A Randomized Comparison of Self-Expanding Transcatheter and Surgical Aortic Valve Replacement in Patients With Severe Aortic Stenosis Deemed High-Risk for Surgery - CoreValve High Risk


The introduction of transcatheter aortic valve replacement (TAVR) has resulted in a paradigm shift in the treatment of patients with severe aortic stenosis. In the PARTNER studies, balloon-expandable TAVR with the Edwards Sapien valve was superior to medical management in the treatment of inoperable patients (Cohort B), and noninferior to surgical AVR for the treatment of patients with a high surgical risk (Society of Thoracic Surgeons [STS] score ≥8%) (Cohort A). This led to Food and Drug Administration (FDA) approval for this device in these patient subgroups.

More recently, self-expanding TAVR with the Medtronic CoreValve met its primary efficacy endpoint in the Extreme risk trial (similar to PARTNER Cohort B). The current trial sought to investigate the safety and efficacy of TAVR with the CoreValve compared with surgical AVR in patients who were at high risk for surgery (similar to PARTNER Cohort A).

Contribution to the Literature: The CoreValve High-Risk trial indicates that TAVR with the self-expanding CoreValve is noninferior to surgical AVR in the treatment of patients with high-risk aortic stenosis up to 3 years of follow-up.

Study Design

  • Randomized
  • Blinded
  • Parallel

Patient Populations:

  • NYHA functional class II or greater
  • Severe aortic stenosis: aortic valve area (AVA) ≤0.8 cm2 or AVA index ≤0.5 cm2/m2 AND mean gradient >40 mm Hg or peak velocity >4 m/sec at rest or with dobutamine stress echocardiogram
  • Risk of death at 30 days after surgery was ≥15% and risk of death or irreversible complications within 30 days was <50%
  • Surgical risk assessment included consideration of STS Predicted Risk of Mortality estimate and other risk factors not captured in the STS risk model
  • Number screened: 995
  • Number of enrollees: 795
  • Duration of follow-up: 12 months, 2 years, 3 years
  • Mean patient age: 83 years
  • Percentage female: 47%


  • Recent active gastrointestinal bleed (<3 months), stroke (<6 months), or myocardial infarction (≤30 days)
  • Any interventional procedure with bare-metal stents (<30 days) and drug-eluting stents (<6 months)
  • Creatinine clearance <20 ml/min
  • Significant untreated coronary artery disease
  • Left ventricular ejection fraction <20%
  • Life expectancy <1 year due to comorbidities

Primary Endpoints:

  • All-cause mortality at 1, 2, and 3 years

Secondary Endpoints:

  • Prespecified hierarchical testing
  • Δ mean gradient baseline to 1 year (noninferior)
  • Δ effective orifice area baseline to 1 year (noninferior)
  • Δ NYHA class baseline to 1 year (noninferior)
  • Δ Kansas City Cardiomyopathy Questionnaire baseline to 1 year (noninferior)
  • Difference in major adverse cardiac and cerebrovascular event rate at hospital discharge or 30 days, whichever is later (superiority)
  • Δ Short Form-12 baseline to 30 days (inequality)

Drug/Procedures Used:

Patients meeting inclusion criteria were randomized in a 1:1 fashion to either TAVR or surgical AVR. The CoreValve system comes in 4 sizes (23 mm, 26 mm, 29 mm, and 31 mm) and can be used to treat annuli between 18 and 29 mm. The delivery sheath is 18F, and the valve can be delivered either via transfemoral, direct aortic route, or subclavian route. Patients assigned to surgical AVR were treated by means of conventional open-heart techniques with the use of cardiopulmonary bypass. The choice and size of the surgical prosthetic valve were left to the discretion of the surgeon.

Concomitant Medications:

Aspirin was used in SAVR patients. In the TAVR patients, dual antiplatelet therapy with aspirin at a dose of at least 81 mg daily, and clopidogrel at a dose of 75 mg daily, was recommended before the procedure and for 3 months after the procedure, followed by aspirin or clopidogrel monotherapy at the same dose indefinitely. In the event that warfarin was indicated for other reasons, aspirin, at a dose of at least 81 mg daily, and warfarin were administered indefinitely without clopidogrel.

Principal Findings:

A total of 747 patients were randomized, 390 to TAVR and 357 to surgical AVR. Baseline characteristics were similar between the two arms. Approximately 40% had diabetes mellitus and 12% had prior stroke. The mean STS score was 7.4%, with a corresponding mean logistic EuroSCORE of 18.1.

Other conditions contributing to a high-risk designation included oxygen-dependent respiratory insufficiency (12%), use of immunosuppressive medications (9%), and frailty based on prespecified criteria. This included a Charlson comorbidity score of >5 (severe) in 56%, Katz index with ≥1 activities of daily living deficit in 12% and assisted living status (10%). Approximately 86% of the patients had New York Heart Association (NYHA) class III or IV symptoms.

At 1 year, TAVR was superior to surgical AVR for the primary endpoint of all-cause mortality (14.2% vs. 19.1%, p < 0.0001 for noninferiority; p = 0.04 for superiority). Mortality at 30 days was lower than predicted in both arms (3.3% vs. 4.5%). All strokes were numerically lower in the TAVR arm at 30 days (4.9% vs. 6.2%) and at 1 year (8.8% vs. 12.6%, p = 0.1). Major strokes were similar at 30 days (3.1% vs. 3.9%) and at 1 year (5.8% vs. 7.0%, p = 0.59).

Other endpoints at 30 days including vascular complications (5.9% vs. 1.7%, p = 0.003) and need for permanent pacemaker (19.8% vs. 7.1%, p < 0.001) were worse in the TAVR arm, while severe bleeding (13.6% vs. 35.0%, p < 0.0001) and new-onset or worsening atrial fibrillation (11.7% vs. 30.5%, p < 0.0001) were better. Symptoms were improved similarly in both arms. The mean gradient was lower in the TAVR arm at 1 year (9.1 mm Hg vs. 12.4 mm Hg), and the effective orifice area was larger (1.91 cm2 vs. 1.57 cm2).

At discharge, paravalvular aortic regurgitation was noted in nearly 42.1% of patients in the TAVR arm compared with 3.3% in the SAVR arm, although it was moderate or severe in only 7.8% of TAVR patients. At 1 year, this was somewhat diminished in the survivors to 6.1%. 

Two-year outcomes: All-cause mortality: 22.2% vs. 28.6%, p = 0.04; strokes: 10.9% vs. 16.6%, p = 0.05; MACCE: 29.7% vs. 38.6%, p = 0.01; permanent pacemaker implantation: 25.8% vs. 12.8%, p < 0.001; reintervention: 2.5% vs. 0.4%, p = 0.02; NYHA class III/IV symptoms: 7.9% vs. 9.5%; AV area: 1.87 vs. 1.51 cm2; moderate to severe paravalvular AR: 6.5% vs. 0.6%.

Three-year outcomes: All-cause mortality: 32.9% vs. 39.1%, p = 0.07; strokes: 12.6% vs. 19.0%, p = 0.034; MACCE: 40.2% vs. 47.9%, p = 0.03; permanent pacemaker implantation: 28.0% vs. 14.5%, p < 0.001; reintervention: 2.5% vs. 0.4%, p = 0.02; AV area: 1.79 vs. 1.53 cm2; mean gradient: 7.62 vs. 11.4 mm Hg, p < 0.0001; moderate to severe paravalvular AR: 5.9% vs. 0%.

Cost-effectiveness analysis: Index hospitalization costs were higher with TAVR versus SAVR ($69,592 vs. $58,332, difference = $11,260, p < 0.0001). Total follow-up costs were, however, lower in the TAVR arm at 12 months ($28,766 vs. $30,819, difference = $2,053, p = 0.52). Cost-effectiveness analysis demonstrated an incremental cost-effectiveness ratio (ICER) of $67,059 per quality-adjusted life-year (QALY) gained, with a greater ICER in iliofemorally treated patients ($55,535 per QALY gained).


The results of the CoreValve High-Risk trial indicate that TAVR with the self-expanding CoreValve is noninferior to surgical AVR in the treatment of patients with high-risk aortic stenosis up to 3 years of follow-up. ICER analysis suggests acceptable cost-effectiveness per current US standards; however, a reduction in index hospitalization costs could significantly improve this profile. These are very important results, and helped to expand the treatment armamentarium for the management of this sick patient population. It is unclear if this trial will lower the bar for eligible patients (currently STS score of ≥8%) going forward. The results of other ongoing trials in intermediate-risk patients are eagerly awaited.

Although they have several similarities, there are also some important differences between the current trial and PARTNER Cohort A. In the latter, TAVR with the balloon-expandable Edwards Sapien valve was noninferior, but not superior, to surgical AVR for the high-risk patient subset. Also, the patients included in this trial were not as high-risk as those included in PARTNER Cohort A (mean STS 7.4 vs. 11.7). There was also a significantly higher need for a permanent pacemaker with the CoreValve, but a much better vascular complication rate profile (due to use of 18F sheaths for CoreValve vs. 22F and 24F sheaths in PARTNER).

Finally, transapical TAVR appeared to have worse outcomes compared with transfemoral TAVR vs. surgical AVR in PARTNER Cohort A; the CoreValve does not currently have a transapical option. Outside of a direct head-to-head comparison, it is thus hard to make too many inferences regarding superiority of one device over another.


Deeb GM, Reardon MJ, Chetcuti S, et al. Three-Year Outcomes in High-Risk Patients Who Underwent Surgical or Transcatheter Aortic Valve Replacement. J Am Coll Cardiol 2016;Apr 3:[Epub ahead of print].

Editorial: Kumbhani DJ. Three-year results of a TAVR trial in high surgical risk patients. J Am Coll Cardiol 2016;Apr 3:[Epub ahead of print].

Presented by Dr. G. Michael Deeb at the American College of Cardiology Annual Scientific Session, Chicago, IL, April 3, 2016.

Adams DH, Popma JJ, Reardon MJ, et al., on behalf of the U.S. CoreValve Clinical Investigators. Transcatheter Aortic-Valve Replacement With a Self-Expanding Prosthesis. N Engl J Med 2014;370:1790-8.

Presented by Dr. David Adams at the American College of Cardiology Annual Scientific Session, Washington, DC, March 29, 2014.

Presented by Dr. Matthew Reynolds at the Transcatheter Cardiovascular Therapeutics meeting (TCT 2014), Washington, DC, September 13, 2014.

Presented by Dr. Michael J. Reardon at ACC.15, San Diego, CA, March 15, 2015.

< Back to Listings