Ischaemic stroke and systemic embolism are the major potentially preventable complications of atrial fibrillation (AF) leading to severe morbidity and mortality. Anticoagulation using vitamin K antagonists (VKA) or non-vitamin K oral anticoagulants (NOACs) is mandatory for stroke prevention in AF. Following approval of the four NOACs dabigatran, rivaroxaban, apixaban, and edoxaban, the use of VKA is declining steadily. Increasing age with thresholds of 65 and 75 years is a strong risk factor when determining annual stroke risk in AF patients. Current recommendations such as the “2016 Guidelines for the management of atrial fibrillation” of the European Society of Cardiology and the “2019 AHA/ACC/HRS Focused Update” by the American College of Cardiology, the American Heart Association, and the Heart Rhythm Society strengthen the importance of anticoagulation and detection of bleeding risks, of which older age is an important one. While patients aged ≥ 75 years are usually underrepresented in randomised clinical trials, they represent almost 40% of the trial populations in the large NOAC approval studies. Therefore, a sufficient amount of data is available to assess the efficacy and safety for this patient cohort in that specific indication. In this article, the evidence for stroke prevention in AF using either VKA or NOACs is summarised with a special focus on efficacy compared to bleeding risk in patients aged ≥ 75 years. Specifically, we used a model of increased weighing of intracranial bleeding to illustrate the potential benefit of NOACs over VKA in the elderly population. In brief, there are at least two tested strategies with apixaban and edoxaban which even confer an additional clinical net benefit compared with VKA. Furthermore, elderly subgroups of trials for combined antithrombotic treatment following percutaneous coronary interventions in anticoagulated patients are analysed.