As more people are living longer lives, the number of older adults with atrial fibrillation (AF), a common age-associated condition, will increase. Advanced age coincides with changes in kidney and hepatic function, body mass and composition, and use of concomitant therapies with the potential for drug interactions. In recent years, helpful information for age-specific dosing has become more common, older adults are more likely to be included in clinical trials for AF, and subgroup analyses are routinely performed to confirm results apply to older subgroups. These shifts are timely and welcome. However, there remains a sizable group of elderly patients with frailty at risk for stroke from AF who are not prescribed anticoagulants. Hesitancy stems from concern about bleeding that outweighs any concern for stroke or systemic embolism. Independent predictors of not using anticoagulation include prior bleeding, history of fall, older age, polypharmacy, and frailty. [1][2][3] Many older adults with AF also have 1 or more geriatric syndromes (ie, falls, loss of independence, incontinence, and cognitive impairment), which also make it less likely they will be prescribed anticoagulation. 4 Older adults with frailty are mostly excluded from clinical trials but represent a unique group where decisions are fraught with angst. The stakes are high for stroke and for bleeding, but gaps in evidence tilt decisions toward the conservative path-not adding 1 more medication. So what evidence is needed to inform use of anticoagulation in a nonagenarian with frailty unlike those in the pivotal trials?The pivotal trials for direct oral anticoagulants, such as the Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation-Thrombolysis in Myocardial Infarction 48 (ENGAGE AF-TIMI 48) trial, 5,6 which studied edoxaban, had no age exclusions and a sizeable proportion of participants were 75 years and older (approximately 40%). ENGAGE AF-TIMI 48 randomized patients into 3 arms: high-dose edoxaban (60 mg), low-dose edoxaban (30 mg), or warfarin. There were no interactions by age or weight in ENGAGE AF-TIMI 48, 5,7 nor in a large metaanalysis that looked across direct oral anticoagulants. 8 Statistical methods can confirm directionally similar findings in age groups, but age in trial populations is an imprecise proxy for age-associated vulnerability in practice. The protocol also recommended dose reduction if anticipated drug exposure would be increased based on reduced creatinine clearance or low body weight, and the majority of patients who met these dose reduction criteria were 75 years or older. 6 Extrapolations of data from trials is more complicated owing to dose adjustment based on subgroup criteria or within subgroups that include age. Prevention of stroke