Atrial fibrillation (AF) is associated with a five-fold increased risk of cerebrovascular stroke. While the risk of stroke in patients with known AF can be calculated via AF stroke prediction risk scores, the prediction of AF occurrence per se in individual patients remains difficult. We will recruit 250 patients with an implanted dual-chamber Pacemaker (PM) or an implantable cardioverter Defibrillator (ICD) with atrial lead to ensure continuous rhythm monitoring during follow up. At baseline, we will gather clinical and device data of patients. For AF prediction based on surface Electrocardiography (ECG), 24-h Holter ECG monitoring will be performed. After 6 months, we will assess the occurrence of AF during the study period via interrogation of PM/ICD. The resulting data will be used to develop algorithms including clinical, device and ECG data for prediction of the development of AF in individual patients. If possible, we will develop a risk score of high accuracy by combination of demographical, clinical and technical parameters of device patients. The resulting risk score could potentially help to facilitate the decision if anticoagulation is necessary in patients with either risk of AF or embolic stroke of unknown origin. Furthermore, Hayn et al. (AIT Austrian Institute of Technology) are currently developing algorithms to predict the occurrence of AF surface ECG data. It is an additional aim of this project to support the development of this algorithm in pacemaker and ICD patients and to increase the accuracy of AF prediction with clinical parameters and other parameters available to patients with implanted pacemaker (PM) or implanted cardioverter-defibrillator (ICD).