Anticoagulation Should Not Be Used in Most Patients With Stroke With Infective Endocarditis
One of the most worrisome consults to the stroke service is the patient with neurological symptoms in the setting of infective endocarditis (IE). Stroke is the main neurological complication of IE and in 50% to 75% is the presenting feature. The most effective strategy for the prevention of a first or recurrent stroke is the prompt institution of appropriate antibiotic therapy, which reduces the risk to 1% to 3% within the first week.1 Ischemic stroke is 3-fold more common than hemorrhagic stroke. Embolic patterns are seen on diffusion-weighted MR in 92% of patients with stroke or encephalopathy. Initiating anticoagulation may seem reasonable to prevent embolization of infected or bland platelet–fibrin valvular vegetations—perhaps enhanced in the presence of antiphospholipid antibodies—and it is the standard of care for most cardioembolic stroke; however, the stroke risk is similar in the presence or absence of anticoagulants at onset of IE. The most compelling reason to avoid anticoagulation in Staphylococcus aureus IE is the predominance of early intracranial hemorrhage and hemorrhagic conversion of ischemic stroke.1–3 In Tornos' study of IE from S. aureus, the risk of stroke with native valve endocarditis was 22%; all were ischemic at onset and 57% became hemorrhagic in the absence of anticoagulation. Patients with prosthetic valve endocarditis were converted from oral anticoagulants to intravenous heparin on admission. Their risk of stroke was substantially higher at 52%; half were hemorrhagic at onset and the remainder converted within 48 to 72 hours. In Heiro's study, the 90-day mortality rate with S. aureus IE was 57% of those on anticoagulant therapy versus 20% those without. Typically patients with intracranial hemorrhage were not offered cardiac surgery and died. Anticoagulants should not be initiated for patients with IE with the goal of reducing the risk of stroke. Anticoagulants should be stopped as soon as a diagnosis of IE is suspected, particularly with neurological symptoms to suggest embolism or if S. aureus infection is possible. If these concerns can be excluded, a short-acting anticoagulant such as intravenous heparin can be initiated while anticipating a surgical decision.
The poor outcome in S. aureus IE and hemorrhagic stroke could be improved by identifying a lesion amenable to endovascular or surgical treatment. Ruptured mycotic aneurysms, documented in <2% of clinical series and 5% to 10% of tissue specimens, are caused by a septic arteritis resulting from embolization to the arterial wall. These lesions are best detected by the “gold standard” cerebral angiogram. Less invasive studies such as contrast-enhanced CT brain scans may be preferred in critically ill patients but only detect 20% of angiographically documented mycotic aneurysms.
Prosthetic valve endocarditis due to S. aureus is more likely to be complicated by central nervous system complications, heart failure, and persistent bacteremia despite adequate antibiotic therapy. Our patient presents with several of the risk factors for mortality from IE, a neurological complication, infection with S. aureus, and prosthetic valve endocarditis. Additional risk factors include intracranial hemorrhage, advanced age, septic shock, cardiac complications, and the need for urgent surgery. Although surgical risk is increased, multiple retrospective studies have demonstrated a survival benefit with surgery. The risk of stroke for noninfected cardiac valves ranges from 6% to 12% with cardiopulmonary bypass and includes embolization, hypoperfusion, heparin-related hemorrhagic complications, and cerebral edema from disruption of the blood–brain barrier. For these reasons, some advocate delaying surgery for 2 to 3 weeks for IE complicated by stroke. Recent series indicate that early surgery can be accomplished with a similar risk in patients with an otherwise uncomplicated partial and complete middle cerebral artery territory ischemic stroke. Patients with intracranial hemorrhage fare poorly with a surgical mortality risk of 30% to 40% and an even higher mortality rate with medical therapy alone.3,4
Because patients with stroke and S. aureus prosthetic valve endocarditis can be precipitously unstable and progress before the consult note is typed, our responsibility as a consultant to the cardiac team must be more than making a diagnosis and documenting the neurological deficit. The benefits of diffusion-weighted MRI as the more sensitive modality for detecting cerebral embolism in patients with stroke symptoms or encephalopathy must be weighed against the challenges of monitoring a potentially hemodynamically unstable patient. We should counsel against antithrombotic therapies and advocate for urgent cerebral angiography in the setting of intracranial hemorrhage or cerebral embolism with a plan for definitive endovascular treatment of any mycotic aneurysm. We should provide a neurological preoperative clearance, encouraging surgery when indicated while a stroke is still ischemic and otherwise uncomplicated lest delays lead to hemorrhagic transformation that substantially increases the risk of death with surgery and medical therapy.4,5
The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association. This article is Part 2 in a 3-part series. Parts 1 and 3 appear on pages 1795 and 1799, respectively.
- Received March 10, 2011.
- Revision received April 6, 2011.
- Accepted April 8, 2011.
- © 2011 American Heart Association, Inc.