For the roughly one-third of atrial fibrillation patients who are also frail or over 75, the choice of blood thinner is not merely a pharmacological preference — it carries profound implications for stroke risk, bleeding events, and functional independence. This narrative review synthesizes pivotal trial data and major real-world registries to clarify which anticoagulation strategies actually serve these high-complexity patients.
The analysis draws on randomized clinical trial evidence and registry outcomes comparing direct oral anticoagulants (DOACs) against vitamin K antagonists (VKAs, primarily warfarin) in patients aged 75 or older or meeting clinical frailty criteria. Across this literature, DOACs consistently matched or exceeded VKA efficacy for stroke and systemic embolism prevention while delivering a meaningfully lower intracranial hemorrhage burden — a particularly consequential advantage given that elderly patients face steeper case-fatality rates from intracranial bleeding. Among the DOAC class, apixaban and edoxaban at dose-adjusted regimens emerged with the most favorable efficacy-to-bleeding tradeoff. A critical counterpoint, however, comes from the FRAIL-AF trial: switching clinically stable elderly patients who were already well-managed on long-term VKA therapy to a DOAC generated an early surplus of clinically relevant non-major bleeding events, complicating any blanket transition policy.
This review arrives at a moment when cardiologists face growing pressure to standardize anticoagulation across aging populations, yet the FRAIL-AF signal demands caution. The field has long assumed that newer agents are always preferable, but this data suggests that stable INR control in an elderly patient on warfarin may not warrant disruption. Frailty, polypharmacy, and declining renal function all modulate both drug metabolism and bleeding risk in ways that point-of-prescription algorithms rarely fully capture. The review's primary limitation is its narrative rather than systematic meta-analytic structure, which constrains the precision of pooled effect estimates. Nonetheless, its clinical synthesis reinforces that individualized risk stratification — not age alone — should drive anticoagulation decisions in this population. Incremental rather than paradigm-shifting, this work is nonetheless a clinically valuable consolidation of a complex evidence base.