Daniel J Sexton, MD
Thomas L Ortel, MD, PhD
Lawrence LK Leung, MD
Stephen A Landaw, MD, PhD
INTRODUCTION — Anticoagulants were first used in patients with infective endocarditis (IE) over 60 years ago. Subsequently, several investigators reported that the use of heparin when combined with penicillin in the treatment of patients with IE was associated with a high incidence of hemorrhagic stroke. By the mid-1950s, most experts had concluded that the use of anticoagulants in patients with IE was excessively dangerous, and that its routine use should be abandoned.
However, an increasing proportion of patients with IE have infected prosthetic cardiac valves, many of whom require long-term anticoagulation therapy to prevent valve dysfunction. The therapeutic dilemma in the management of such patients is as follows:
- Continuing anticoagulant therapy in the face of a prosthetic valvular infection poses the risk of hemorrhagic transformation of an embolic stroke or accentuation of bleeding from septic arteritis or mycotic aneurysms should they occur.
- Stopping anticoagulant therapy introduces the risk of clots forming on the prosthetic valve, leading to the risk of further emboli or prosthetic value thrombosis.
(See “Risk of intracerebral hemorrhage in patients treated with warfarin” and “Antithrombotic therapy in patients with prosthetic heart valves”.)
The use of anticoagulation in patients with IE will be reviewed here. It should be emphasized that our recommendations and those included in published guidelines for the use of anticoagulation therapy in IE are primarily based on older studies, anecdotal evidence, and expert opinion. Thus, these recommendations use extrapolation of general medical principles and available information, rather than randomized controlled trials to reach conclusions [1,2]. Overall discussions of the treatment and complications of prosthetic valve endocarditis are presented separately. (See “Antimicrobial therapy of prosthetic valve endocarditis” and “Complications and outcome of infective endocarditis”.)
TREATMENT AND DISEASE RISKS — It is not uncommon for patients with IE to suffer an embolic stroke before the initiation of any form of therapy or within a few days following the diagnosis of IE . In such cases, clinicians are faced with the dual problem of treating a patient with IE and an acute or subacute stroke. In addition to producing cerebral infarction, infected emboli can rarely cause septic cranial arteritis, mycotic aneurysms, and brain abscesses .
Risk of embolic stroke in IE — The ability of echocardiography to identify patients with IE at risk for embolism is low for the following reasons [3-5]:
- More than 90 percent of patients with IE have valvular vegetations detectable by transesophageal echocardiography.
- Approximately 15 to 35 percent of all patients with IE develop clinically evident systemic emboli. If more sensitive tests such as magnetic resonance imaging are used, a much higher proportion of patients with IE have evidence of emboli. As an example, in one prospective study of 65 patients with left-sided endocarditis, 35 percent of patients had evidence of a symptomatic cerebrovascular complication; another 30 percent had clinically “silent” evidence of a cerebrovascular complication using magnetic resonance imaging of the brain .
Overall, the presence of a vegetation detected by echocardiography has little utility in assessing the actual risk that this vegetation will result in embolism. (See “Diagnostic approach to infective endocarditis”, section on ‘Role of TEE’.)
Even though the presence of vegetations by echocardiography is not particularly useful in predicting the risk of embolism in an individual patient, the following risk factors have been suggested as useful indicators of a higher than normal risk of embolism. (See “Complications and outcome of infective endocarditis”, section on ‘Predictors of embolization’.)
- Size of vegetation >10 mm
- Vegetation location on aortic or mitral valve, especially the anterior mitral leaflet
- Virulent organisms (eg, S. aureus)
- Visibility of the vegetation(s) on TTE and TEE rather than only on TEE
Although this list may be useful in assessing embolic risk in population-based studies, its utility in making decisions in an individual patient is limited. In addition, the results of some studies assessing the beneficial effect of anticoagulation in patients with IE are contradictory. As examples:
- One prospective cohort study of 175 patients with IE due to S. aureus concluded that patients receiving anticoagulation at the time of admission for IE were significantly less likely to have experienced a major cerebral event at the time of admission compared with those without anticoagulation (15 versus 37 percent; adjusted OR 0.27; 95% CI 0.075-0.96) . The decrease was primarily due to a decrease in cerebral thromboembolic episodes in those patients on anticoagulant therapy.
- A randomized, placebo-controlled study has shown that anti-platelet therapy with aspirin had no effect on the risk of stroke in patients with IE .
- A third study has shown that patients with IE due to S. aureus had an increased risk of stroke if they were concurrently receiving anticoagulants, although this was primarily due to hemorrhagic rather than thromboembolic stroke (discussed below) .
Native valve IE — There is no evidence that prophylactic anticoagulation reduces the incidence of emboli in patients with native valve IE. As a result, the 2008 ACCP Guidelines recommended against routine antithrombotic therapy unless a separate indication exists .
In contrast, there are compelling data to suggest that prompt use of antibiotics significantly reduces the incidence of emboli in patients with IE. This was illustrated in a report of patients with native valve IE, in whom the rate of embolization fell from 5 to 13 per 1000 patient days during the first week of antibiotic therapy to 1.2 to 1.7 per 1000 patient days after two weeks of therapy [10,11]. Similar findings were noted in another large case series in which a prompt reduction in emboli was seen shortly after antibiotic therapy was initiated . (See “Complications and outcome of infective endocarditis”, section on ‘Effect of antibiotic therapy on embolic risk’.)
Prosthetic valve IE — With the exception of patients with bioprosthetic valves who are in normal sinus rhythm, patients with prosthetic heart valves have an increased risk of thromboembolism that can be markedly reduced with appropriate anticoagulation. (See “Complications of prosthetic heart valves”, section on ‘Systemic embolization’.)
Patients with prosthetic valve IE have the same risk of cerebral emboli as other patients with IE. There is conflicting evidence on the effectiveness of anticoagulation in reducing this risk of embolization, as illustrated by the following observations:
- In a report of 52 patients with prosthetic valve IE, central nervous system embolic events occurred less frequently in patients who received adequate anticoagulation (3 of 38 versus 10 of 14, respectively) . A suggestion of similar benefit from anticoagulation was also noted in another series .
- However, anticoagulation either failed to reduce the incidence of emboli [3,15] or was associated with a higher risk of serious CNS bleeding in other studies [4,16,17].
As is the case with native valve IE, prompt antibiotic treatment appears to be more important in preventing neurologic complications in patients with prosthetic valve IE than anticoagulation .
Mechanisms of intracerebral hemorrhage — Three potential mechanisms exist for intracerebral hemorrhage in patients with IE:
- Hemorrhagic transformation of an ischemic infarct
- Septic erosion of arteritic vessels without aneurysm formation
- Rupture of mycotic aneurysms
Hemorrhagic transformation and septic arteritis are more common causes of intracerebral hemorrhage associated with IE than mycotic aneurysm rupture. The latter is relatively uncommon in the modern era of antimicrobial therapy [18,19]. Rupture of mycotic aneurysms typically causes intraparenchymal hemorrhage, but limited subarachnoid hemorrhage can also occur. (See “Overview of infected (mycotic) arterial aneurysm”.)
Septic arteritis without aneurysm, when present, typically occurs in the small cortical arterial branches located in the spaces of cortical sulci , and thus is generally invisible with CT or MR angiographic imaging. The appearance of arteritis is nonspecific at best and undetectable in many cases, even when conventional angiography is used. Thus, it is not surprising that no studies have evaluated the optimal treatment of septic arteritis in patients with IE.
Risk of anticoagulation — The hazards of anticoagulation were outlined in a study of 35 patients with native valve left-sided S. aureus IE and 21 patients with S. aureus left-sided prosthetic valve IE. None of the 35 patients with native-valve IE was receiving anticoagulants at the time of diagnosis; in contrast, 19 of 21 patients with prosthetic valve IE were receiving warfarin therapy at the time of diagnosis. Although the INR was not reported for these patients, the prothrombin time data suggested that they were not excessively anticoagulated . Heparin was substituted for warfarin during the hospitalization of these patients . The incidence of embolic episodes, including all major neurologic complications, was comparable in patients with native valve IE and patients with prosthetic valve IE (63 versus 67 percent). However, 6 of 12 patients with neurologic complications and concurrent prosthetic valve IE developed hemorrhagic brain lesions, all of whom died. In contrast, only one patient in the nonanticoagulated native valve group developed hemorrhagic transformation of a presumed embolic stroke.
These findings led the authors to conclude that the incidence of hemorrhagic transformation of embolic strokes was so high in patients with prosthetic valve S. aureus IE that anticoagulation therapy should be discontinued until the septic phase of the disease is overcome.
In contrast, another study found that anticoagulant therapy was not associated with an increased risk for cerebral hemorrhage . Out of 175 consecutive patients with S. aureus aortic and/or mitral valve IE patients, 40 were on anticoagulant therapy at the time of admission. Out of a total of four patients sustaining a hemorrhagic stroke at the time of admission, none was on anticoagulant therapy. One additional patient was diagnosed with cerebral hemorrhage during the course of hospitalization; this patient was on anticoagulation at the time of the event. These authors concluded that anticoagulation did not increase the risk of cerebral hemorrhage in S. aureus IE, and that it should be continued when clinically indicated .
Use of aspirin — Results of a few small studies have suggested that the use of aspirin might attenuate the virulence of S. aureus infections and reduce the frequency of embolic events [20,21]. Similarly, a large retrospective study found a reduced embolic risk among patients who received daily antiplatelet therapy before the onset of IE .
However, no benefit from aspirin was seen in a randomized trial involving 115 patients with IE . The rate of embolic events was not significantly different in patients treated with aspirin compared with those receiving placebo (28 versus 20 percent with placebo, odds ratio 1.6; 95% CI 0.68-3.9). There was a trend toward a higher incidence of bleeding in patients taking aspirin (odds ratio 1.9; 95% CI 0.76-4.9). However, there were no differences between the two groups in the development of new intracranial lesions, vegetation resolution, or valvular dysfunction. Concomitant use of systemic anticoagulation by 23 percent of the patients was a potentially important limitation of this study; most of the patients receiving anticoagulants had prosthetic heart valves.
Patients had been symptomatic for approximately one month before the diagnosis of IE and the subsequent use of aspirin in this study . In fact, aspirin might have been more effective in reducing embolic events if it had been started earlier in the course of IE. This supposition was examined in a retrospective analysis of data from the 84 patients ineligible for the randomized study because they were taking long-term aspirin (average daily dose 325 mg/day) prior to the diagnosis of IE. Results of this analysis included :
- The incidence of embolism in 84 patients taking long-term aspirin before the onset of IE as well as throughout the four-week observation period (21 percent) was similar to that seen in 55 patients randomly assigned to treatment with placebo (20 percent, unadjusted odds ratio 0.80; 95% CI 0.36-1.8).
- As also noted in the randomized study, there was a trend toward excess bleeding in long-term aspirin recipients compared with those receiving placebo (unadjusted odds ratio 2.3; 95% CI 0.97-5.7).
The authors of this study concluded that the need for continuation of aspirin should be carefully assessed in patients with a diagnosis of IE who are already taking aspirin. The risk versus benefit of continuing or discontinuing aspirin should be carefully weighed in order to minimize the potential risk of bleeding during the acute phase of IE, should aspirin be continued .
Based on the findings of the single randomized trial discussed above , we and others do not recommend routine use of aspirin to prevent embolism in patients with IE .
Mechanical heart valves — There is no strong consensus among experts regarding the decision to continue or temporarily discontinue anticoagulation when IE develops in patients already receiving anticoagulation for mechanical heart valves. Two divergent approaches have been recommended:
- Continue anticoagulation unless the patient develops a stroke — The rationale for this approach is based upon studies reporting a decreased rate of neurologic complications with adequate anticoagulation [13,14], as well as the well-known high risk of thrombosis of the mechanical valve if anticoagulation is stopped. Many experts believe that long-term anticoagulation therapy can be safely and appropriately continued in patients with prosthetic valve IE if there is no evidence of a CNS complication consistent with emboli. If anticoagulation is continued, some experts recommend converting from warfarin to intravenous heparin in order to achieve better control of anticoagulation during the active phase of infective endocarditis. (See ‘Patients with stroke’ below.)
- Temporarily discontinue anticoagulation — The rationale for this approach is based upon observational evidence from a single study suggesting an appreciable risk of fatal brain hemorrhage in patients with mechanical heart valves and left-sided S. aureus IE who are continued on anticoagulant therapy . Discontinuing anticoagulation until blood cultures are negative and the clinical signs of active infection or ongoing sepsis have resolved is a reasonable approach in such patients. However, this position is not universally held and there are no studies to guide the physician concerning the safest time to resume anticoagulation if this option is chosen.
The 2008 ACCP Guidelines suggest that the vitamin K antagonist (VKA) should be discontinued in patients with an infected prosthetic valve who are taking the VKA at the time of initial presentation. Unfractionated heparin should be substituted for a VKA until it is clear that invasive procedures will not be required and the patient has stabilized without signs of central nervous system involvement (eg, intracranial bleeding due to a ruptured mycotic aneurysm or hemorrhagic transformation of a bland infarct). VKA therapy can be reinstituted when the patient is deemed stable without contraindications or neurologic complications (Grade 2C) .
Patients with stroke — Anticoagulation should be temporarily discontinued in all patients with a mechanical valve and IE who develop central nervous system symptoms compatible with embolism or stroke, despite the potential risk of further thromboembolism from their prosthetic heart valves. Anticoagulant therapy is contraindicated if there is any evidence on imaging studies of hemorrhagic transformation of the stroke.
However, there are no studies that give guidance as to the best form of brain imaging to use in order to detect hemorrhagic transformation, nor are there published criteria that help determine when reinstitution of anticoagulation therapy will be safe after a stroke.
The optimal time to restart anticoagulation after stroke in patients with IE and mechanical heart valves is unknown. Some guidance may be found from data collected from patients without IE or hypertension who have an embolic stroke. If there is no evidence for hemorrhage on CT scan one to two days following the stroke, anticoagulation can be resumed immediately for smaller lesions, and within one week in those with large areas of infarction [1,24,25].
However, given the risk of inducing intracerebral hemorrhage with anticoagulation in patients with prosthetic heart valve IE and an embolic event, a more conservative approach seems appropriate for patients with endocarditis:
- If there is a compelling need to reinstitute anticoagulant therapy during the first two weeks following the onset of stroke, serial brain imaging with CT or MRI should be performed to exclude hemorrhagic transformation or intraparenchymal hemorrhage .
- Angiography, preferably noninvasive contrast angiography with CT, may be necessary to exclude a mycotic aneurysm in selected patients who have evidence of hemorrhage by routine brain imaging studies and who are candidates for definitive neurosurgical or embolization therapy [27,28]. If CT angiography does not show a mycotic aneurysm in such patients, a conventional invasive angiogram is required as the definitive test. If a symptomatic mycotic aneurysm is detected by angiography, resumption of anticoagulation should be postponed until definitive treatment has been achieved. (See “Overview of infected (mycotic) arterial aneurysm”.)
- Serial neuroimaging and angiography are not indicated for patients without hemorrhagic transformation who do not require anticoagulation.
Native or bioprosthetic valves — The indications for anticoagulation in a patient with either a native valve IE or bioprosthetic valve IE and cerebral embolism are not clear. Anticoagulation is not routinely recommended because there is persuasive evidence of the risk of intracerebral hemorrhage without solid countervailing evidence of benefit.
However, clinicians may be reasonably guided to use or withhold anticoagulation according to the following additional considerations. (See “Antithrombotic therapy in patients with prosthetic heart valves”, section on ‘Bioprosthetic valves’ and “Risk of intracerebral hemorrhage in patients treated with warfarin”, section on ‘ICH risk with antithrombotic therapy’.)
- The perceived risk of additional emboli
- The presence or absence of a left atrial thrombus
- The perceived or expected efficacy of the antibiotic treatment
- The presence of hypertension, atrial fibrillation, or other comorbid features (eg, venous thromboembolism)
Concurrent indication for anticoagulation — The use of anticoagulation in patients with IE and other concurrent conditions (eg, deep vein thrombosis, surgical repair or replacement of an infected valve) is unresolved. Anticoagulant therapy can reasonably be initiated in individual patients after brain imaging studies have excluded the presence of asymptomatic or subclinical stroke. However, in general, anticoagulation should be discontinued if signs of central nervous system embolism occur. An inferior vena caval filter is an option for selected patients with deep vein thrombosis who are not candidates for anticoagulation. (See “Inferior vena cava filters”.)
Heparin versus warfarin — There is an unresolved and ongoing controversy over whether heparin or warfarin should be used when anticoagulation is warranted for the treatment of patients with IE [3,13,14,29,30]. We generally prefer warfarin with an International Normalized Ratio (INR) maintained between 2.5 to 3.0. However, when there is increased concern about the risk of intracerebral bleeding, or for those patients requiring placement of central venous access or other procedures, heparin anticoagulation has the advantage of permitting more rapid reversal of anticoagulation.
Care must be taken with the use of warfarin in patients receiving antibiotics, as these may change the metabolism of this agent. As an example, rifampin induces cytochrome enzymes that accelerate the metabolism of warfarin. Thus, rifampin therapy will markedly increase the daily dose of warfarin required for effective anticoagulation, and the discontinuation of rifampin will necessitate a reduction in the daily dose. For some patients who will be on a prolonged course of antibiotics, conversion of the patient to therapeutic-dose low-molecular weight heparin may be an alternative strategy for effective, safe anticoagulation. (See “Therapeutic use of warfarin”, section on ‘Drug interactions’.)
- Salem DN, O’Gara PT, Madias C, et al. Valvular and structural heart disease: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008; 133:593S.
- Baddour LM, Wilson WR, Bayer AS, et al. Infective endocarditis: diagnosis, antimicrobial therapy, and management of complications: a statement for healthcare professionals from the Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease, Council on Cardiovascular Disease in the Young, and the Councils on Clinical Cardiology, Stroke, and Cardiovascular Surgery and Anesthesia, American Heart Association: endorsed by the Infectious Diseases Society of America. Circulation 2005; 111:e394.
- Davenport J, Hart RG. Prosthetic valve endocarditis 1976-1987. Antibiotics, anticoagulation, and stroke. Stroke 1990; 21:993.
- Pruitt AA, Rubin RH, Karchmer AW, Duncan GW. Neurologic complications of bacterial endocarditis. Medicine (Baltimore) 1978; 57:329.
- Di Salvo G, Habib G, Pergola V, et al. Echocardiography predicts embolic events in infective endocarditis. J Am Coll Cardiol 2001; 37:1069.
- Snygg-Martin U, Gustafsson L, Rosengren L, et al. Cerebrovascular complications in patients with left-sided infective endocarditis are common: a prospective study using magnetic resonance imaging and neurochemical brain damage markers. Clin Infect Dis 2008; 47:23.
- Rasmussen RV, Snygg-Martin U, Olaison L, et al. Major cerebral events in Staphylococcus aureus infective endocarditis: is anticoagulant therapy safe? Cardiology 2009; 114:284.
- Chan KL, Dumesnil JG, Cujec B, et al. A randomized trial of aspirin on the risk of embolic events in patients with infective endocarditis. J Am Coll Cardiol 2003; 42:775.
- Tornos P, Almirante B, Mirabet S, et al. Infective endocarditis due to Staphylococcus aureus: deleterious effect of anticoagulant therapy. Arch Intern Med 1999; 159:473.
- Steckelberg JM, Murphy JG, Ballard D, et al. Emboli in infective endocarditis: the prognostic value of echocardiography. Ann Intern Med 1991; 114:635.
- Dickerman SA, Abrutyn E, Barsic B, et al. The relationship between the initiation of antimicrobial therapy and the incidence of stroke in infective endocarditis: an analysis from the ICE Prospective Cohort Study (ICE-PCS). Am Heart J 2007; 154:1086.
- Paschalis C, Pugsley W, John R, Harrison MJ. Rate of cerebral embolic events in relation to antibiotic and anticoagulant therapy in patients with bacterial endocarditis. Eur Neurol 1990; 30:87.
- Wilson WR, Geraci JE, Danielson GK, et al. Anticoagulant therapy and central nervous system complications in patients with prosthetic valve endocarditis. Circulation 1978; 57:1004.
- Leport C, Vilde JL, Bricaire F, et al. Fifty cases of late prosthetic valve endocarditis: improvement in prognosis over a 15 year period. Br Heart J 1987; 58:66.
- Yeh TJ, Anabtawi IN, Cornett VE, Ellison RG. Influence of rhythm and anticoagulation upon the incidence of embolization associated with Starr-Edwards prostheses. Circulation 1967; 35:I77.
- Carpenter JL, McAllister CK. Anticoagulation in prosthetic valve endocarditis. South Med J 1983; 76:1372.
- Lieberman A, Hass WK, Pinto R, et al. Intracranial hemorrhage and infarction in anticoagulated patients with prosthetic heart valves. Stroke 1978; 9:18.
- Hart RG, Kagan-Hallet K, Joerns SE. Mechanisms of intracranial hemorrhage in infective endocarditis. Stroke 1987; 18:1048.
- Masuda J, Yutani C, Waki R, et al. Histopathological analysis of the mechanisms of intracranial hemorrhage complicating infective endocarditis. Stroke 1992; 23:843.
- Kupferwasser LI, Yeaman MR, Shapiro SM, et al. Acetylsalicylic acid reduces vegetation bacterial density, hematogenous bacterial dissemination, and frequency of embolic events in experimental Staphylococcus aureus endocarditis through antiplatelet and antibacterial effects. Circulation 1999; 99:2791.
- Kupferwasser LI, Yeaman MR, Nast CC, et al. Salicylic acid attenuates virulence in endovascular infections by targeting global regulatory pathways in Staphylococcus aureus. J Clin Invest 2003; 112:222.
- Anavekar NS, Tleyjeh IM, Anavekar NS, et al. Impact of prior antiplatelet therapy on risk of embolism in infective endocarditis. Clin Infect Dis 2007; 44:1180.
- Chan KL, Tam J, Dumesnil JG, et al. Effect of long-term aspirin use on embolic events in infective endocarditis. Clin Infect Dis 2008; 46:37.
- Immediate anticoagulation of embolic stroke: a randomized trial. Cerebral Embolism Study Group. Stroke 1983; 14:668.
- Immediate anticoagulation of embolic stroke: brain hemorrhage and management options. Cerebral Embolism Study Group. Stroke 1984; 15:779.
- Scott WR, New PF, Davis KR, Schnur JA. Computerized axial tomography of intracerebral and intraventricular hemorrhage. Radiology 1974; 112:73.
- Salgado AV, Furlan AJ, Keys TF. Mycotic aneurysm, subarachnoid hemorrhage, and indications for cerebral angiography in infective endocarditis. Stroke 1987; 18:1057.
- Brust JC, Dickinson PC, Hughes JE, Holtzman RN. The diagnosis and treatment of cerebral mycotic aneurysms. Ann Neurol 1990; 27:238.
- Delahaye JP, Poncet P, Malquarti V, et al. Cerebrovascular accidents in infective endocarditis: role of anticoagulation. Eur Heart J 1990; 11:1074.
- Horstkotte D, Piper C, Niehues R, et al. Late prosthetic valve endocarditis. Eur Heart J 1995; 16 Suppl B:39.