Anticoagulation in patients with stroke with infective endocarditis.

Anticoagulation is a controversial issue in Staphylococcus aureus infective endocarditis (IE) because these patients are believed to be particularly susceptible to hemorrhagic transformation of embolic lesions. However, the evidence supporting the deleterious effect of anticoagulation is at best incomplete and the adverse effect of such treatment has been questioned by most recent research. An increasing number of patients with IE receive anticoagulant treatment because of mechanical prosthetic valves, atrial fibrillation, pulmonary embolism, and factor V Leiden mutation as well as other hypercoagulability disorders. These patients carry an increased risk of thromboembolism and the decision to terminate anticoagulant treatment should therefore balance the risks and benefits of such treatment. Cerebrovascular complications (CVCs) in IE are most often the consequence of septic embolization from vegetations located at the heart valves and occur in approximately 25% of the patients with S. aureus IE. CVCs cover a wide clinical spectrum, 1 of the most feared being cerebral hemorrhage, which is almost always associated with a poor outcome. Cerebral hemorrhage is mainly caused by hemorrhagic transformation of an ischemic stroke, whereas acute pyogenic arteritis or rupture of mycotic aneurysms is rare. In a series of studies conducted in the period 1975 to 1997, cerebral hemorrhage has been reported in 13% to 17% of the patients with S. aureus IE, which is 1 of the key arguments against anticoagulation in this group of patients.1 Thus, in a number of these studies, cerebral hemorrhage has been linked to uncontrolled infection and anticoagulation, whereas other studies have failed to find such an association. For example, in a study by Wilson et al evaluating the impact of anticoagulation in patients with prosthetic valve endocarditis, the authors showed that the incidence of CVC and death was higher if anticoagulation was discontinued supporting the use of anticoagulation in this group of patients.2 Unfortunately, these studies share a common set of limitations because they all have been conducted in an era before the introduction of the Duke criteria with poor or no access to high-quality neuroimaging (CT or MRI) and echocardiography. It is generally accepted that the clinical presentation of IE may be suggestive but mostly is unspecific and that the diagnosis of IE primarily depends on blood cultures positive for microorganisms that commonly cause IE and echocardiographic findings. Accordingly, studies including patients only based on the clinical evidence of IE may be inclined to include the sickest patients with IE, thus overestimating the incidence of CVC and cerebral hemorrhage. A most recent study by the current author showed that although stroke was a common complication in patients with S. aureus IE, the incidence of cerebral hemorrhage was low (3%), and there was no association between this feared complication and anticoagulation.3 Contrary anticoagulation was associated with a reduction in vegetation size and CVC on admission in patients with native valve S. aureus IE.3 Because a vegetation by all means is a coagulum that forms on the damaged heart valve and consists of damaged tissue, bacteria, fibrin, and platelets, a protective