PARTNER 2

Published

Leon MB, et al. "Transcatheter or Surgical Aortic-Valve Replacement in Intermediate-Risk Patients". The New England Journal of Medicine. 2016. 374(17):1609-20.
PubMed • Full text • PDF

Clinical Question

In patients with symptomatic severe AS who are intermediate-risk surgical candidates (expected periprocedural mortality ~ 4-8%), is transcatheter aortic-valve implantation (TAVI) noninferior to surgical aortic valve replacement (AVR) with regard to all-cause mortality and disabling stroke?

Bottom Line

In patients with symptomatic severe AS who are intermediate-risk surgical candidates (expected periprocedural mortality ~ 4-8%), TAVI was noninferior to surgical AVR with respect to all-cause mortality and disabling stroke at 2 years.

Major Points

It is well-established that aortic valve replacement (AVR) prolongs survival in patients with severe aortic stenosis (AS). However, the condition often presents at an advanced age when patients are poor surgical candidates due to multiple comorbidities. Transcatheter aortic valve implantation (TAVI), an emerging technology in which a self-expanding aortic valve prosthesis is inserted using a transcatheter approach, is a less invasive alternative to open surgical AVR that has been shown to be an effective option both in patients who are at prohibitive surgical risk (20% absolute reduction in mortality vs. medical management; PARTNER B) and in patients who are high surgical risk (similar mortality to surgical AVR; PARTNER A). Whether these encouraging results with TAVI could be extended to the large population of AS patients at intermediate-risk for surgical complications was unknown.

The 2016 Placement of Aortic Transcatheter Valves 2 (PARTNER 2) trial randomized 2032 patients with symptomatic severe AS at intermediate perioperative risk (estimated perioperative mortality ~ 4-8%) to TAVI or surgical AVR. At 2 years, TAVI was noninferior to surgical AVR with respect to the primary endpoint, with very similar rates of all-cause mortality or disabling stroke in both groups. Furthermore, TAVI and surgical AVR resulted in similar rates of improvement in clinical heart failure as well as echocardiographic valvular function. Importantly, there was a borderline significant (p=0.06) interaction between TAVI access approach and the primary endpoint, with a 4% absolute decrease in the primary endpoint with transfemoral TAVI and a 4% absolute increase in the primary endpoint with transthoracic TAVI, suggesting that transfemoral TAVI may actually be superior to surgical AVR while transthoracic TAVI may be inferior (with the pooled combination of the two resulting in noninferiority). Complications more prevalent with open AVR included a 30% higher rate of major bleeding, a 2% higher rate of acute kidney injury, and a 16% higher rate of incident atrial fibrillation. Conversely, TAVI was associated with a 3% increase in major vascular complications.

Based largely on the results of PARTNER A and PARTNER B, the 2014 ACC/AHA guidelines uniformly recommend TAVI for patients at either prohibitive or high-risk for surgical complications. These guidelines continue to recommend surgical AVR for patients at low- or intermediate-risk, although these recommendations may need to be revised given the results of PARTNER 2 which provide strong evidence that TAVI results in equivalent material outcomes (i.e., mortality and stroke) when compared to the surgical approach. The interaction between TAVI access site and outcomes suggests that transfemoral TAVI may even be superior to surgical AVR in this population (and transthoracic TAVI may be inferior), although these subgroups were underpowered and these hypothesis-generating findings require further prospective study. In the meantime, taken together the results of PARTNER 2 suggest that the TAVI is noninferior to surgical AVR in intermediate-risk patients, and that the decision to proceed with either approach may need to be made on a case-by-case basis while considering factors such as availability of transfemoral access, patient preference, and provider experience.

Guidelines

There have been no guideline recommendations based on the results of this study.

Design

• Prospective, multicenter, randomized, open-label, comparative trial
• N=2032
  • TAVI (n=1011)
  • Surgical AVR (n=1021)
• Setting: 57 centers in US and Canada
• Enrollment: December 2011 - November 2013
• Duration follow-up: 2 years
• Analysis: Intention-to-treat
• Primary outcome: All-cause mortality or disabling stroke at 2 years

Population

Inclusion Criteria

• Severe AS defined as:
  • Aortic valve area <0.8 cm² AND either of:
    • Mean aortic valve gradient ≥40 mmHg
    • Peak aortic jet velocity of ≥4 m/sec
• NYHA II, III, or IV symptoms
• Candidate for AVR
• Intermediate-risk for operative complications defined as either of:
  • STS risk score ≥ 4 (out of 100, higher is greater surgical risk)
  • STS risk score < 4 with heart team determination of intermediate-risk due to comorbidities not represented in STS risk score

Exclusion Criteria

• Unicuspid, bicuspid, or non-calcified AV
• MI in the prior month or CAD requiring revascularization
• Stroke or TIA in prior 6 months
• AS and aortic regurgitation mixed disease with regurgitation with severity >3+
• Invasive cardiac procedures in the prior 30 days
• DES placed in prior 6 months
• Complex CAD (unprotected LM or SYNTAX score > 32)
• Any prior prosthetic valve or ring
• Severe mitral annular calcification
• MR with severity >3+
• Leukopenia, anemia, thrombocytopenia, coagulopathy, or bleeding diathesis
• Requirement for inotropes or devices to support mechanical hemodynamics
• Need for emergency surgery
• HCM with or without obstruction
• LVEF <20%
• Thrombus, vegetation, or intracardiac mass
• PUD or UGIB in previous 3 months
• ASA, heparin, ticlopidine, clopidogrel, or contrast media sensitivities that would not be amenable to pre-medication
• Aortic annulus <18 mm or >25 mm
• Creatinine >3 mg/dL or RRT
• Life expectancy <2 years due to non-cardiac conditions
• Aortic disease
• Participation in another study for medications or devices
• Active bacterial endocarditis within 6 months

Baseline Characteristics

From the TAVI group.

• Demographics: Age 82 years, male 54.2%
• Risk Indices: STS score 5.8
• Cardiac History: NYHA II-III 77.3%, CAD 69.2%, Previous MI 18.3%, Previous CABG 23.6%, Previous PCI 27.1%, AF 31.0%, PPM 11.7%, LVEF 56.25%, moderate or severe MR 16.8%
• Other Comorbidities: Cerebrovascular disease 32.1%, PAD 27.9%, COPD 31.8%, CKD 5.0%, Frailty 44.4%, Liver disease 1.9%
• Aortic Valve: AVA 0.7cm2, Mean gradient 44.9mmHg

Interventions

• At enrollment, patients underwent mapping to determine candidacy for transfemoral vs. transthoracic TAVI approach
• Subjects then underwent 1:1 randomization stratified by planned approach (transfemoral vs. transthoracic) to:
  • TAVI - Esophageal echocardiographically-guided, percutaneous placement of a trileaflet bovine valve with stainless steel support frame deployed by balloon during rapid RV pacing
    • Heparin was used periprocedurally
    • ASA and clopidgorel were continued for at least 1 month
  • Surgical AVR - Surgical aortic valve replacement
• Major outcomes assessed at 30 days, 1 year, and 2 years by blinded adjudicators
• Patients could undergo concomitant CABG or PCI for CAD requiring revascularization at discretion of heart team

Outcomes

Presented as TAVI vs. surgical AVR.

Primary Outcome

All-cause mortality or disabling stroke at 2 years

30 days: 62 (6.1%) vs. 80 (8.0%) [p=0.11]

1 year: 145 (14.5%) vs. 160 (16.4%) [p=0.24]

2 years: 192 (19.3%) vs. 202 (21.1%) [p=0.33]

Secondary Outcomes

All-cause mortality

30 days: 39 (3.9%) vs. 41 (4.1%) [p=0.78]

1 year: 123 (12.3%) vs. 124 (12.9%) [p=0.69]

2 years: 166 (16.7%) vs. 170 (18.0%) [p=0.45]

Disabling stroke

30 days: 32 (3.2%) vs. 43 (4.3%) [p=0.20]

1 year: 49 (5.0%) vs. 56 (5.8%) [p=0.46]

2 years: 59 (6.2%) vs. 61 (6.4%) [p=0.83]

Any stroke

30 days: 55 (5.5%) vs. 61 (6.1%) [p=0.57]

1 year: 78 (8.0%) vs. 79 (8.1%) [p=0.88]

2 years: 91 (9.5%) vs. 85 (8.9%) [p=0.67]

Major vascular complication

30 days: 80 (7.9%) vs. 51 (5.0%) [p=0.008]

1 year: 84 (8.4%) vs. 54 (5.3%) [p=0.007]

2 years: 86 (8.6%) vs. 55 (5.5%) [p=0.006]

Life threatening or disabling bleeding

30 days: 105 (10.4%) vs. 442 (43.4%) [p<0.001]

1 year: 151 (15.2%) vs. 460 (45.5%) [p<0.001]

2 years: 169 (17.3%) vs. 471 (47.0%) [p<0.001]

Acute kidney injury

30 days: 13 (1.3%) vs. 31 (3.1%) [p=0.006]

1 year: 32 (3.4%) vs. 48 (5.0%) [p=0.07]

2 years: 36 (3.8%) vs. 57 (6.2%) [p=0.02]

New atrial fibrillation

30 days: 91 (9.1%) vs. 265 (26.4%) [p<0.001]

1 year: 100 (10.1%) vs. 272 (27.2%) [p<0.001]

2 years: 110 (11.3%) vs. 273 (27.3%) [p<0.001]

Subgroup Analysis

All analysis are for primary outcome of all-cause mortality

Access Approach (interaction p=0.06)

Transfemoral: 16.8% vs. 20.4% (HR=0.79; 95% CI 0.62-1.00)

Transthoracic: 27.7% vs. 23.4% (HR=1.21; 95% CI 0.84-1.74)

Criticisms

• Excluded specific groups that may benefit from TAVI, including those with CAD, severe PVD, and LVEF <20%
• Interaction between TAVI approach and primary endpoint composed of underpowered subgroups, potentially underestimating the heterogeneity between these populations.
• Study operators and patients not blinded to treatment allocation, but outcomes were assessed by blinded adjudicators.
• Although designed as a trial of intermediate-risk patients, 12% of the study population had an STS risk score > 8, which has been used to define high risk in previous studies, potentially biasing toward noninferiority of TAVI (TAVI is known to be effective in these patients).

Funding

• Study sponsor and maker of SAPIEN XT valve Edward Lifesciences designed, funded, participated in selection and management of sites, collection of data, and data monitoring.
• Study sponsor statistician performed the primary analysis.