110 CHAPTER 7 to balance the different prescribed NOACs as much as possible during patient accrual. NOAC dosing and dose adjustments in principle followed the summary of product characteristics, unless the treating physician deliberately opted for a different dose. All patients were followed after one, three, six, nine, and twelve months by telephone interviews, and when the occurrence of any of our predefined outcomes was suspected, additional medical information was retrieved. Outcomes The primary outcome was the occurrence of a major or clinically relevant non-major (CRNM) bleeding complication (whichever came first). For bleeding complications, we used the definitions of the International Society of Thrombosis and Haemostasis.14,15 A major bleeding complication was defined as a fatal bleeding; any bleeding in a critical area or organ (intracranial, intraspinal, intraocular, retroperitoneal, intra-articular, pericardial, or intramuscular leading to a compartment syndrome); bleeding leading to a fall in haemoglobin level of ≥2 g/dL (1.24 mmol/L); or bleeding leading to a transfusion of ≥2 units of whole blood or red blood cells. A CRNM bleeding complication was defined as any bleeding not being major but including at least one of the following items: bleeding prompting a face-to-face consultation; bleeding requiring a medical intervention by a healthcare professional; or bleeding leading to hospitalisation or increased level of care. Secondary outcomes included all-cause mortality, major bleeding complications (separate from CRNM bleeding complications); CRNM bleeding complications (separate from major bleeding complications); the occurrence of all-cause thromboembolic events (ischaemic stroke; transient ischaemic attack; peripheral arterial thromboembolism); the composite of thromboembolic events and major or CRNM bleeding; and the composite of ischaemic and haemorrhagic stroke. Statistical analysis The yearly incidence of major and CRNM bleeding complications was assumed to be 10% to 15% in frail older patients with AF using a VKA.16 A relative reduction of 20% to 30% was expected on the occurrence of these bleeding complications when switching to a NOAC. At a 2-sided α-level of 0.05, a 1:1 allocation ratio and 1,250 patients in each treatment arm, the power was at least 0.80 if the incidence of major or CRNM bleeding complications on VKA treatment was between 11% (with an incidence of our composite outcome on NOAC treatment of 7%) and 15% (with an incidence of our composite outcome on NOAC treatment of 11.2%). All analyses were performed on an intention-to-treat (ITT) basis. In patients randomly assigned to the intervention arm, a variable amount of time occurred between the moment of randomisation and the actual start of the NOAC. In line with the ITT
RkJQdWJsaXNoZXIy MTk4NDMw