Percutaneous Left Atrial Appendage Closure: The CHAMPION-like vs CLOSURE-like Patient Phenotype Framework.

AIMS: The near-simultaneous publication of CHAMPION-AF and CLOSURE-AF produced apparently discordant evidence regarding percutaneous left atrial appendage closure (LAAC) for stroke prevention in atrial fibrillation. We aimed to reconcile these findings through a phenotype-based framework and identify the determinants of net clinical benefit. METHODS AND RESULTS: This narrative review integrates randomized trials, registries, consensus documents, and guidelines, focusing on patient selection, procedural risk, post-implant antithrombotic therapy, endpoint construction, competing mortality, and time to benefit. CHAMPION-AF evaluated LAAC against direct oral anticoagulation in an elective, anticoagulation-eligible population with lower bleeding risk, demonstrating non-inferior thromboembolic protection and less non-procedural bleeding. Conversely, CLOSURE-AF enrolled an older, multimorbid population with substantial renal dysfunction, previous bleeding, and high competing mortality. Although implantation success and stroke rates were comparable between treatment groups, early procedural and antithrombotic-related bleeding, a composite endpoint incorporating procedural events, and limited survival reduced the opportunity for delayed benefit. The highly selective recruitment pathway further limits generalisability. These findings define a continuum from the "CHAMPION-like" patient, with sufficient physiological reserve and life expectancy to realise long-term benefit, to the "CLOSURE-like" patient, in whom frailty and competing risk may outweigh the advantage of appendage exclusion. CONCLUSION: LAAC efficacy is not determined by device performance alone but by matching the intervention to patient phenotype and time horizon. Selection should integrate thromboembolic and bleeding risk, frailty, renal function, post-implant treatment tolerance, and expected survival. Future trials require transparent screening data, time-sensitive endpoints, landmark analyses, and competing-risk methods.