Cover Story: TAVR, and more TAVR, at ACC.17

This year’s ACC Scientific Session could have been subtitled, “PCSK9s and the TAVR/SAVR update meeting,” what with all the important data presented on both topics. For those in the cath lab, the most exciting trial presented was SURTAVI, followed by some new data on neurological outcomes after transcatheter aortic valve replacement (TAVR) and surgical aortic valve replacement (SAVR) and a slew of registry and observational data that further finessed our understanding of the transcatheter option. The implications of these data and more will be the topic of a “deep dive” article in the next issue of Cardiology: Interventions. See our feature article for a round-up of other interventional data from ACC.17, including a full dose of fractional flow facts.


TAVR continues its relentless descent down the risk ladder: SURTAVI showed clearly that TAVR using the self-expanding CoreValve or second-generation Evolut R stands head-to-head with SAVR in patients with symptomatic severe aortic stenosis (AS) deemed at intermediate surgical risk.

“The implications of this to me, as a cardiac surgeon who has been practicing for 34 years and as a TAVR implanter, is that this outstanding data should, and hopefully will, lead to very rapid approval of this valve,” said principal investigator, Michael J. Reardon, MD, FACC.

Seems the U. S. Food and Drug Administration (FDA) was listening in (or waiting for the findings to become public): they announced approval of the CoreValve Evolut PRO valve just a few days after this Late-Breaking Clinical Trial report, albeit for the treatment of symptomatic severe AS in patients who are at high or extreme risk for SAVR.

SURTAVI enrolled 1,746 patients at 87 centers in the U.S., Europe and Canada, and randomly assigned them to TAVR or SAVR. Participants had mortality risk ≥3 and <15 percent based on The Society of Thoracic Surgeons Predicted Risk of Mortality (STS PROM) score and overall clinical status including frailty, disability and comorbidity factors. Participants were NYHA functional class II or greater, but those with a SYNTAX score >22 were excluded. Mean patient age was 80 years and 56 percent were men. Mean STS PROM score was 4.4 percent.

At two years, the estimated incidence of the primary endpoint of all-cause mortality or disabling stroke was 12.6 and 14.0 percent in the TAVR and SAVR groups (p < 0.005 for non-inferiority). Two-year mortality was 11.4 and 11.6 percent, respectively. The disabling stroke rates were 2.6 and 4.5 percent, respectively, which strongly trended towards favorability for TAVR, noted Dr. Reardon in his presentation of the trial results.

Thirty-day mortality was 2.2 percent for TAVR and 1.7 percent for SAVR, the latter figure being indicative of “very good surgery by excellent surgeons,” said Reardon.

TAVR was associated with higher rates of residual aortic regurgitation and the need for pacemaker implantation (25.9 vs. 6.6 percent with SAVR; p < 0.05), as well as lower mean gradient and larger aortic-valve areas (AVAs) compared with SAVR. Indeed, the AVAs for TAVR were significantly better than for SAVR at all follow-up visits.

Only 16 percent of patients (mostly U.S. patients enrolled after April 2015) received the newer, second-generation CoreValve Evolut R valve, which has a smaller sheath size and pacemaker requirement than the CoreValve classic, and which may further improve outcomes than were seen in this trial. Reardon noted the need for a pacemaker was not associated with increased mortality in SURTAVI.

In comparison with TAVR, SAVR was associated with higher rates of acute kidney injury, atrial fibrillation and transfusion requirements. Structural valve deterioration at 24 months was not seen in either group.


Subclinical leaflet thrombosis of bioprosthetic aortic valves was identified in 12 percent of patients post TAVR and 4 percent post SAVR (p = 0.001) in registry findings reported by Raj R. Makkar, MD, FACC. Anticoagulation with either direct oral anticoagulants (DOACs) or warfarin was effective in preventing or treating subclinical leaflet thrombosis, but dual antiplatelet therapy (DAPT) was not, leaving experts questioning the appropriate anticoagulation regimen post bioprosthetic valve implantation.

“The choice of therapy, whether it is TAVR or SAVR is better guided by clinical outcomes data rather than a single imaging finding such as subclinical leaflet thrombosis…,” said Makkar. “We believe that despite the excellent clinical outcomes with new-generation valves, our study findings can help further optimize adjunctive pharmacotherapy, which has the potential to further improve valve hemodynamics and clinical outcomes.”

The RESOLVE and SAVORY single-center prospective registries provided interpretable CT imaging on 890 TAVR and SAVR patients, performed using a dedicated 4-D volume-rendered imaging protocol at varying intervals after valve replacement. Subclinical leaflet thrombosis was identified as the presence of reduced leaflet motion, along with corresponding hypo-attenuating opacities seen in the corresponding leaflets. Restricted movement on CT imaging is thought to be indicative of possible thrombosis at the valve site.

Aortic valve gradients of >20 mm Hg and increases in aortic valve gradients of >10 mm Hg were seen in 14 percent of those with subclinical leaflet thrombosis compared with only 1 percent of those with normal leaflet motion (p < 0.0001), but stroke rates did not differ between groups. Transient ischemic attack (TIA) was increased in those with reduced leaflet motion (p = 0.0005), as was the composite of stroke and TIA (p = 0.001).

Analysis of a small group of patients (n = 58) who underwent a second CT scan three months after the first scan showed subclinical leaflet thrombosis resolved over time in all patients who were started on warfarin or DOACs after the first CT scan. In those not started on anticoagulation, 91 percent had persistent or progressively reduced leaflet motion at the time of the second scan (p < 0.001).

“If a patient is started on anticoagulation, warfarin or a DOAC, it [stent thrombosis] goes away, and if it’s stopped, it [stent thrombosis] comes back. There is at least an association with TIA and the composite of stroke and TIA. But before this is actionable, we need randomized trials, which are being done,” said Deepak L. Bhatt, MD, MPH, FACC. “An important message that shouldn’t get lost, is that TAVR is a great option for the right patients, and this [stent thrombosis] shouldn’t detract from that, but this does mean that perhaps before we are doing TAVR on 50-year-olds, we really want to have some longer-term data and sort out whether we need to anticoagulate a proportion of them or not.”

Cerebral Protection in SAVR

Both SAVR and TAVR are dogged by concern regarding post-procedural brain injury, stroke and reductions in neurocognitive functioning. In a study presented by Michael J. Mack, MD, FACC, cerebral embolic protection during SAVR, using two mechanistically different devices, appeared to do little to reduce the risk of stroke. Similar findings have been reported for TAVR.

The study involving 383 patients undergoing SAVR who were randomly assigned to procedures that included use of the Embol-X intra-aortic filter (n = 133), the aspiration-based CardioGard embolic protection cannula (n = 118) or SAVR without embolic protection (n = 132). Both devices have FDA clearance.

The primary endpoint of freedom from clinical or radiographic central nervous system (CNS) infarction at seven days did not differ with either device (27.1 percent for Embol-X, 32.7 percent for CardioGard, and 34.8 percent for controls; p nonsignificant for both devices vs. controls).

Trial enrollment was stopped at 77 percent of intended enrollment at the recommendation of the Data Safety and Monitoring Board due to anticipated futility.

Investigators saw clear capture of embolic debris in most patients, a reduction in delirium in the two study arms and an observed difference in infarct size distribution with fewer large volume infarcts. The association between the clinical and radiographic indications of brain injury seen in these patients are the subject of an ongoing investigation.

“Although this appears to be a negative study, as a vascular neurologist, I’m very excited by it,” said panel discussant David Thaler, MD, PhD. “It’s got to be a good thing to stop all that junk from going up in to the head… I think the problem with this and other studies is the ability to detect the outcome,” he suggested. If there was a sensitive biomarker of brain injury, a “brain creatinine” he suggested, then this might have been a positive study.

Microbleeds in TAVR

Adding grist to the post aortic valve replacement brain debate was a trial showing that more than one in five patients have cerebral microbleeds during TAVR or in the immediate perioperative period. Before TAVR, MRI scanning showed at least one microbleed in 26 percent of patients. At three days after the procedure, a total of 40 percent of patients had microbleeds and 23 percent had new microbleeds that were not present before TAVR.

Cerebral microbleeds are extravasations of blood components through fragile cerebral microvascular walls. They are known as markers of risk for stroke, dementia and cognitive impairment, reported Eric Van Belle, MD, PhD, FACC, but this is the first time it’s been shown that a cardiovascular or endovascular procedure could impact or trigger their occurrence.

The single-center trial enrolled 84 patients (mean age 80 years; 50 percent women) with AS undergoing implantation of the Edwards Sapien heart valve delivered transfemorally. Patients had neurologic assessments and cerebral MRI scans before and after their TAVR procedure. Survivors (n = 75) had a second neurologic assessment at six months post TAVR.

Five percent of patients had cerebral emboli detected before the procedure and this proportion increased to 64 percent post procedure. The rates of stroke/TIA were low, as would be expected in this kind of population, noted Van Belle.

The presence of microbleeds was associated with poorer neurological outcomes at six months. Factors associated with a higher risk of microbleeds included longer procedural times, history of bleeding and prolonged exposure to anticoagulation.

This trial was small, stated Van Belle, but highlights the need for further study to investigate the occurrence and impact of cerebral bleeding associated with aortic valve interventions. “In particular, we believe that it will be really important, because we have a lot of studies dedicated to new anticoagulation medication, to have such MRI sub-studies as part of these studies,” said Van Belle. It’s also important, he noted, to gain a better understanding of this TAVR complication in younger patients who may undergo the procedure in greater numbers as the acceptable risk indications fall.

Featured Interventional Clinical Research

ADVICE: Real-World TAVR Versus SAVR

Are the results seen in the randomized TAVR versus SAVR trials generalizable to “real-world” clinical settings? That was the question asked by the ADVICE trial, and the answer appears to be a solid yes. In an unselected population of real-world patients and hospitals in the U.S., TAVR was a safe and effective alternative to SAVR for intermediate- and high-risk patients, offering similar outcomes and a less demanding early recovery.

J. Matthew Brennan, MD, presented the results of this comparative effectiveness study that included 9,464 intermediate- and high-risk (STS PROM ≥3 percent) patients. Propensity matching was used to control for baseline differences.

In-hospital death was lower for TAVR (3.0 vs. 5.0 percent for SAVR; p < 0.001), as was new dialysis (1.7 vs. 3.2 percent (p < 0.001). However, new pacemaker implantation was higher (12.8 vs. 6.3 percent; p < 0.001). The one-year rates of death and stroke were both similar between groups, with no interactions noted between the two procedures for mortality or stroke in relation to higher or lower STS PROM scores.

Only 41 percent of SAVR patients were discharged home, compared with 70 percent of TAVR patients, and the mean length of stay was eight days and four days, respectively (p < 0.001 for both comparisons).

STS/ACC TVT Registry

Data from the STS/ACC TVT registry linked to Medicare claims data were used to compare the one-year safety and efficacy of TAVR with valve-in-valve (ViV) implantation for a failed surgical prosthesis with native valve (NV) TAVR. To control for baseline differences, a 2:1 NV TAVR versus ViV TAVR matched cohort was created using STS PROM for hostile chest, porcelain aorta and inoperability/extreme risk designation.

For the matched cohort of 750 ViV TAVR patients and 1,495 NV TAVR patients, the in-hospital and 12-month death rates were significantly lower for ViV compared with NV (Figure 1), as were stroke rates (2.0 vs. 4.3 percent at one year; p = 0.01).

Median age was 79 for ViV and 84 years for NV. When morality was considered by age <80 years or ≥80 years, no difference was seen between ViV and NV TAVR. Mortality was similar when considered in relation to capacity to perform a five-minute walk test in less than six seconds.

For ViV TAVR compared with NV TAVR, post-TAVR echocardiography showed higher aortic valve mean gradients (16 vs. 8 mm Hg) and lower rates of moderate or severe aortic regurgitation (4.6 vs. 8.3 percent; p < 0.001 for both comparisons).

“Results of this study suggest that valve-in-valve TAVR should be considered as the treatment of choice for inoperable and probably for high-risk patients,” said principal investigator E. Murat Tuzcu, MD, FACC.

Procedural and Clinical Outcomes

Because of its unique morphology, congenital bicuspid AS was excluded from the randomized trials that established the safety and efficacy of TAVR. Turns out the TAVR option, particularly using the newer generation devices, is a good one for younger patients with bicuspid AS, according to a 33-site international registry.

Outcomes of 561 patients with bicuspid AS and 4,546 patients with tricuspid AS were compared after propensity-score matching (546 pairs). Sung-Han Yoon, MD, presented procedural and clinical outcomes during a Featured Clinical Research session.

Patients with bicuspid AS had significantly lower device success rates (85.3 vs. 91.4 percent; p = 0.002) and, overall, somewhat poorer procedural outcomes. However, their 30-day clinical outcomes were very similar with those seen in tricuspid AS patients.

But here’s the kicker: when the investigators singled out those who received newer aortic valve implants, as opposed to the early-generation devices—so, the Evolut R instead of the CoreValve, for example—the device success rates and procedural outcomes were virtually identical with those for tricuspid AS. Conversely, when only those who received early-generation devices were identified, the procedural outcomes were significantly better for those with tricuspid AS.

One-year all-cause mortality did not differ between bicuspid and tricuspid AS, across early- and new-generation devices.

Early-generation devices (Sapien XT/CoreValve) were implanted in 320 patients with bicuspid AS and 321 patients with tricuspid AS, whereas new-generation devices (Sapien 3/Lotus/Evolut R) were implanted in 226 and 225 patients with bicuspid and tricuspid AS, respectively.

Clinical Topics: Cardiac Surgery, Invasive Cardiovascular Angiography and Intervention, Valvular Heart Disease, Aortic Surgery, Cardiac Surgery and VHD, Interventions and Structural Heart Disease

Keywords: Cardiology Interventions, ACC Publications, Aortic Valve, Aortic Valve Stenosis, Heart Valve Prosthesis, Registries, Transcatheter Aortic Valve Replacement

< Back to Listings