LAAC vs. DOACs: Long-Term Results of PRAGUE-17

Quick Takes

  • Among patients with atrial fibrillation at elevated risk of stroke and bleeding, percutaneous left atrial appendage closure is associated with similar rates of a combined endpoint of stroke, cardiovascular death, and bleeding compared to treatment with a direct oral anticoagulant at 3.5 years of follow-up.

Study Questions:

What are 4-year outcome results of PRAGUE-17 (Left Atrial Appendage Closure vs. Novel Anticoagulation Agents in Atrial Fibrillation) trial?


PRAGUE-17 was a randomized noninferiority trial comparing percutaneous left atrial appendage closure (LAAC; Watchman or Amulet) with direct oral anticoagulants (DOACs) in patients with nonvalvular atrial fibrillation (AF) and with increased risk of cardioembolism and bleeding. The primary endpoint was a composite of cardioembolic, cardiovascular death, clinically relevant bleeding, or procedure-/device-related complications.


A total of 402 patients with AF were randomized (average CHA2DS2-VASc score was 4.7 and HASBLED was 3.1). After 3.5-year median follow-up, LAAC was noninferior to DOACs for the primary endpoint (p for noninferiority = 0.006). The corresponding subdistribution hazard ratios were 0.68 for cardiovascular death, 1.14 for all stroke/transient ischemic attack, 0.75 for clinically relevant bleeding (p nonsignificant for all), and 0.55 (p = 0.039) for nonprocedural clinically relevant bleeding. The primary endpoint outcomes were similar in the per-protocol and on-treatment analyses.


At 3.5 years of follow-up, LAAC remains noninferior to DOACs for preventing major cardiovascular, neurological, or bleeding events.


Prior studies have demonstrated that LAAC was noninferior to warfarin and noninferior to DOACs at short- and medium-term follow-up for preventing major neurological, cardiovascular, or bleeding events in patients with AF. The original PRAGUE-17 trial showed noninferiority of LAAC to DOACs at a median follow-up of 20 months in terms of a combined composite endpoint of stroke, transient ischemic attack, systemic embolism, cardiovascular death, and major or nonmajor clinically relevant bleeding. The present report shows that this noninferiority persists at least up to 3.5 years. Unfortunately, PRAGUE-17 was not powered to detect differences between individual components of the composite endpoint. The composite endpoint treats stroke and death on the same level as major and even nonmajor “clinically significant” bleeding, rendering the results harder to interpret.

Keywords: Anticoagulants, Arrhythmias, Cardiac, Atrial Appendage, Atrial Fibrillation, Embolism, Hemorrhage, Ischemic Attack, Transient, Neurology, Secondary Prevention, Stroke, Vascular Diseases, Warfarin

< Back to Listings