Transcatheter Aortic-Valve Replacement in Low-Risk Patients at Five Years (PARTNER 3)
Authors, Journal, Affiliations, Type, DOI
- Mack MJ, Leon MB, Thourani VH, Pibarot P, Hahn RT, Genereux P, Kodali SK, Kapadia SR, Cohen DJ, Pocock SJ, Lu M, White R, Szerlip M, Ternacle J, Malaisrie SC, Herrmann HC, Szeto WY, Russo MJ, Babaliaros V, Smith CR, Blanke P, Webb JG, Makkar R; for the PARTNER 3 Investigators
- N Engl J Med. 2023;389:1949–60
- Affiliations: Baylor Scott and White Health (Mack); Columbia University / Cardiovascular Research Foundation (Leon, Hahn, Kodali, Smith); Marcus Heart Valve Center / Piedmont Heart Institute (Thourani); Emory University (Babaliaros); Laval University (Pibarot); St. Paul's Hospital, UBC (Blanke, Webb); Morristown Medical Center (Genereux); Cleveland Clinic (Kapadia); London School of Hygiene and Tropical Medicine (Pocock); Edwards Lifesciences (Lu, White); Cedars-Sinai Medical Center (Makkar); and others
- Type: Multicenter randomized controlled trial — 5-year pre-specified extension of the PARTNER 3 trial; 71 sites; open-label; 1:1 randomization
- DOI: 10.1056/NEJMoa2307447
- Funded by Edwards Lifesciences (sponsor participated in site selection, data collection, monitoring, and statistical analysis)
Overview
PARTNER 3 is a multicenter RCT comparing TAVR (SAPIEN 3 balloon-expandable valve, transfemoral) vs surgical aortic-valve replacement (SAVR) in 1,000 patients with severe symptomatic AS at low surgical risk (mean STS-PROM 1.9%). At 5 years, neither pre-specified primary endpoint reached statistical significance: the composite of death/stroke/rehospitalization was 22.8% (TAVR) vs 27.2% (SAVR; P=0.07), and the hierarchical win ratio was 1.17 (P=0.25). The early TAVR superiority seen at 1 and 2 years was attenuated by year 5, with late mortality trending numerically toward surgery in the landmark years 1–5 analysis. Valve durability (bioprosthetic-valve failure 3.3% vs 3.8%) was similar, while new-onset AF (13.7% vs 42.4%) remained markedly lower with TAVR and valve thrombosis (2.5% vs 0.2%) was more frequent.
Keywords
Transcatheter aortic-valve replacement; surgical aortic-valve replacement; aortic stenosis; low surgical risk; SAPIEN 3; bioprosthetic valve failure; valve durability; valve thrombosis; atrial fibrillation; PARTNER 3; win ratio; VARC-3
Key Takeaways
Study Design
- Multicenter RCT at 71 clinical sites; 1000 patients randomized 1:1; TAVR n=503 assigned (n=496 as-treated); SAVR n=497 (n=454 as-treated)
- TAVR: SAPIEN 3 balloon-expandable valve via transfemoral access
- SAVR: commercially available bioprosthetic valve (surgeon's choice)
- Eligibility: severe symptomatic AS + STS-PROM <4% + Heart Team approval + transfemoral anatomy suitable; bicuspid AV excluded
- Mean age 73 years; 69.3% male; mean STS-PROM 1.9%
- Follow-up at baseline, discharge, 30 days, 6 months, 1 year, then annually to 5 years
- Vital status available for 98.0% (TAVR) and 97.1% (SAVR) via vital-status sweep
Primary Endpoints at 5 Years
- First primary endpoint (nonhierarchical composite — death from any cause, stroke, or rehospitalization related to procedure/valve/heart failure): TAVR 22.8% vs SAVR 27.2%; difference −4.3 pp (95% CI −9.9 to 1.3; P=0.07; HR 0.79 [0.61–1.02])
- Second primary endpoint (hierarchical composite win ratio — death, disabling stroke, nondisabling stroke, rehospitalization days): win ratio 1.17 (95% CI 0.90–1.51; P=0.25)
- Neither endpoint met pre-specified superiority threshold; the trend favors TAVR driven by lower rehospitalization
- Restricted mean event-free survival longer by 103 days (95% CI 26–180) with TAVR
Components of the First Primary Endpoint
- Death from any cause: TAVR 10.0% vs SAVR 8.2% (odds ratio 1.24 [0.79–1.97]; NS; nonproportional hazards confirmed)
- 48 TAVR deaths (26 CV, 22 non-CV) vs 34 SAVR deaths (21 CV, 13 non-CV)
- 3 TAVR vs 1 SAVR death from Covid-19
- Stroke: TAVR 5.8% vs SAVR 6.4% (HR 0.87 [0.51–1.48]; NS); most strokes ischemic in origin
- Rehospitalization: TAVR 13.7% vs SAVR 17.4% (HR 0.75 [0.54–1.05]; NS)
Temporal Dynamics — The Attenuation Signal
- At 1 year: composite endpoint significantly favored TAVR (previously published)
- At 5 years: initial superiority lost; difference no longer significant (P=0.07)
- Landmark analysis years 1–5: TAVR 15.7% vs SAVR 13.7% (HR 1.17 [0.81–1.70]) — event accumulation now trends toward TAVR being worse after year 1
- More deaths occurred in TAVR arm from year 1 to 5, from both CV and non-CV causes
- Whether Covid-19 disproportionately contributed cannot be definitively determined
Secondary Endpoints
- New-onset atrial fibrillation: TAVR 13.7% vs SAVR 42.4% — dramatically and persistently lower with TAVR
- Permanent pacemaker implantation: TAVR 13.5% vs SAVR 10.4% — more frequent with TAVR
- Clinically significant valve thrombosis (VARC-3): TAVR 2.5% (n=12) vs SAVR 0.2% (n=1)
- Hemodynamic staging: Stage 1 (no deterioration) in 4, Stage 2 (moderate) in 5, Stage 3 (severe) in 3 TAVR patients
- None of the 13 thrombosis patients died; 7 had dyspnea/exertional symptoms; 3 had stroke (1 disabling, 2 nondisabling)
- Possible contribution from differential anticoagulation patterns in early post-procedure period (speculative)
- Serious bleeding: TAVR 49 patients vs SAVR 64 patients — less with TAVR
- Aortic-valve reintervention: TAVR 2.2% vs SAVR 2.6% (similar)
Valve Hemodynamics and Durability at 5 Years
- Mean aortic-valve gradient: TAVR 12.8±6.5 mmHg vs SAVR 11.7±5.6 mmHg (stable from year 2; similar)
- Mean aortic-valve area: TAVR 1.9±0.5 cm² vs SAVR 1.8±0.5 cm²
- Bioprosthetic-valve failure (VARC-3, any cause): TAVR 3.3% vs SAVR 3.8% (HR 0.86 [0.42–1.77]; NS)
- Irreversible stage 3 hemodynamic deterioration: TAVR 1.1% vs SAVR 1.0%
- Structural valve deterioration alone: TAVR 1.4% vs SAVR 2.0%
- Valve-related death: TAVR 0.0% vs SAVR 0.2%
- Paravalvular AR (mild or greater): TAVR 20.8% vs SAVR 3.2%
- Total AR (mild or greater): TAVR 24.5% vs SAVR 6.3%
- Mild paravalvular AR at 30 days was NOT associated with higher 5-year mortality in TAVR group (HR 0.78 [0.42–1.45])
- Alive with normally functioning valve: TAVR 86.3% vs SAVR 87.4% (similar)
Functional and Health Status
- NYHA class I or II at 5 years: TAVR 84.4% vs SAVR 86.0%
- KCCQ-OS score at 5 years: TAVR 86.2 vs SAVR 85.9 (similar; both within MCID of each other)
- Alive with KCCQ-OS ≥75 at 5 years: TAVR 71.0% vs SAVR 71.9%
- TAVR had larger QoL gains at year 1 (attributable to less-invasive procedure and shorter recovery); this advantage fully converged by year 5
Limitations of the Document
- Carefully defined trial population: excluded poor transfemoral access, bicuspid AV, and other anatomical/clinical exclusions — limits generalizability to broader clinical populations
- Disproportionate loss to follow-up in SAVR group (TAVR 94.6% vs SAVR 88.3% at 5 years); vital-status sweep partially mitigated mortality bias but cannot correct underreporting of nonfatal events
- Missing NYHA class, KCCQ, and echocardiographic data at later timepoints could not be fully accounted for with multiple imputation
- Covid-19 pandemic follow-up overlap; 3 TAVR vs 1 SAVR Covid-19 deaths; contribution to late mortality trend uncertain
- Funded by Edwards Lifesciences; sponsor involved in data collection and statistical analysis
- Single SAPIEN 3 device — results are not necessarily applicable to other TAVR platforms or self-expanding valves
Key Concepts Mentioned
- concepts/TAVI — primary subject: TAVR (SAPIEN 3) vs SAVR 5-year outcomes in low-risk patients; durability data
- concepts/Aortic-Stenosis — underlying disease; intervention selection and timing in low-surgical-risk patients
- concepts/Structural-Valve-Deterioration — bioprosthetic-valve failure (VARC-3), valve thrombosis, paravalvular AR, and valve durability data at 5 years
Key Entities Mentioned
- Edwards Lifesciences — trial sponsor; manufacturer of SAPIEN 3 transcatheter heart valve
Wiki Pages Updated
wiki/sources/tavr-partner3-5yr-nejm-2023.md— created (this file)wiki/concepts/TAVI.md— added PARTNER 3 5-year data; updated contradictionswiki/concepts/Aortic-Stenosis.md— updated TAVI vs SAVR section with 5-year specificswiki/concepts/Structural-Valve-Deterioration.md— added PARTNER 3 valve thrombosis and durability datawiki/sourceindex.md— added entrywiki/wikiindex.md— updated TAVI concept row