcomes based on the heart team’s assessment of SAVR risk, a better understanding emerges of factors that are related to key outcomes, with the caveat that these are observational data, and the causes of trends can be suggested, but not proven.

This risk-based analysis also suggests the major degree to which the rates of some TAVR outcomes metrics are associated with patient factors. The rates of mortality, bleeding, stroke, and the need for a permanent pacemaker are all associated with the heart team’s risk categorization of their patients. On the other hand, the degree of post-TAVR aortic regurgitation, high residual gradient, and an increased valve gradient at 1 year do not. Likewise, the rates of needing CPB and conversion to OHS are similar across all risk groups, but fortunately very low in current TAVR practice.

In addition, this risk-based analysis provides insight into the extent to which TAVR outcomes have improved over the years independent of patient risk. Specifically, rates of mortality (30-day and 1-year), bleeding, and moderate/severe AR post-TAVR have declined substantially among patients in the high/extreme-risk category. This implies a substantial improvement in the quality of care.

Although TAVR is now approved in the United States for patients in all SAVR risk categories, it remains useful for heart teams to continue to perform comprehensive pre-procedural assessment of patient characteristics and the associated risk categories. Identification of patients at higher risk for complications may potentially modify the nature of procedure performance and post-procedure care. In addition, heart teams better understand their program’s performance compared with national benchmarks of quality metrics and case mix.

Changes in procedure performance

Multiple dynamic factors of TAVR performance are also identified and quantified. Two major shifts involve anesthesia and vascular access that are often interconnected.

Conscious sedation

The shift from general anesthesia to conscious sedation is noteworthy because it is typically accompanied by no longer routinely using transesophageal echocardiography, intubation, and other invasive monitoring techniques (16).

Shifts in vascular access

The ability to treat most patients, 95% of all patients, with transfemoral access has been associated with lower bleeding rates, reduced LOS, and discharges to home. In those requiring alternative access, there has been a dramatic shift to subclavian-axillary access, plus other approaches less commonly used, and away from transapical and direct aortic approaches.

Major complications

The infrequent need for CPB and conversion to OHS represent another reason TAVR has become lower risk. But the occasional occurrence of emergency transitions to mechanical support and OHS during TAVR, across all risk groups, justifies that all patients undergo TAVR in facilities with teams capable of performing these potentially life-saving procedures.

Clinical challenges remain, and the volume of patients classified as high/extreme risk remains substantial; these data should focus attention on strategies to further improve outcomes. It is noteworthy that there has been a decline in rates of stroke and perhaps in the need for permanent pacemaker, but these declines are small.

Stroke

The self-reported rates of stroke are low, but underreporting of clinically mild strokes is a known phenomenon (19). A small decline in stroke rates appears associated with the addition of large numbers of patients in the intermediate-risk category and, more recently, low-risk patients. An analysis suggests stroke rates are related to risk category. Further reports from the Registry on stroke and the impact of the use of cerebral protection are forthcoming; issues of ascertainment bias will confound site-reported stroke rates, and the multivariable analysis of patient and procedure factors on stroke rates will be important.

Need for permanent pacemaker

Pacemaker rates remain high. It is perplexing that there has been only a small decline in recent years with the inclusion of large numbers of patients in the intermediate-risk category despite the finding that pacemaker rates are related to patient risk category. This issue is clinically problematic, given the short hospital LOS after TAVR and an increasing proportion of patients require a pacemaker after hospital discharge, during which time, there is some risk of sudden death and other complications due to conduction system block.

Conclusions

The Registry, as an example of an innovative national learning health care system, provides in this report, and in the numerous previous publications using Registry data, a comprehensive observational warehouse of patient-level data from over a quarter of a million patients, with analyses that provide great insight into the state of TAVR in the United States. TAVR is the dominant form of AVR, with sites in all 50 states. Outcomes out to 1 year have steadily improved. Further growth is expected with recovery of the health care system in the new world of COVID-19.

Footnotes