Background and Purpose—Cerebral microbleeds (CMBs) predispose patients to intracerebral hemorrhage. Preclinical models to examine the effects of antithrombotic treatments on the development of clinically overt intracerebral hemorrhage are needed. We examined the natural course of CMB development and the effects of long-term anticoagulation with warfarin or dabigatran on cerebral micro- and macrohemorrhage in mice overexpressing the APP23 (amyloid precursor protein).

Methods—Repeated susceptibility-weighted magnetic resonance imaging was performed in APP23 mice at the age of 18 and 21 months, respectively. After establishing stable long-term anticoagulation effects of warfarin and dabigatran on number and total volume of CMBs, the outcome parameters were compared with nonanticoagulated control.

Results—CMBs were equally located in lobar and deep brain regions, and number and total volume of CMBs increased over time. Anticoagulation with either warfarin or dabigatran did not increase CMBs in APP23 significantly. Mice treated with warfarin numerically had a higher mortality (nonanticoagulated: 31%; dabigatran: 35% versus warfarin: 55%; P=0.21). In postmortem brains of prematurely dying animals warfarin caused significantly more frequently large intracerebral hemorrhage than control and dabigatran.

Conclusions—Anticoagulation with warfarin or dabigatran for 3 to 4 months does not promote the formation of CMBs in aged APP23 mice. Nevertheless, warfarin but not dabigatran is associated with a higher risk of extensive intracerebral hemorrhage, suggesting that this model may allow preclinical safety evaluation of antithrombotic therapies.