Anticoagulation in Patients With Stroke With Infective Endocarditis
The Sword of Damocles
Stroke complicating infective endocarditis (IE) poses a therapeutic dilemma, particularly in patients with prosthetic valves (PV). Most experts recommend against the use of anticoagulation because of increased risk of intracerebral hemorrhage (ICH). Some worry that stopping anticoagulation increases the risk of clot formation on the PV and further embolization. Brain embolization in the setting of IE may result in a wide spectrum of clinical scenarios with different risk/benefit profiles to anticoagulation. Dr Rasmussen, a cardiologist, states that the potential harmful effect of anticoagulation is overestimated and that the beneficial effect of ongoing anticoagulation to prevent early recurrent stroke counterbalances the low risk of ICH. He, therefore, recommends in our case switching from warfarin to heparin due to the likelihood of urgent surgery. Conversely, Dr Sila, a neurologist, argues that anticoagulation should not be initiated and patients already on anticoagulantion should have it stopped as soon as a diagnosis of IE is suspected given the low rate of early recurrent stroke and high incidence of early ICH and hemorrhagic transformation (HT) in patients with stroke with Staphylococcus aureus IE. These diametrically opposed perspectives by our experts may reflect not only divergent approaches to different stages in the natural history of IE, but also “consultation bias” with a broader spectrum of IE seen by cardiologists, including less severe or even neurologically asymptomatic patients and more selected, often severely affected, patients seen by neurologists.
The beneficial or deleterious effect of anticoagulation in patients with IE is determined by a multitude of clinical, bacteriologic, radiological, and echocardiographic variables that may tilt the balance of the risk toward early recurrent stroke or ICH. These include native versus PV IE, size of the vegetation and its location on the mitral or aortic valve, virulence of the infective organism, size of the infarct(s), and presence of HT or mycotic aneurysms. The evidence against anticoagulation is largely anecdotal and based on retrospective nonrandomized studies reported in the late 1990s, including mainly severe patients with clinical evidence of IE, showing an increased risk of hemorrhagic complications and mortality especially in those patients on anticoagulants.1 On the other hand, more recent prospective studies show no association between anticoagulation and ICH. However, these studies were nonrandomized and included less severe patients with a high rate of native valve IE diagnosed at an early stage, which may have resulted in a lower incidence of ICH.
Rapid institution of effective antibiotic therapy represents the cornerstone of medical treatment of IE to reduce the mortality and morbidity from embolic complications and heart failure. Although the risk of early recurrent stroke is 1% to 3% in the overall IE population, it is certainly higher in patients with PV. Antibiotic therapy may reduce vegetation size; however, its efficacy in decreasing further embolism in patients with PV endocarditis may be limited. Although there is little evidence that anticoagulation reduces the risk of embolization in patients with native valve, the data are conflicting for PV IE. Our protagonists agree that surgery, when warranted, should be performed as soon as possible because the risk of complications is comparable to those operated 2 to 3 weeks later; we concur.
The decision of anticoagulation before valve replacement is a real Damocles' sword over our patient. Risk stratification and decision-making should be based on the individual risk balance of potentially devastating HT and further embolic ischemic stroke because S. aureus is the most aggressive form of IE in terms of both ischemic and hemorrhagic stroke risk. Potential mechanisms for ICH in patients with IE include HT of the ischemic infarct, rupture of mycotic aneurysms, or erosion of the septic arteritic vessels, the former being the most common. Although spontaneous HT after reperfusion is part of the natural history of cardioembolic stroke, symptomatic and in some cases devastating HT is more frequently seen in patients with IE, especially when S. aureus infection and anticoagulation concur. The risk of HT mainly depends on the size of the infarcted tissue. Therefore, it seems reasonable to withdraw anticoagulation, at least temporarily, in patients with large territorial infarctions. However, patients with smaller infarcts are also at risk of ICH. Numerous cerebral microbleeds seen on T2*-weighted MRI are highly prevalent in patients with IE, which may reflect a subacute microvascular inflammation that heralds mycotic aneurysms.2 Therefore, emergent MRI evaluation, including diffusion-weighted imaging, MR angiography, and T2*-weighted sequences, may help to identify patients with IE at risk of hemorrhagic complications. On the other hand, extensive echocardiographic evaluation may help to identify those at risk of early recurrent embolism who require continuing anticoagulants. Large (>10 mm) and mobile vegetations have been associated with multiple brain and spleen embolisms.
We agree that the risk of devastating ICH due to anticoagulation after IE-related stroke, particularly when small, may be overestimated; however, it is not nil. Therefore, we support a more conservative approach unless there is overwhelming evidence for high risk of recurrent embolization. There is certainly a need for randomized studies in patients with IE with stroke to determine the risk/benefit of anticoagulation using multimodal MRI and echocardiography for patient selection and stratification, but is this ever likely to occur?
The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association. This article is Part 3 in a 3-part series. Parts 1 and 2 appear on pages 1795 and 1797, respectively.
- Received April 4, 2011.
- Accepted April 6, 2011.
- © 2011 American Heart Association, Inc.