A trial fibrillation (AF) increases the risk of stroke, disability, dementia, and death, with a characteristic profile of more severe, disabling, and recurrent stroke compared with stroke without AF. [1][2][3][4] With aging populations, the prevalence of AF is projected to increase ≥2-fold by 2050, accompanied by an increase in the frequency of AF-associated stroke. 5 Accurate health economic data are important to inform health policy decisions to respond to this societal increase in AF prevalence, such as population screening programs. Reliable data on the cost of stroke associated with AF (AF-stroke) are also essential for rigorous cost-effectiveness studies of new oral anticoagulant agents, which guide reimbursement decisions for healthcare providers internationally. [6][7][8][9] Existing hospital-based studies have described high costs of stroke associated with AF. [10][11][12][13][14][15] However, existing studies have been limited by the reporting of acute hospital costs only, or by the inclusion of selected AF-stroke subgroups, such as hospitalized patients, nonaphasic patients, those with first-ever stroke only, or 1-year survivors. No cost studies of AF-stroke have been performed in unselected patients, including costs of community healthcare and indirect costs associated with lost productivity, leading to substantial underestimation of the Background and Purpose-No economic data from population-based studies exist on acute or late hospital, community, and indirect costs of stroke associated with atrial fibrillation (AF-stroke). Such data are essential for policy development, service planning, and cost-effectiveness analysis of new therapeutic agents. Methods-In a population-based prospective study of incident and recurrent stroke treated in hospital and community settings, we investigated direct (healthcare related) and indirect costs for a 2-year period. Survival, disability, poststroke residence, and healthcare use were determined at 90 days, 1 year, and 2 years. Acute hospital cost was determined using a case-mix approach, and other costs using a bottom-up approach (2007 prices). Results-In 568 patients ascertained in 1 year (2006), the total estimated 2-year cost was $33.84 million. In the overall sample, AF-stroke accounted for 31% (177) of patients, but a higher proportion of costs (40.5% of total and 45% of nursing home costs). On a per-patient basis compared with non-AF-stroke, AF-stroke was associated with higher total (P<0.001) and acute hospital costs (P<0.001), and greater nursing home (P=0.001) and general practitioner (P<0.001) costs among 90-day survivors. After stratification by stroke severity in survivors, AF was associated with 2-fold increase in costs in patients with mild-moderate (National Institutes of Health Stroke Scale, 0-15) stroke (P<0.001) but not in severe stroke (National Institutes of Health Stroke Scale ≥16; P=0.7). Conclusions-In our population study, AF-stroke was associated with substantially higher total, acute hospital, nursing home, and general practitioner cos...