Objectives: To better understand the tradeoffs between the efficacy of anticoagulation with warfarin and its side effects in the oldest old with nonrheumatic atrial fibrillation (AF).
Design: Cost-effectiveness analysis.
Setting: Published literature, including meta-analyses when available, and web-based sources.
Participants: Those aged 65 to 100 with AF.
Intervention: Anticoagulation with warfarin.
Measurements: Quality-adjusted life expectancy and cost.
Results: Anticoagulation is not effective in persons with AF who do not have other risk factors, even in the oldest old. The best argument for its use (prolongation of life at an acceptable cost) can be made in those at major risk for stroke because of previous stroke or transient ischemic attack, diabetes mellitus, and hypertension, but poor quality of life before anticoagulation and comorbidities that carry their own risks of dying diminish benefits. The decision to anticoagulate the oldest old with AF must take into consideration the risk of hemorrhagic stroke and of death from hemorrhagic stroke that existed before anticoagulation, the increased risk of hemorrhagic stroke and of death from hemorrhagic stroke while anticoagulated, and the efficacy of anticoagulation. Cost-effectiveness is also influenced by the cost of warfarin, the risk of major extracranial bleeding, the risk (natural and anticoagulated) of death from hemorrhagic stroke, the rate of ischemic stroke, the cost of major extracranial bleeding and hemorrhagic strokes, the cost of nursing home care, and the fraction of patients with stroke who need nursing home care.
Conclusion: There is no compelling evidence to date that anticoagulation prolongs quality-adjusted life expectancy in the oldest old with nonrheumatic AF. More studies that better estimate the risk of intracranial bleeding with and without anticoagulation in the oldest old are needed before recommendations can be made. The oldest old who are most likely to benefit are those who have a high risk of stroke secondary to risk factors other than age alone, although quality of life and life expectancy related to these risk factors limit obtained benefit. Recommendations that all older persons with AF should be anticoagulated are premature.