We thank Bellelli et al for their interest in our article. We agree that nursing home residents are atypical of participants in randomized controlled trials and patients in routine practice. This is why we excluded them in our primary analysis and headline results. However, for completeness, we did include them in a sensitivity analysis and discovered this did not affect our overall findings.
The comment on the similar rates of stroke and bleeding is irrelevant. The correct comparison is with what these rates would have been if the patients were not taking anticoagulation medication (4.5% for stroke and 1% for major bleeding, based on randomized controlled trials1). Although bleeding risk increases with age, so too does the rate of ischemic stroke.1
Evans A, Kalra L. Anticoagulation Treatment in Disabled Patients—Reply. Arch Intern Med. 2002;162(1):108. doi: