Clinical manifestations

Presentation and course

Infective endocarditis is a systemic disease and the neurologic complications of infective endocarditis must be considered in the context of the overall illness.

Systemic features of infective endocarditis. At present, development of symptoms of infective endocarditis usually arise acutely or subacutely (85) and the interval between a defined inciting event (eg, dental procedures) in patients with streptococcal endocarditis has been found to be 2 weeks or less in 84% of cases (16). The interval between symptom onset and diagnosis, however, tends to be in the range of 5 weeks and may be considerably longer (87). Fever is present in roughly 80% of patients, with related symptoms of chills, night sweats, or anorexia present in 25% to 40% (09; 46; 85; 13). Fever and other systemic signs may be absent in elderly or immunocompromised patients (46). Cardiac murmurs are present in up to 85% of cases but may not be easily auscultated and may be absent in right-sided endocarditis. Changing cardiac murmur or de novo appearance of a regurgitant murmur indicating valvular destruction have been considered classic signs of endocarditis. These signs are now infrequent, occurring in under 10% of patients. However, 90% of these patients will develop congestive heart failure. Splenomegaly can be present in up to 60% of patients but may be difficult to palpate (09). Embolic infarction – which may be painful or painless – can occur in virtually any organ including lungs, kidney, or spleen. Splenic embolization may also result in abscesses, which are more usually painful (50).

Embolization into the coronary arteries is especially likely to occur with aortic vegetations and may result in myocardial infarction (90). Septic emboli may result in abscesses or mycotic aneurysms systemically as well as within the central nervous system. Focal injury to kidneys or other organs may also result from the immune response induced by the infection (50).

Several findings are considered classic for infective endocarditis, although none of them are specific. These include digital clubbing; splinter hemorrhages in fingernails and toenails, petechial hemorrhages involving skin, conjunctivae, or buccal mucosa; embolic retinal lesions showing a pale center surrounded by hemorrhage (Roth spots); Osler nodes, which are 2 to15 mm painful subcutaneous nodules occurring on pads of fingers or toes caused by immune complex deposition; and Janeway lesions, representing nonpainful macular lesions also involving palms or soles (09; 112). At present, Roth spots, Osler nodes, and Janeway lesions are rarely encountered.

Neurologic manifestations of infective endocarditis. The great majority of neurologic events are due to septic embolization: this may result in ischemic stroke, vascular wall infection from the septic embolus causing vessel rupture with or without preceding mycotic aneurysm, or extension of infection outside the vessel to cause meningitis or brain abscess (44; 13; 50). The likelihood of embolization correlates with blood flow so that embolic events and their consequences are most frequent in the distribution of the middle cerebral artery followed by the anterior cerebral artery and the posterior circulation (44; 48). Most studies have shown the rate of cerebral complications in active infective endocarditis to be in the range of 16% to 33% (40; 49; 82), although others have reported a higher incidence (82; 56; 117). Ischemic stroke occurs in 20% to 40% of patients, intracerebral hemorrhage in up to 27%, and infectious intracranial aneurysms in 2% to 4% (40; 82; 56; 117). Asymptomatic events may be detected by MRI even more frequently (30; 21; 16). Duval and colleagues, in reviewing MRI images from 127 patients with definite or possible endocarditis, detected cerebral lesions in 106 individuals (82%) (34). In this study, ischemic lesions were detected in 68 patients, microhemorrhages in 74 patients, and silent aneurysms in 10 patients. In a minority of patients, the neurologic presentation of infective endocarditis may be that of confusion, depression, or encephalopathy.

Ischemic strokes. These may be large or small vessel and may be single or multiple. Large vessel stroke may result in typical ischemic syndromes of aphasia, neglect, or hemiparesis. In 10% to 15% of cases, strokes involve the posterior circulation. Embolization to small vessels may result in extremely focal deficits such as one-and-one half syndrome, internuclear ophthalmoplegia, Parinaud syndrome, or thalamic injury with unilateral asterixis (44; 38; 48; 125).

Hemorrhage can result from hemorrhagic conversion of infarcts or from rupture of an infected vessels with or without mycotic aneurysm. The resultant hemorrhage may be parenchymal, subarachnoid, or, in rare cases, subdural (27; 40; 50; 12; 117).

Mycotic aneurysms and microbleeds are more frequently associated with the viridans group of streptococci and tend to develop along branch points of arteries, often distant from the circle of Willis. The aneurysms can develop following direct bacterial invasion of the arterial wall, from embolic occlusion of the vasa vasora supplying larger vessels, or immune complex deposition with resultant injury to the arterial wall (50; 56). Mycotic aneurysms and cerebral microbleeds may be detected in patients without overt clinical symptoms but may also result in subarachnoid or parenchymal hemorrhage. Premonitory signs and symptoms precede a catastrophic hemorrhage from a mycotic aneurysm in a significant proportion of patients. Thus, onset of severe headache, meningismus, seizures, and focal deficits such as cranial neuropathies in a patient with known endocarditis should all prompt immediate investigation. Aneurysms may reduce in size or undergo apparent resolution with antibiotic treatment. However, this occurs in only one third of cases and aneurysmal rupture may occur months to years after the endocarditis has been successfully treated (08; 101).

Meningitis and brain abscess. Central nervous system invasion by bacteria in infective endocarditis may also produce meningitis or brain abscess. Meningitis is now unusual in most clinical series, occurring in 3% to 5% of patients (40; 82; 116). Both meningitis and brain abscess are most frequently associated with Staphylococcus aureus but can occur with Streptococcus pneumoniae and multiple other organisms (71; 82; 31; 116).

Presentation of meningitis or brain abscess in the setting of endocarditis does not differ from that seen in meningitis or brain abscess in individuals without cardiac valve infection. Although infrequent, Streptococcus pneumoniae endocarditis can occur in a clinical triad of pneumonia, aortic valve endocarditis, and meningitis (Austrian syndrome) with rapid destruction of the aortic valve and frequent adverse outcome (31; 97).

Encephalopathy. Acute encephalopathy during infective endocarditis can be the presenting symptom of infective endocarditis and may herald the development of meningitis or meningoencephalitis (40; 117). Encephalopathy can be directly related to central nervous system involvement, including multifocal ischemic injury, intracranial hemorrhage, or infection per se, but can also reflect the systemic complications of the disease. Confusion, personality change, headache, neck stiffness, seizures, and a declining level of consciousness all may suggest this process. In rare instances, infective endocarditis may present as severe depression without other changes in cognition (44; 96).

Other neurologic syndromes may accompany infective endocarditis. Headache (25% to 42% of patients with neurologic complications), seizures (1% to 15%), myelopathy, and ocular involvement can all develop in patients with endocarditis (82). Ophthalmic changes include the appearance of conjunctival petechiae, conjunctival hemorrhages, and Roth spots (112). Ocular involvement may also manifest as unilateral visual loss from central retinal artery occlusion or retinal hemorrhage; hemianopia may result from ischemia in the optic radiations or occipital cortex. Diplopia can arise secondary to brainstem or specific cranial nerve involvement (58; 44). Occasionally infective endocarditis can result in spondylodiscitis, with or without epidural abscess. For this reason, the possibility of underlying endocarditis should always be kept in mind in patients with unexplained spinal disc space infections or with epidural abscess (83; 29; 10). Rarely, emboli from infective endocarditis may cause a peripheral mononeuropathy (57).

Prognosis and complications

Overall in-hospital mortality in infective endocarditis overall is roughly 20% (higher in some studies), increasing to 25% to 30% by 6 months (13; 23). Patients with valvular destruction resulting in perivalvular fistulae have a mortality of 40% (13). Patients with neurologic involvement do less well, with a mortality rate of approximately 45% (40). Adverse prognostic factors include age over 70 years, congestive heart failure, cerebrovascular events, Staphylococcus aureus infection, prosthetic valve endocarditis, and infective endocarditis acquired in a healthcare setting (09; 85; 40). Mortality in intravenous drug abusers with infective endocarditis – even with surgery – can approach 50%, in part because of poor patient compliance with treatment (13; 43).

Biological basis

Etiology and pathogenesis

Causative organisms. Staphylococci, streptococcal organisms, and enterococci account for 80% to 90% of cases of infective endocarditis (13). In a series of 497 patients with infective endocarditis, Selton-Suty and colleagues detected the following isolates (108):

*HACEK organisms: Haemophilus sp., Aggregiatibacter sp., Cardiocterium, Eikenella corrodans, Kingella sp. (108; 13)

In the more recent EURO-ESC study, infective endocarditis was associated with staphylococci in 44.1% of patients, oral streptococci in 12.3%, enterococci in 15.8%, and Streptococcus gallolyticus in 6.6% (45). Gram-negative organisms such as Escherichia coli and Proteus sp. are rarely associated with infective endocarditis (69). Although Candida species are the most common cause of fungal endocarditis, Aspergillus can cause endocarditis as well, in particular in immunosuppressed organ transplant recipients and patients with prior valvular surgery (04; 54). Cases have also been associated with extracorporeal membrane oxygenation and a single case has been reported in an otherwise healthy patient who had undergone splenectomy (01; 88). The agent of Q fever, Coxiella burnetti, is a rare cause of endocarditis and should be kept in mind as a possible agent in individuals who have contact with sheep, goats, or cattle (71; 45; 69; 132).

The predominant organism in intravenous drug abusers is Staphylococcus aureus, accounting for roughly 70% of cases, followed by streptococci and enterococci (118; N'Guyen et al. 2017). A minority of cases are caused by Pseudomonas species, Candida albicans, or other fungi. Polymicrobial infections are more common in intravenous drug users than in nonusers (118; 28).

Pathogenesis. Although transient bacteremia is relatively common in normal individuals, normal cardiac valves are normally resistant to infection and infective endocarditis is an infrequent event (50): development of endocarditis usually requires underlying valvular abnormality or injury. An important exception to this is S. aureus, which can attack normal valves. In developing countries, rheumatic carditis remains a major risk factor for infective endocarditis and the mitral valve is most commonly affected. In developed countries, rheumatic heart disease is the underlying condition in less than 5% of patients, and important underlying conditions include the presence of prosthetic valves or cardiac devices such as permanent pacemakers or cardioverter-defibrillators, or congenital heart disease (50). Injectable drug use, with valvular damage caused by repeated injections of particulate matter, remains an important cause and is particularly associated with right-sided endocarditis. Valvular injury from any of these causes can result in interstitial valvular edema and valvular deposition of fibrin and platelets, a condition termed “nonbacterial thrombotic endocarditis,” which can provide a substrate for subsequent infection following bacteremia and development of valvular vegetations (13; 50). The major neurologic sequelae of infective endocarditis, as discussed above, are the result of emboli from these vegetations.

The severe inflammatory response elicited by infective endocarditis can also result in a number of apparently autoimmune events. Virtually all patients with infective endocarditis have circulating immune complexes (50) and patients may develop leukoclastic or ANCA-associated vasculitides (37). A subset of patients will develop anticardiolipin antibodies (110). In older literature, Winkelman and Eckel described a productive endarteritis of small cortical vessels in brains of patients dying from endocarditis (133). The roles of these immune and inflammatory events in encephalopathy or other neurologic sequelae of infective endocarditis remain uncertain.

Epidemiology

Infective endocarditis is an infrequent condition, with an incidence of 1.5 to 11.6 cases per 100,000 person-years (13; Holland et al. 2016). The condition is predominantly associated with 4 groups of conditions: congenital or acquired valvular disease; prosthetic valves; indwelling cardiac appliances including pacemakers and extracorporeal membrane oxygenation; and intravenous drug abuse. The EURO-ENDO registry, involving a prospective cohort of 3116 adult patients, recorded native valve endocarditis in 1764 (56.6%) patients, prosthetic in 939 (30.1%), and device-related in 308 (9.9%) (45). In that study, infective endocarditis was community-acquired in 2046 (65.66%) patients and involved native valves in 56.6% of patients, prosthetic valves in 30.2%, and device-related infections in 9.9% (45). Additional risk factors include human immunodeficiency virus infection and prolonged exposure to healthcare settings (85; 50). In the prospective cohort study by Murdoch and colleagues, the most common underlying condition was degenerative (predominantly mitral or aortic) valvular disease (43% of cases) followed by the presence of a prosthetic valve (23% of cases) (85). Intravenous drug abuse accounted for 10% of cases, with an equal number associated with infections resulting from chronic intravenous access. Only 3% of their patients had rheumatic valvular disease. Significant underlying conditions include diabetes (16% of patients in Murdoch’s series) and hemodialysis. Although left-sided endocarditis is much more common in general, the distribution of valvular involvement differs strikingly in intravenous drug users: in this population 79% of cases have been reported to involve right-sided valves (as opposed to 13% in non-IV drug abusers), with 16% of cases involving left sided valves and 5% valves on both sides of the heart (118; 28). Left-sided valvular involvement has been reported to be higher, however, in intravenous drug users who develop neurologic complications (75).

Increases in cases of infective endocarditis related to injection drug use has been documented in multiple countries (127; 92; 78; 102; 105) and has been associated with a rise in prescriptions for hydromorphone (131; 114); in the United States this has been particularly prominent in non-Hispanic white patients in northeastern and southern states (105).

Prevention

Identifying patients with conditions that predispose to the development of infective endocarditis has been considered an important means of preventing disease and antibiotic prophylaxis has traditionally been given prior to dental procedures or other potential causes of bacteremia. However, a beneficial effect of antibiotic prophylaxis prior to dental procedures remains controversial (129). A cohort study evaluating nearly 139,000 adults with prosthetic heart valves and an antibiotic prophylaxis rate of 50% for invasive dental procedures showed that invasive dental procedures may contribute to development of infective endocarditis, but there was no significant reduction in infective endocarditis rate with antibiotic prophylaxis (126). At present, good oral hygiene is emphasized and both the American Heart Association and the European Society for Cardiology currently recommend prophylaxis prior to dental procedure in patients who have cardiac conditions placing them at high risk for adverse outcome if infective endocarditis should occur (129). High-risk cardiac conditions in which antimicrobial prophylaxis for endocarditis may be indicated include prosthetic heart valves (including bioprosthetic and homograft valves), a prior history of infective endocarditis, complex cyanotic congenital heart diseases, and surgically constructed systemic or pulmonary conduits. Moderate risk conditions include other congenital cardiac malformations, acquired valvular dysfunction, hypertrophic cardiomyopathy, and mitral valve prolapse with documented valvular regurgitation, or valvular thickening, or both. Jeppson and colleagues reported a patient who developed multivalvular right- and left-sided endocarditis after elective pregnancy termination without preexisting cardiac disease, emphasizing the importance of prophylactic antibiotics in such select cases (55).

Differential diagnosis

The classical clinical indicators of infective endocarditis should suggest the diagnosis: these include cardiac murmur, fever, and elevated C-reactive protein and erythrocyte sedimentation rate, which may reach levels over 100 mm/hr. However, the systemic and neurologic consequences of infective endocarditis are protean and these classical symptoms, including murmur, are not present in all patients. For example, single or multiple cerebral or brainstem infarcts, which may be the presenting manifestations of endocarditis, are more commonly due to atrial fibrillation or, less often, aortic atheromata. Furthermore, many symptoms and signs found in infective endocarditis, including elevated C-reactive protein and erythrocyte sedimentation rate, can be duplicated by other conditions. These include vasculitides accompanying collage-vascular diseases such as polyarteritis nodosa; the hypercoagulable state seen with anticardiolipin antibody syndrome occurring alone or with other collagen vascular diseases; Libman-Sachs endocarditis, seen with systemic lupus erythematosus and/or anticardiolipin antibody syndrome; Loeffler endocarditis, a rare consequence of hypereosinophilic syndrome with cardiac involvement (Churg-Strauss syndrome); marantic (nonbacterial) endocarditis accompanying malignancy (02; 53; 106), Behçet syndrome (33); or conditions such as hepatitis B antigenemia (41). Clinical and MRI changes resembling CNS infarcts, at times with accompanying rise in erythrocyte sedimentation rate, can also be seen in intravascular lymphomatosis (14; 70; 111).

Diagnostic workup

The diagnostic evaluation of an individual with neurologic complications of infective endocarditis involves 2 separate processes: (1) diagnosis of the endocarditis itself and identification of the infectious organism; and (2) delineation of the nature and extent of neurologic involvement.

Diagnosis of infective endocarditis. Another finding includes hypointense signal spots on T2*-weighted MRI. These were detected in 84% of patients in 1 series. Of these, 46% had ischemic lesions, 22% subarachnoid hemorrhage, 9% intraparenchymal hemorrhage, and 9% infectious aneurysms, and the strong association between T2*-weighted hypointense signal spots and cerebral findings may suggest this as additional diagnostic criteria for infective endocarditis (39). In another series, cerebral microhemorrhages were found in 57% of patients imaged within 1 week of hospitalization for infective endocarditis (60).

Interestingly, the role of presurgical evaluation is still not entirely clear either. The diagnosis of infective endocarditis requires a combination of clinical, laboratory, and echocardiographic data. The gold standard for definitive diagnosis has been 2 positive blood cultures in a patient without another identifiable source of infection. However, this standard has been modified and supplanted by the Duke and then modified Duke criteria (68). In these criteria, the most important is the presence of positive blood culture(s), with predisposing factors, fever, and vascular phenomena representing minor criteria. Blood cultures thus remain the most important laboratory study in the diagnosis of infective endocarditis (13; 50). Echocardiography and/or other cardiac imaging studies are second in importance. Hematological studies, erythrocyte sedimentation rate, and C-reactive protein provide indication of ongoing infection.

(68; 13)

General physical examination. Clinical manifestations of infective endocarditis include fever in up to 90% of patients, chills, anorexia, and weight loss. Symptoms can be subacute and insidious, adding to difficulty and delay in diagnosis. Other symptoms include malaise, headache, myalgias and arthralgias, and night sweats. Cardiac murmur occurs in up to 85% of patients but may not be detected by auscultation. Most significant is appearance of a new regurgitant murmur indicating developing valvular insufficiency. Splenomegaly and cutaneous manifestations are hallmarks of infective endocarditis. These include splinter hemorrhages; petechiae of the skin, mucous membranes, palate and conjunctivae; Janeway lesions, which are nonerythematous papules on palms and soles; Osler nodes, which are tender subcutaneous nodules on palms, fingers, and toes; and Roth spots, exudative hemorrhagic lesions of the retina with pale centers.

Blood work and inflammatory markers. These include CBC, erythrocyte sedimentation rate, and C-reactive protein. Patients may exhibit a normochromic/normocytic anemia. Urinalysis can show proteinemia and, importantly, microscopic hematuria or RBC casts (13; 50). Infective endocarditis may be accompanied by a number of inflammatory markers including positive rheumatoid factor, hypoglobulinemia, cryoglobulinemia, circulating immune complexes, lupus anticoagulant, anticardiolipin antibodies, and false positive test for syphilis. Lupus anticoagulant and anticardiolipin antibodies may improve or resolve with antibiotic therapy (72). It must be kept in mind that not all patients with infective endocarditis show signs of inflammation or inflammatory markers: in a retrospective review of 469 patients with left-sided endocarditis, Ribeyrolles and colleagues identified 13 patients (2.3%) with C-reactive protein values between 4.7 and 14.2 mg/L (100). The main presentations were apyretic heart failure and stroke. Eighty-five percent of these patients were without fever and all patients had normal white blood cell and neutrophil counts. Sixty-two percent of these patients had severe valvular lesions on transesophageal echocardiography, including severe valvular regurgitation, valvular perforations, and paravalvular abscess (100).

Blood cultures. These represent the essential diagnostic test for infective endocarditis. Recommendations per the American Heart Association are that at least 3 sets of blood cultures should be obtained from different venipuncture sites, with the first and last samples drawn at least 1 hour apart.

Culture-negative endocarditis. This is most common where blood has been sterilized by previous antibiotic treatment. Cultures may also be initially negative in cases of endocarditis due to fastidious organisms, such as HACEK bacteria, defective streptococci-Gemelia, Granulicatella, Abiotrophia sp., Propionibacterium acnes, or Candida species: in these cases, prolonged incubation will usually allow identifying the causative pathogen in a few days. A minority of cases are due to intracellular bacteria that cannot be routinely cultured in blood with currently available techniques. Examples include Bartonella sp. or Coxiella burnetti, which may be diagnosed serologically or by PCR or culture of excised heart valve tissue (123; 89). Tropheryma whipplei, a rare cause of culture-negative endocarditis, may be diagnosed by histological or PCR examination of intestine or excised heart valve, or by serum PCR (64; 77; 86). The newer technique, metagenomic next generation sequencing, which isolates nonhost species of nucleic acids and compares them to a database of infectious organism nucleic acid sequences, may provide a powerful additional tool in the diagnosis of culture-negative endocarditis (22; 24; 26; 61).

Echocardiography. These studies are second in importance only to blood cultures and are crucial in identifying valvular lesions. Transthoracic echocardiography has a sensitivity of only about 70% for detecting vegetations on native valves and 50% for detecting vegetations on prosthetic valves. Transesophageal echocardiography is complementary to transthoracic echocardiography where transthoracic echocardiography is negative and the recommended first-line imaging test for patients with prosthetic valves and no contraindications to the test (13; 50). Although transesophageal echocardiography has a higher sensitivity than transthoracic echocardiography (up to 96% for vegetations on native valves and 92% for those on prosthetic valves), it can still fail to detect valvular vegetations. False-negative findings are likelier in patients who have preexisting severe valvular lesions, prosthetic valves, cardiac implanted electronic devices, small vegetations, or abscesses, or if a vegetation has already broken free and embolized (07; 13; 50; 63; 129). Distinguishing between vegetations and thrombi, cardiac tumors, and myxomatous changes can be difficult.

Transcranial Doppler ultrasound. Transcranial Doppler ultrasound provides a noninvasive method of detecting emboli within the cerebral circulation and has been widely used in study of stroke patients and can be an important tool in predicting patients at risk for neurologic complications in infective endocarditis. A study by Huang and colleagues suggested that transcranial doppler may be useful in noninvasive detection of intracranial embolic events and monitoring effect of antibiotic treatment (51). The study may also provide valuable data regarding the timing of valve replacement surgery (67).

Cardiac CT and MRI. 4D cardiac CT can detect local extension of infection, including abscess, fistula, and pseudoaneurysm. It can also incidentally detect pulmonary emboli. It is useful in suspected infective endocarditis in patients with negative transthoracic echocardiography and contraindications to transesophageal echocardiography. 4D CT can be coupled with CT angiography for a noninvasive way to evaluate the coronary arteries perioperatively without the risks associated with catheterization (62; 63; 79). Cardiac MRI is less well studied. It is potentially more sensitive than echocardiography for detecting vegetations, especially in patients with poor echocardiography images. However, it is unclear if it is better than CT and cannot be readily used in patients with noncompatible metallic hardware.

FDG-PET and leukocyte scintigraphy. The utility of FDG-PET is founded on the uptake of 18F-fluorodeoxyglucose by cells, with higher uptake taking place in cells with higher metabolic activity, including as in areas of inflammation. Similarly, leukocyte scintigraphy relies on the use of radiolabeled leukocytes (ie, leukocytes previously extracted from the patient, labelled, and reintroduced into the patient) to allow for localization of inflamed tissue. The most significant contribution of FDG-PET may be the ability to detect infective endocarditis early, when echocardiography is initially negative. When abnormal FDG uptake was included in the modified Duke criteria, it increased the sensitivity to 97% for detecting infective endocarditis on admission (103; 79). Both FDG-PET and leukocyte scintigraphy have a high sensitivity, specificity, and negative predictive value for cardiac implanted electronic device infection. They may help to differentiate thrombus from infected vegetation. FDG-PET and leukocyte scintigraphy are especially useful for patients with a prosthetic valve or cardiac implanted electronic device; in 1 study in patients with prosthetic valves and suspected infective endocarditis, FDG-PET was found to have a sensitivity of up to 91% and a specificity of up to 95% (121).

Brain imaging. As with stroke, CT scan is often done initially to rule out hemorrhage or hemorrhagic stroke. Beyond this, however, contrast-enhanced MRI is preferred to detect parenchymal lesions. Both CT angiography and MR angiography can be used for initial evaluation of intracranial vasculature but may miss mycotic aneurysms seen on catheter angiography (52); for this reason 4-vessel catheter angiography remains the gold standard still diagnosis of mycotic aneurysms given their frequent distal location within the cerebral arterial tree (52).

Vessel wall imaging may provide an additional noninvasive means of detecting areas of actual vascular infection. Although it is common practice to evaluate for mycotic aneurysms in setting of hemorrhagic lesions or high suspicion, routine screening has not been proven to improve outcome: in a study of preoperative MRI findings, there was not an association of findings with postoperative outcomes regardless of timing of valvular surgery (19; 18).

Lumbar puncture. Lumbar puncture should be performed if meningitis is suspected and should be considered is indicated in patients with disorientation, altered level of consciousness, or significant headache. The study has little value in stroke and is relatively contraindicated as an initial study if brain abscess, other space-occupying lesion, or severe cerebral edema are suspected. A specific CSF formula, however, does not exist to differentiate endocarditis from other infectious etiologies.

Management

Treatment of infective endocarditis with neurologic complications requires consideration of 4 separate areas: antibiotic treatment, surgical repair or replacement of infected valves, treatment of the neurovascular complications of endocarditis including stroke, mycotic aneurysm, and hemorrhage, and treatment of meningitis or brain abscess. A closely collaborative team approach consisting of neurologists, infectious disease specialists, interventional radiologists, and cardiothoracic surgeons is essential in providing effective care (35; 36).

Antibiotic therapy. Antibiotics are the mainstay of treatment for infective endocarditis. Neurologic manifestations mainly occur before antimicrobial treatment is begun, and rapid diagnosis and initiation of antimicrobial therapy is still the most effective means to prevent neurologic complications. The risk of stroke after antibiotic initiation decreases 0.5% to 0.3% per day (47).

Antibiotic treatment is eventually determined by antibiotic sensitivities as determined by blood culture. Empiric therapy, before antibiotic sensitivities are known, can be as follows (07; 13; 50; 20):

Therapy is typically continued for 4 to 6 weeks depending on the specific organism (07). Fungal endocarditis, though rare, has high mortality and demands a high index of suspicion, especially in patients with predisposing host conditions. Candida endocarditis should be treated immediately with liposomal amphotericin B or caspofungin with optional addition of flucytosine (07; 54; 107). Voriconazole has also been used as treatment. Treatment of Aspergillus endocarditis is difficult. Amphotericin has been used in combination with voriconazole (83; 54; 107).

Cardiac surgery. Cardiac valve surgery in patients with infective endocarditis is needed in about 40% to 50% of patients with infective endocarditis and is performed for 3 main indications: heart failure due to valve dysfunction, uncontrolled infection, and prevention of embolism. Absolute indications for cardiac surgery include refractory congestive heart failure, perivalvular or myocardial abscess, rupture of a sinus of Valsalva aneurysm, antibiotic failure, unstable prostheses, fungal endocarditis, recurrent emboli despite adequate antimicrobial therapy, culture-negative infective endocarditis with recalcitrant fever for more than 10 days despite antimicrobial therapy, and relapse of infective endocarditis despite aggressive antibiotic therapy, especially in prosthetic valve endocarditis (50). Early valvular surgery is essential in meningitis due to Candida or Aspergillus endocarditis. Postsurgical outcome may be worse in patients who have neurologic complications of infective endocarditis as opposed to those without such complications (40). Cerebral embolism may factor into the decision about timing of valve replacement, and neurologists frequently are asked to comment about the risk of bleeding into an ischemic stroke during cardiac bypass or about the effect of urgent cardiac surgery on recent cerebral infarction.

Prosthetic valve endocarditis presents particular challenges and cardiac surgeons should be involved as soon as the diagnosis of prosthetic valve endocarditis is made to determine best treatment strategies in a multidisciplinary manner (06).

Traditionally, valvular surgery was delayed until patients with infective endocarditis-related stroke were neurologically stable, typically for about 2 weeks or longer; this was based on earlier studies that suggested that patients operated on within the first 2 weeks after stroke may suffer further neurologic deterioration and that valve replacement 3 or more weeks after stroke showed no excess neurologic morbidity (42). More recent data, however, has indicated that among patients with neurologic complications in whom valve surgery was indicated but not performed, mortality by far exceeded that of patients in whom surgery was performed despite the neurologic complication (68 vs. 20%, respectively) (109). In recent years as well, there is cumulative evidence that even early valve surgery may be beneficial, at least in select cases, and not necessarily associated with increased morbidity or mortality. In a retrospective study of 137 patients with endocarditis and nonsevere stroke (NIHSS ≤10), Murai and colleagues showed that patients undergoing early surgery had a greater survival rate free of infective endocarditis-related death (84). Another study showed patients with septic emboli who received early surgery showed no significant difference in incidence of postoperative neurologic events and mortality when compared to those without emboli (59). In a small series of early valve surgery (within 10 days) for ongoing sepsis or emboli in patients with ischemic or hemorrhagic stroke and comprehensive imaging to identify actionable neurologic findings, 11 of the 12 patients had good outcome (98). Transesophageal echocardiographic studies may define a group of patients whose size and configuration of vegetations dictates emergent cardiac surgery (99). However, data are controversial and in particular, the impact of early valve surgery on the outcome of Staphylococcus aureus (SA) prosthetic valve infective endocarditis (PVIE) remains unresolved. In a prospective, multinational, observational cohort of 1025 patients with infective endocarditis, Lalani and colleagues found that after adjustment for differences in clinical characteristics and survival bias, early valve replacement was not associated with lower mortality compared with medical therapy (65). Chirouze and associates, reviewing patients in that cohort with Staphylococcus aureus prosthetic valve endocarditis, reported that early valve surgery performed within the first 60 days of hospitalization was not associated with reduced 1-year mortality (25). These data suggest that the decision to pursue early valve surgery should be individualized for each patient and should be based on infection-specific characteristics rather than solely on the microbiology of the infection causing prosthetic valve infective endocarditis (25). Major predictors of embolic events include intravenous drug use, Staphylococcus aureus infection, mitral valve vegetation, and vegetation size greater than 10 mm (134). In a metaanalysis of early versus late surgery in patients with endocarditis and neurologic events, Tam and colleagues found that data supported delaying surgery by 7 to 14 days, if possible, in infected endocarditis complicated by ischemic stroke and at least 21 days in hemorrhagic stroke (122). Randomized trials are needed for definitive guidance in this area and newer surgical or interventional procedures such a valve reconstruction using novel biomaterials may improve outcomes (76; 129). Although the presence of intracranial hemorrhage often defers surgeons from offering valvular surgery to patients with infectious endocarditis, data suggest that among patients with intracranial cerebral hemorrhage and an indication for cardiac surgery, those managed with a conservative approach had a higher mortality than those managed surgically (104).

Treatment of thromboembolic complications. Here, management strategies are less well established and controversies exist. Major areas of treatment controversy include: (1) the indications for and risk of prophylaxis and treatment for endocarditis-associated cerebral embolism, (2) the appropriate management of infectious intracranial (mycotic) aneurysms, and (3) the role and timing of cardiac surgery in active endocarditis.

Generally, antiplatelet and anticoagulant prophylaxis is not recommended due to risk of (intracranial) hemorrhage (82; 113). If anticoagulation is urgently indicated, such as in patients with mechanical (mitral) valves, practice varies. Patients with mechanical prosthetic valves who develop endocarditis are often continued on their anticoagulation in absence of intraparenchymal hemorrhage. However, the risk of hemorrhage if embolism occurs is then high. Anticoagulation should be withheld for at least 48 hours in prosthetic valve patients suffering a cerebral embolism with endocarditis. Patients with cardiogenic brain embolism should be monitored for signs of deterioration that suggest a hemorrhagic transformation, and a follow-up imaging study in 1 to 2 weeks is advisable in order to rule out abscess formation or evidence of a mycotic aneurysm. Some guidelines recommend to replace oral anticoagulation, if anticoagulation is absolutely needed, with intravenous heparin but larger outcome data are lacking (46).

Treatment with intravenous thrombolysis and/or endovascular thrombectomy remains controversial in patients with embolic stroke due to infective endocarditis because of the increased risk of hemorrhage. Furthermore, the diagnosis of endocarditis may be difficult in the time frame where decision making is essential regarding intravenous thrombolysis and endovascular thrombectomy. Walker and colleagues found that thrombolysis in the setting of infective endocarditis carried a mortality of 70% (128) and AHA/ASA guidelines currently state that because of the risk of intracranial hemorrhage, intravenous rTPA (Alteplase) should not be administered to patients with infective endocarditis and ischemic symptoms (32). Although studies are limited, thrombectomy appears to have less risk of intracranial hemorrhage, and in 1 study overall outcome was similar to that of thrombectomy for stroke in atrial fibrillation (05; 105; 11; 74; 115). MRI with gradient echo or, ideally, susceptibility-weighted images could be useful in determining patients at particular risk. However, there are also reports of severe vasospasm after mechanical thrombectomy in infective endocarditis; this was thought to be due to thrombectomy-induced injury to the frail vessel wall (91).

Treatment of mycotic aneurysms. Mycotic aneurysms – which may be single or multiple – have been reported to occur in roughly 9% of patients with infective endocarditis (52). Approximately one quarter of these will resolve with antibiotic therapy alone (101) but the remainder may enlarge and rupture. Four-vessel angiography remains the standard for the detection and follow-up. Angiography should be performed in cases of CT-documented hemorrhage and persistent focal headache with cerebrospinal fluid pleocytosis (52). The indications for interventional radiology or neurosurgical repair of aneurysm must be individualized by the patient’s medical condition and by organism, location, number, and size. Unruptured aneurysms can be treated medically and followed with serial angiography or can be treated with coiling or, less often, surgical clipping in cases of aneurysms that have ruptured or are enlarging (17; 113; 03). The use of antiplatelet agents is thought to be necessary to prevent rethrombosis in stent placement if performed.

Treatment of meningitis and brain abscess. These are similar to treatment of patients without infective endocarditis and are discussed elsewhere in Medlink Neurology.

Outcomes

Overall in-hospital mortality in infective endocarditis is roughly 20%, increasing to 25% to 30% by 6 months (13; Chen et al. 2020). Five-year mortality is as high as 40% and can approach 70% in patients requiring intensive care unit hospitalization (80). Patients with neurologic involvement have a mortality rate of approximately 45% (40). Altered mental status at infective endocarditis onset, indicative of brain injury, is a major determinant of short-term outcome (16). Other adverse prognostic factors include age over 70, congestive heart failure, cerebrovascular events, Staphylococcus aureus infection, prosthetic valve endocarditis, and infective endocarditis acquired in a healthcare setting (09; 85; 40; 109). As mentioned above, mortality in intravenous drug abusers with infective endocarditis – even with surgery – can approach 50%, in part because of poor patient compliance with treatment (13; 43). Mortality in Candida endocarditis approaches 40% and that of Aspergillus endocarditis 60% (54; 66).

Special considerations

Pregnancy

Incidence, clinical symptomatology, and treatment of abscess in pregnancy do not differ from those seen in nonpregnant individuals, nor do the complications of infective endocarditis affect the fetus. All antibiotics cross the placenta to various degrees, but only tetracycline has been documented to affect fetal development (81). Although few solid data regarding the teratogenic potential of other antibiotics exist, experience suggests that the penicillins (except ticarcillin), cephalosporins, and aminoglycosides can all be safely used in pregnancy (81). Therapy for endocarditis should, therefore, continue during pregnancy with the understanding that higher drug levels may be required, given the increased volume of distribution and rate of clearance that normally occur in pregnant women.

Anesthesia

Patients with infective endocarditis who undergo cardiac surgery in the setting of a recent CNS complication may pose a significant perioperative risk from both a cardiac and a neurologic standpoint. No particular anesthetic agents are contraindicated, and antibiotics are continued through the perioperative period in their usual doses (119).