High-quality randomized trials demonstrate that warfarin therapy dramatically reduces the risk for ischemic stroke associated with atrial fibrillation (1). These trials and observational studies also indicate that the benefit of anticoagulation is markedly reduced at INRs lower than 2.0 (2). Our study showed that the risk for intracranial hemorrhage increases with age and with INRs above 3.5. However, patients receiving anticoagulation did not have a reduced risk for intracranial hemorrhage at INRs less than 2.0. Thus, it appears that maintaining the INR in the 2.0 to 3.0 range maximizes the benefits of warfarin while minimizing the risks.

Unfortunately, older patients are at higher risk for both ischemic stroke and intracranial hemorrhage; erratic anticoagulation exacts a greater penalty. As Dr. Ruiz-Ruiz notes, numerous factors can lead to difficult warfarin management in the elderly, including polypharmacy, multiple comorbid conditions, and physical and mental frailty. We agree that the anticoagulation decision must be individualized and must engage the patient or patient caregiver. Appropriate decision making should account for whether patients can be safely maintained within a therapeutic INR range of 2.0 to 3.0.

Although intracranial hemorrhage risk increases at older ages, other validated clinical predictors of intracranial hemorrhage are few. As a consequence, individualized risk assessment often represents guesswork. The preponderance of evidence favors the use of warfarin in elderly patients with atrial fibrillation. However, the large proportion of elderly patients with atrial fibrillation, their increased risk for intracranial hemorrhage, and the devastating consequences of intracranial hemorrhage in patients taking warfarin all highlight the need to find better predictors of this condition. Such knowledge would make individualized decisions about anticoagulation in atrial fibrillation much more rational and effective.