Clinical manifestations and diagnosis of infective endocarditis

INTRODUCTION In the modern era, infective endocarditis (IE) presents most frequently as acute rather than chronic disease. The diagnosis is usually based upon a combination of factors, including a careful history and physical examination, blood cultures and other selected laboratory studies, echocardiography, and chest radiography

The clinical manifestations and diagnosis of IE will be reviewed here. Treatment of IE is discussed separately. (See "Antimicrobial therapy of native valve endocarditis" and "Antimicrobial therapy of prosthetic valve endocarditis".)

HISTORY AND PHYSICAL EXAMINATION Infective endocarditis (IE) usually presents acutely [1]. A careful clinical history should be performed, including surgical and medical procedures and recent bacterial infection. Special attention should be given to preexisting cardiac lesion(s) and clues pointing toward a recent source of bacteremia, such as indwelling prosthetic devices (including intravascular catheters, orthopedic hardware, and cardiac devices) or intravenous drug use.

The physical examination should include a careful cardiac examination for signs of new regurgitant murmurs or heart failure. (See "Auscultation of cardiac murmurs".)

A vigorous search should be undertaken for the clinical stigmata of endocarditis, including evidence of small and large emboli with special attention to the fundi, conjunctivae, skin, and digits. Peripheral cutaneous or mucocutaneous lesions of IE include petechiae, splinter hemorrhages, Janeway lesions, Osler's nodes, and Roth spots.

Petechiae and splinter hemorrhages are nonspecific for IE. Petechiae are the most common skin manifestation (picture 1). They may be present on the skin (usually on the extremities) or on mucous membranes such as the palate. Lesions on the conjunctivae usually present as hemorrhages best seen with eversion of the upper or lower eyelids. Splinter hemorrhages are also nonspecific for IE; they are nonblanching, linear, reddish-brown lesions found under the nail bed (picture 2).

Janeway lesions, Osler's nodes, and Roth spots are more specific findings of IE; they occur most frequently in the setting of protracted bacteremia and therefore are relatively uncommon in the modern era. Janeway lesions are macular, nonpainful, erythematous lesions on the palms and soles (picture 3). Osler's nodes are painful, violaceous nodules found in the pulp of fingers and toes and are seen more often in subacute than acute cases of IE (picture 4). Roth spots are exudative, edematous hemorrhagic lesions of the retina.

A neurologic evaluation should be undertaken for evidence of focal neurologic impairment; it is also important as a baseline examination, should neurologic deficits develop later. (See "Complications and outcome of infective endocarditis", section on 'Neurologic complications'.)

Patients with IE may have involvement of other organs due to metastatic infection (eg, vertebral osteomyelitis), embolic events (eg, focal neurologic deficits, renal infarct, splenic infarct), or a systemic immune reaction (eg, glomerulonephritis). In right-sided endocarditis, septic pulmonary emboli may be seen (image 1). (See "Complications and outcome of infective endocarditis" and "Renal disease in the setting of infective endocarditis or an infected ventriculoatrial shunt".)

DIAGNOSIS The diagnosis of infective endocarditis (IE) is usually based upon a combination of factors, including a careful history and physical examination, blood cultures and other selected laboratory studies, echocardiography, electrocardiography (ECG), and chest radiography.

The modified Duke criteria are the accepted criteria for diagnosis of IE; these are summarized in the Tables (table 1 and table 2) (calculator 1). (See "Infective endocarditis: Historical and Duke criteria", section on 'Duke criteria'.)

The diagnosis of IE is generally straightforward in the setting of positive blood cultures for a pathogen likely to cause endocarditis, together with evidence of endocardial involvement. However, it can be difficult to distinguish between IE and an alternate source of infection in a bacteremic patient with underlying heart disease.

In addition, some patients with IE do not have positive blood cultures (ie, culture-negative endocarditis), and up to 25 percent of patients have no identifiable cardiac lesion at initial presentation. The presence of atypical features may result in a delayed diagnosis.

Laboratory studies

Blood cultures

Culture collection Blood cultures should be obtained prior to initiation of antibiotic therapy. At least three sets of blood cultures should be obtained; in patients who have received recent antimicrobial therapy, additional blood cultures may be useful. If the tempo of illness is subacute and the patient is not critically ill, it is reasonable to delay initiation of antimicrobial therapy while awaiting the results of blood cultures and other diagnostic tests. In the setting of acute illness, three sets of blood cultures should be obtained over a one-hour period prior to initiation of empiric antimicrobial therapy. (See "Blood cultures for the detection of bacteremia".)

The diagnostic yield of more than three sets of blood cultures is minimal in the absence of recent antimicrobial therapy. In one series including 206 cases of endocarditis, the initial blood culture in patients with streptococcal endocarditis was positive in 96 percent of cases, and one of the first two blood cultures was positive in 98 percent [2]. In patients with IE caused by organisms other than streptococcus, the first blood culture was positive in 82 percent of cases and one of the first two cultures was positive in 100 percent of cases [2].

A minimum of 10 mL (and preferably 20 mL) of blood should be obtained from adults and 0.5 to 5 mL from infants and children. In one study, blood cultures inoculated with at least 5 mL of blood had a significantly higher detection rate for bacteremia than bottles inoculated with less than 5 mL of blood (92 versus 69 percent) [3]. The estimated yield of blood cultures in bacteremic adults increased approximately 3 percent per mL of blood cultured.

Each set of cultures should be obtained from separate venipuncture sites, and, ideally, blood cultures should not be obtained from a vascular catheter. Blood cultures may be collected at any time; they do not need to be obtained at the time of fever or chills since patients with IE typically have continuous bacteremia [4].

Interpreting culture results The Duke criteria for the diagnosis of endocarditis define the following organisms as "typical causes" of IE (table 1 and table 2) [5]:

Staphylococcus aureusViridans streptococci and Streptococcus bovisEnterococciHACEK group organismsThe risk of endocarditis in patients with S. aureus bacteremia (regardless of source or residence at the time of onset) is particularly high. As a result, all patients with S. aureus recovered from the blood should be clinically evaluated for IE. (See "Clinical approach to Staphylococcus aureus bacteremia in adults".)

The likelihood of endocarditis is variable depending on bacteria species. As examples:

Bacteremia due to Streptococcus sanguis is more often indicative of endocarditis than bacteremia due to Streptococcus milleri (also known as S. anginosus).Bacteremia due to group A or C streptococci is seldom associated with IE, whereas group G streptococcal infection is more frequently associated with endocarditis [6]. (See "Group C and group G streptococcal infection".)Bacteremia due to Enterococcus faecalis has been more frequently associated with IE than bacteremia due to other enterococcal species [7].The microbiology of IE is discussed further separately. (See "Epidemiology, risk factors and microbiology of infective endocarditis", section on 'Microbiology'.)

False-positive culture results occasionally occur despite careful techniques for collection and processing. Organisms for which it can be difficult to distinguish between pathogenicity and contamination include Propionibacterium acnes, Corynebacterium species, Bacillus species, and coagulase-negative staphylococci. In general, the likelihood of pathogenicity is increased if the organism is observed in multiple blood cultures obtained by independent venipunctures. Recovery of these organisms from a single blood culture or a minority of blood culture bottles likely reflects a false-positive result. (See "Blood cultures for the detection of bacteremia", section on 'Contamination'.)

The definition of persistent bacteremia depends on the likelihood of the organism to cause endocarditis [5]. For organisms likely to cause endocarditis (such as S. aureus or viridans streptococci), two positive blood cultures collected more than 12 hours apart may be presumed to represent true bacteremia. For organisms that are more commonly skin contaminants, three or a majority of four or more positive blood cultures (with the first and last samples collected at least one hour apart) may be presumed to represent true bacteremia.

Culture-negative endocarditis Culture-negative endocarditis should be considered in patients with negative blood cultures and persistent fever with one or more clinical findings consistent with infective endocarditis (eg, stroke or other manifestations of emboli). Culture-negative IE should also be considered in patients with vegetation on echocardiogram and no clear microbiologic diagnosis. Issues related to culture-negative IE are discussed separately. (See "Epidemiology, microbiology, and diagnosis of culture-negative endocarditis".)

Other laboratory tests The utility of other laboratory tests in the diagnosis of endocarditis is limited. The following findings may be observed among patients with IE but are relatively nonspecific:

An elevated erythrocyte sedimentation rate and/or elevated C-reactive proteinA normochromic normocytic anemiaThe white blood cell count may be normal or elevated in patients with subacute presentations of endocarditis; however, most patients with staphylococcal endocarditis have leukocytosis, and some may have thrombocytopenia.Hyperglobulinemia, cryoglobulinemia, circulating immune complexes, hypocomplementemia, elevated rheumatoid factor titers, and false-positive serologic tests for syphilis occur in some patients.An elevated rheumatoid factor titer in patients without a known prior rheumatologic disorder is a minor criteria in the Duke diagnostic scheme (table 2) [5]. (See "Origin and utility of measurement of rheumatoid factors".)

Most patients with endocarditis have an abnormal urinalysis, as manifested by microscopic or gross hematuria, proteinuria, and/or pyuria. Each of these findings lacks specificity. However, the presence of red blood cell casts on urinalysis is generally indicative of glomerulonephritis (often in association with hypocomplementemia) and is a minor diagnostic criterion for IE (table 2) [5]. (See "Renal disease in the setting of infective endocarditis or an infected ventriculoatrial shunt".)

Some causes of culture-negative IE may be identified via serology; these include Coxiella burnetii, Bartonella spp, Chlamydia spp, Legionella spp, and Brucella [8]. (See "Epidemiology, microbiology, and diagnosis of culture-negative endocarditis", section on 'Diagnosis'.)

Cardiac studies

Echocardiography An echocardiogram should be performed in all patients with a moderate or high suspicion of endocarditis (table 1 and table 2 and algorithm 1) [9-11]. Among patients with a low clinical probability of endocarditis, the diagnostic yield of transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) is low and neither should be performed [10]. (See "Infective endocarditis: Historical and Duke criteria" and "Role of echocardiography in infective endocarditis".)

Echocardiography allows detection and characterization of vegetations on valves or other sites (such as pacemaker wires), evaluation for valvular dysfunction, assessment of hemodynamic severity, and detection of associated abnormalities such as shunts or abscesses. It is also useful for follow-up evaluation of patients with virulent organisms, severe hemodynamic instability, persistent or recurrent bacteremia or fever, or other clinical deterioration.

In general, TTE is the first diagnostic test for patients with suspected IE. Detection of a vegetation is a positive test (movie 1 and movie 2 and movie 3). TTE has relatively low sensitivity (29 to 63 percent in different series), although the specificity approaches 100 percent [12]. Thus, the absence of vegetation does not preclude the diagnosis of IE, although the finding of normal valves (both morphology and function) substantially reduces the probability of IE. In one study, 96 percent of patients with normal valves and no vegetation on TTE also had a negative TEE [13].

For patients with a normal TTE, it is reasonable to proceed to TEE in the following circumstances:

A high clinical suspicion of IE (persistently positive blood cultures and/or multiple minor criteria for endocarditis) (table 1 and table 2)Bacteremia due to an organism known to be a common cause of IE such as S. aureus or viridans streptococciA technically limited TTE studyPatients with abnormal findings on TTE who may require further evaluation by TEE include those with significant valvular regurgitation to determine need for surgery, as well as those with one or more risk factors for perivalvular abscess including new conduction delay on ECG, aortic valve endocarditis, and persistent bacteremia or fever despite appropriate antimicrobial therapy.

It is reasonable to forgo TTE and proceed to TEE in the following circumstances:

Presence of prosthetic valve(s), particularly prosthetic aortic or mitral valve, in which shadowing may make TTE visualization difficultA prior valvular abnormality (including previous endocarditis)Limited transthoracic windows (eg, due to obesity, chest wall deformity, or mechanical ventilation)Echocardiography should be performed as soon as possible after the diagnosis of IE is suspected [9,14], even though TTE and TEE can be falsely negative if vegetations are small (and, occasionally, if previously present vegetations have embolized). High suspicion for IE in the setting of negative echocardiography should prompt subsequent repeat study. The optimal timing for repeat studies is uncertain; some favor repeat echocardiography after five to seven days in patients with a high likelihood of endocarditis, when clinical findings suggestive of IE are present in patients for whom initial studies were done during the first few days of symptoms or when the initial study was technically suboptimal.

Echocardiography is warranted even for patients with an associated condition that requires a protracted course of antimicrobial therapy (such as vertebral osteomyelitis). Documenting the presence or absence of vegetation is important for determination of subsequent follow-up.

Electrocardiogram A baseline electrocardiogram should be performed as part of the initial evaluation of all patients with suspected endocarditis, even though this test rarely demonstrates diagnostic findings. The presence of heart block or conduction delay may provide an important clue to extension of infection to the valve annulus and adjacent septum. In addition, the presence of findings consistent with ischemia or infarction may suggest the presence of emboli to the coronary circulation.

Thereafter, a weekly screening electrocardiogram may be useful (in the absence of telemetry monitoring), particularly in the setting for patients with vegetation(s) in close anatomic proximity to conductive tissue, which may increase the risk of developing heart block.

Chest radiograph Chest radiography may demonstrate evidence of septic pulmonary emboli or focal infiltrate, with or without cavitation. Rarely, valve calcification may be seen that may suggest endocarditis in the setting of suggestive history and physical findings.

Valve culture and histopathology Valve culture is useful if there are other clinical signs and symptoms and/or echocardiographic findings that suggest the presence of IE; routine culture of heart valves removed at the time of surgery can lead to false-positive results. In one study including 1030 valves removed at surgery, cultures were positive in 39 percent of patients that met Duke criteria for IE and in 28 percent of patients without other criteria for IE [15].

Histologic demonstration of microorganisms, vegetations, or active endocarditis in cardiac valve tissue obtained at surgery is included in the Duke criteria and may be used to confirm the diagnosis of IE (table 1). The histologic features that characterize endocarditis were defined in a retrospective pathologic analysis of tissue adjoining mechanical cardiac valves in 90 patients who underwent surgical removal of a mechanical valve for suspected IE (21 patients) or noninfectious dysfunction (69 patients) [16]. IE was characterized by microorganisms, vegetations, and neutrophil-rich inflammatory infiltrates with extensive neovascularization. In contrast, tissue adjoining valves from noninfectious complications demonstrated extensive fibrosis and, when present, inflammatory infiltrates that were mainly composed of macrophages and lymphocytes.

Thus, when no microorganisms are detected and vegetations are lacking in tissue adjacent to a mechanical valve, neutrophil-rich inflammation and extensive neovascularization may allow differentiation between IE and inflammatory noninfectious valve processes in patients with mechanical cardiac valves who undergo surgery.

Other tests Radiographic imaging should be tailored to findings on history and physical examination. Patients with back pain should be evaluated for vertebral osteomyelitis and discitis with MRI. Computed tomography (CT) of the torso should be pursued for patients with abdominal pain or costovertebral angle tenderness to evaluate for presence of splenic infarct, renal infarct, psoas abscess, or other intraabdominal sites of infection. Such imaging is also appropriate for patients with documented endocarditis, even in the absence of focal symptoms, to evaluate for subclinical sites of metastatic infection. An initial imaging study performed early in the presentation of acute IE may be negative, and worsening back pain should prompt repeat imaging, especially in the setting of S. aureus infection.

The role of routine brain magnetic resonance imaging (MRI) is uncertain. In one study including 53 patients, early use of cerebral MRI led to the reclassification from possible to definite IE in one-third of cases [17]; further study is needed.

A bedside clinical score may be a useful tool for prediction of endocarditis in some cases. In one study describing such a tool for prediction of endocarditis in patients with enterococcal bacteremia (NOVA: number of positive blood cultures [3/3 or the majority if more than 3], 5 points; unknown origin of bacteremia, 4 points; prior heart valve disease, 2 points; auscultation of a heart murmur, 1 point), a cutoff score 0.12 mcg/mL and 0.5 mcg/mL), and rare strains are considered to be fully resistant with a penicillin MIC >0.5 mcg/mL [5].

Endocarditis due to viridans group streptococci and S. bovis organisms can usually be microbiologically cured if one of four different regimens is used (table 1 and table 2):

For treatment of penicillin-susceptible strains (MIC 0.12 mcg/mL), the American Heart Association (AHA), British Society for Antimicrobial Chemotherapy (BSAC), and European Society for Cardiology (ESC) all recommend a regimen consisting of aqueous crystalline penicillin G at a dose of 12 to 24 million units daily (either continuously or in four to six equally divided doses) for four weeks in patients with endocarditis due to highly penicillin-susceptible streptococci (table 1) [3,6,7]. Penicillin, or an equally effective regimen consisting of ceftriaxone given intravenously or intramuscularly as a 2 g daily dose for four weeks, remain the preferred treatments for older adult patients or in other patients in whom aminoglycoside therapy is considered risky or contraindicated.Selected patients with native valve endocarditis due to penicillin-susceptible strains who do not have evidence of intracardiac or extracardiac complications, or preexisting renal or otic disease, may be treated with shorter courses of combination therapy (table 1). The AHA guidelines recommend a regimen that includes gentamicin plus either aqueous crystalline penicillin G or ceftriaxone for two weeks [3]. The BSAC guidelines recommend a regimen of gentamicin plus penicillin or ceftriaxone for two weeks [7], while the ESC guidelines recommend monotherapy with penicillin G for two weeks [6]. Gentamicin, if used, should be given either as a single dose of 3 mg/kg per day or in two to three equally divided doses adjusted to give a peak serum level of 3 to 4 mcg/mL. We prefer to administer the gentamicin in two to three equally divided doses for hospitalized patients when serum concentrations can be followed and as a single daily dose in outpatients.Gentamicin is more commonly used in clinical practice than streptomycin due to the wider availability of gentamicin serum levels and because dosing regimens for gentamicin are more familiar to most clinicians. When streptomycin is used to treat endocarditis, the suggested dose in patients with normal renal function is 15 mg/kg per day intramuscularly (IM) or intravenously (IV) divided equally into two doses a day for two weeks. Dose intervals need to be prolonged in patients with renal insufficiency.

Patients with prior histories of penicillin allergy can usually be treated with ceftriaxone, if their prior history of penicillin allergy consists of rash without other signs of immediate-type hypersensitivity. Patients with histories of immediate-type hypersensitivity may either be treated with vancomycin (15 to 20 mg/kg/dose every 8 to 12 hours, not to exceed 2 g per dose) for four weeks or desensitized to penicillin and treated with a standard regimen. In patients with streptococcal endocarditis and a history of significant penicillin allergy, some favor treatment with a combination of gentamicin with either vancomycin or teicoplanin [7].

We believe that penicillin is preferable to vancomycin; we are unaware of controlled treatment trials comparing the efficacy of vancomycin to penicillin in cases of suspected drug hypersensitivity. Thus, when expertise is available to safely supervise desensitization protocols, we advise desensitization to penicillin rather than substitution of vancomycin. Following desensitization, the tolerance persists only as long as the patient continues to receive the drug. Once antibiotic therapy is stopped for a period of more than 24 hours, repeat desensitization is required if the particular drug is to be used again. (See "Allergy to penicillins".)

For treatment of endocarditis due to "intermediate" susceptibility viridans streptococci (MIC >0.12 and 0.5 mcg/mL), the AHA guidelines recommend aqueous penicillin G (24 million units daily either continuously or in four to six equally divided doses) or ceftriaxone (2 g IV or IM once daily) for a total of four weeks [3]. Gentamicin should be added to this regimen for the first two weeks (dosed as above). As for other patients with streptococcal endocarditis who have had an immediate-type hypersensitivity reaction to beta-lactams, vancomycin is an acceptable alternative to penicillin (table 2) [3].The 2012 BSAC guidelines recommend that streptococcal endocarditis associated with strains with penicillin MICs >0.12 and 0.5 mg/L be treated with benzylpenicillin for four to six weeks combined with an aminoglycoside for the first two weeks of treatment [7]. Regimens used to treat enterococcal endocarditis are recommended for streptococcal endocarditis due to strains with MIC >0.5 mg/L. Both the ESC and the BSAC guidelines recommend that vancomycin be used for the treatment of streptococcal endocarditis due to strains with MICs >4 mg/L, although they acknowledge that there are no controlled trials to support these consensus opinions.

Endocarditis due to strains of viridans streptococci and streptococcal-like organisms (eg, Abiotrophia defectiva, Granulicatella spp, and Gemella spp) that have penicillin MICs >1 mcg/mL and are considered fully resistant to penicillin should be treated with regimens used to treat enterococcal endocarditis, as recommended by the AHA (table 3) [3,7-9]. A review of 29 patients with endocarditis due to penicillin-resistant viridans streptococci seen over a 39-year period concluded that current treatment guidelines should be successful in most patients. (See 'Enterococci' below.)OTHER STREPTOCOCCAL SPECIES Other streptococcal species (eg, groups A, B, C, G, and Streptococcus pneumoniae) are occasional causes of endocarditis [10,11]. Therapy should always be based on the results of susceptibility testing. Since most of these organisms are highly sensitive to penicillin, regimens used to treat endocarditis due to viridans streptococci are typically effective.

Many strains of groups B, C, and G streptococci are more resistant to penicillin than S. pyogenes. For this reason, some experts recommend adding gentamicin to a penicillin or cephalosporin for the first two weeks of a four to six week course of therapy [11].

Pneumococcus Pneumococcal endocarditis typically follows pneumonia in alcoholic patients and may be complicated by concurrent meningitis [10]. However, increasing numbers of isolates of S. pneumoniae have become relatively or highly resistant to penicillin. These organisms may be simultaneously resistant to other beta-lactams and other antimicrobial agents. Therapy in these cases usually requires the input of an infectious disease specialist or a microbiologist.

Penicillin G (24 million units per day either continuously or in four to six equally divided doses) remains the therapy of choice for patients with endocarditis due to penicillin-susceptible strains of S. pneumoniae and S. pyogenes [3]. The 2012 British Society for Antimicrobial Chemotherapy (BSAC) guidelines recommend that the therapy of pneumococcal endocarditis be guided by the same susceptibility breakpoints used to determine the therapy of endocarditis due to other streptococcal species [7]. (See 'Viridans streptococci and Streptococcus bovis' above.)

A few strains of S. pneumoniae only respond to vancomycin therapy. (See "Resistance of Streptococcus pneumoniae to beta-lactam antibiotics".)

Pneumococcal endocarditis is usually fulminant and causes severe valve damage and embolic complications. The potential complications of pneumococcal endocarditis are illustrated by a case series from France [12]. Transesophageal echocardiography (TEE) was performed in 28 of 30 patients with pneumococcal endocarditis; valvular vegetations, valve perforation, ring abscess, and other complications were detected in 97, 20, 13, and 20 percent, respectively. Early valve replacement may be necessary in patients with pneumococcal endocarditis to prevent complications. (See "Surgery for native valve endocarditis".)

ENTEROCOCCI Enterococci have a narrower spectrum of susceptibility than streptococcal species. In particular, members of the genus Enterococcus are all resistant to, at the least, low concentrations of penicillin. They are also relatively resistant to expanded spectrum penicillins, resistant to cephalosporins, and typically resistant to aminoglycosides at concentrations achieved after standard dosing regimens. However, many strains of enterococci are killed both in vitro and in vivo if penicillin, ampicillin, or vancomycin is given in synergistic combination with an aminoglycoside such as gentamicin.

There are also a number of additional trends in the resistance pattern of enterococci. (See "Mechanisms of antibiotic resistance in enterococci".)

Rare strains of Enterococcus faecalis have acquired genetic material that allows them to produce beta-lactamase [13,14].Other strains have acquired high-level resistance to streptomycin and/or gentamicin. The latter is accompanied by high-level resistance to synergy with tobramycin, netilmicin, and amikacin.High-level resistance to penicillin and ampicillin (minimum inhibitory concentration [MIC] 128 mcg/mL) that is not mediated by penicillinase has increased in some enterococci, primarily strains of Enterococcus faecium.Increasing numbers of enterococci have acquired high-level resistance to vancomycin as well as ampicillin. Some, but not all, of these highly resistant strains are also resistant to the glycopeptide antibiotic teicoplanin.These patterns of resistance have a direct impact on the guidelines for treatment of enterococcal endocarditis (table 3 and table 4 and table 5).

Susceptible strains The majority of cases of enterococcal endocarditis are caused by strains of E. faecalis [15]. The American Heart Association (AHA) and British Society for Antimicrobial Chemotherapy (BSAC) recommend that therapy for E. faecalis with typical low-level penicillin resistance consist of a combination of intravenous aqueous penicillin G or ampicillin plus gentamicin (table 3) [3,7]. In addition, the BSAC recommends either a combination of vancomycin and gentamicin and teicoplanin as alternatives to penicillin or ampicillin in the penicillin-allergic patient if the isolate is susceptible (MIC 4 mcg/L) [7]. The European Society for Cardiology (ESC) recommends combination therapy with either penicillin G plus gentamicin for four weeks or vancomycin plus gentamicin for six weeks (table 3) [16].

Although ampicillin is about one dilution more potent than penicillin in vitro, there is no evidence that this is necessary or beneficial for synergism in most instances. Because of the possibility of increased allergic reactions to ampicillin, some experts prefer penicillin rather than ampicillin for synergistic therapy. Gentamicin therapy is usually given at doses of 1 mg/kg every eight hours to achieve peak serum levels of approximately 3 to 4 mcg/mL and trough serum concentrations 16 mcg/mL) be treated with a combination of gentamicin plus either ampicillin-sulbactam (12 g per day in four equally divided doses) (if the resistance is beta-lactamase mediated) or vancomycin (15 to 20 mg/kg/dose every 8 to 12 hours, not to exceed 2 g per dose) for six weeks (table 4) [3]. The BSAC recommends the combination of gentamicin plus either vancomycin (1 g every 12 hours) or teicoplanin (10 mg/kg once daily, if MIC 4 mcg/L) for at least four weeks [7]. The ESC recommends vancomycin plus gentamicin for six weeks [16].

Unfortunately, most beta-lactamase-producing E. faecalis strains have also been highly resistant to streptomycin and gentamicin.

High-level aminoglycoside or vancomycin resistance E. faecium is the enterococcal strain most likely to manifest vancomycin resistance, but this strain is an uncommon cause of endocarditis. Therefore, literature on the treatment of patients with endocarditis due to vancomycin-resistant enterococci is limited to isolated case reports and extrapolation from in vitro susceptibility studies and animal models of endocarditis. Management of endocarditis due to enterococci with high-level aminoglycoside and/or vancomycin resistance may require individualized consideration by the microbiology laboratory as well as infectious disease consultation.

Results of in vitro susceptibility studies using various combinations of daptomycin and gentamicin have demonstrated conflicting results for both vancomycin-resistant enterococci and methicillin-resistant S. aureus (MRSA). Some studies have shown modest additive effects [19,20]; other studies showed synergy with some but not all strains of MRSA [21]. In view of these findings and the added risk of toxicity to gentamicin therapy, we do not favor using daptomycin and gentamicin in combination unless in vitro studies support addition of gentamicin and other less toxic regimens have failed.

Some strains of enterococci that are streptomycin resistant lack high-level resistance (HLR) to gentamicin and vice versa. However, HLR to gentamicin is conferred by a bifunctional enzyme that also confers high-level resistance to and/or resistance to synergism with tobramycin, netilmicin, and amikacin. When endocarditis is due to enterococci that are highly resistant to both streptomycin and gentamicin, the addition of an aminoglycoside is not beneficial. (See "Treatment of enterococcal infections".)

Success and failure of therapy for endocarditis due to resistant organisms has been described [22-24]. In a retrospective review of 12 cases of native valve vancomycin-resistant enterococcal endocarditis, all patients had significant comorbidities [22]. Three of the 12 patients died. Four patients were successfully treated with six to nine weeks of linezolid (two received linezolid alone, one each received an additional antibiotic, either alatrofloxacin or chloramphenicol).

Optimal therapy for such patients has not been defined, but a number of possible regimens were mentioned in the major society guidelines (table 5). Subsequent to publication of the guidelines, a six-week regimen of ampicillin (2 g IV every four hours) plus ceftriaxone (2 g IV every 12 hours) has been evaluated in patients with E. faecalis endocarditis [25,26] based upon demonstrated efficacy in experimental endocarditis due to E. faecalis strains that were highly resistant to aminoglycosides [27]. Among 246 patients in an observational, nonrandomized multicenter study, 159 were treated with ampicillin and ceftriaxone (AC) and 87 were treated with ampicillin and gentamicin (AG) [26]. AC appeared to be as effective as AG; there were no differences in mortality (during treatment or at three-month follow-up), treatment failure, or relapse. Interruption of treatment due to adverse events occurred more frequently among patients treated with AG, mainly due to renal failure.

In patients who do not respond to antimicrobial therapy, surgical resection of the involved valve may be necessary for cure. (See "Surgery for native valve endocarditis".)

STAPHYLOCOCCAL ENDOCARDITIS The success of therapy for staphylococcal endocarditis is dependent upon numerous factors, including involvement of right- versus left-sided valvular structures, whether the staphylococcus is coagulase negative or positive, the susceptibility of the staphylococcal isolate, and whether the infection occurs on a native or a prosthetic valve.

Rare, occasional strains of S. aureus are found to be penicillin susceptible; the strain should be tested to confirm the minimum inhibitory concentration (MIC) is 0.1 mcg/mL and that the strain does not produce beta-lactamase in vitro. If penicillin susceptibility is documented, penicillin (24 million units/day in four to six divided doses) can be substituted for nafcillin or another semisynthetic penicillin in the regimens discussed below.

Methicillin susceptible Native valve endocarditis due to methicillin-susceptible S. aureus is best treated with a semisynthetic penicillin, such as nafcillin or oxacillin (12 g per day intravenously in four to six equally divided doses) or flucloxacillin (2 g every four to six hours) (table 6).

Low-dose aminoglycosides should NOT be combined routinely with antistaphylococcal penicillins or vancomycin for treatment of left-sided native valve S. aureus endocarditis. Although in vitro and experimental models of endocarditis have demonstrated that combination therapy facilitates more rapid killing of methicillin-susceptible S. aureus than monotherapy, the evidence for clinically significant benefit is minimal. This was illustrated by a randomized trial of 48 patients with methicillin-susceptible S. aureus (MSSA) native valve endocarditis; patients who received nafcillin plus gentamicin for the first two weeks of therapy had more rapid clearing of bacteremia than those who received nafcillin alone [28]. However, cure rates were comparable and combination nafcillin and gentamicin therapy was associated with a higher incidence of renal dysfunction.

Subsequently, a randomized trial including 236 patients with S. aureus bacteremia and endocarditis demonstrated that daptomycin monotherapy is not inferior to low-dose gentamicin plus an antistaphylococcal penicillin or vancomycin, although those in the standard therapy arm experienced significantly more renal impairment than those in the daptomycin arm [29]. An investigation of the safety data from the trial noted significantly greater reduction in creatinine clearance among those who received initial low-dose gentamicin than those who did not (22 versus 8 percent, respectively) [30].

Therefore, initial low-dose aminoglycoside should NOT be combined routinely with antistaphylococcal penicillins or vancomycin for treatment of S. aureus endocarditis. The limited benefit must be weighed against evidence for potential harm, particularly in older adults and in those with diabetes and mild baseline renal dysfunction. (See "Clinical approach to Staphylococcus aureus bacteremia in adults".)

In adults, six weeks of therapy is recommended for complicated right-sided infective endocarditis (IE) and for all left-sided IE; complicated IE is defined as metastatic infections or when the course is otherwise complicated by secondary cardiac problems (eg, heart failure). In patients with uncomplicated right-sided IE, the duration of therapy is two weeks if synergistic therapy can be given. (See 'Uncomplicated right sided' below.)

In children, six weeks of therapy is recommended regardless of the site of infection or absence of complications because of insufficient clinical experience to support a duration of therapy of two weeks for uncomplicated right-sided IE [31].

Penicillin allergy Patients with native valve endocarditis due to S. aureus who have a history of penicillin allergy can be treated with a first-generation cephalosporin, such as cefazolin (2 g intravenously every eight hours), if there is no prior history of penicillin reaction that is typical of an immediate-type allergy. Some experts caution against substituting cefazolin for nafcillin in the treatment of staphylococcal endocarditis [32]. However, to our knowledge, the only study purporting to show that cefazolin was inferior to nafcillin in staphylococcal endocarditis was based on anecdotal observations of two patients with S. aureus endocarditis who either relapsed after treatment or who failed therapy [33]. The American Heart Association (AHA) recommends cefazolin as an alternative in patients with penicillin allergy that is not anaphylactoid type [3]; however, the British Society for Antimicrobial Chemotherapy (BSAC) and the European Society for Cardiology (ESC) recommend vancomycin therapy for all patients with penicillin allergy regardless of type [7,16]. In addition, the BSAC 2012 guidelines recommend adding rifampin (300 to 600 mg orally every 12 hours) to vancomycin when treating methicillin-susceptible native valve endocarditis in patients with penicillin allergy [7].

Vancomycin is an acceptable alternative in the patient with immediate-type penicillin allergy. However, vancomycin should not be used on the basis of convenience related to pharmacokinetics in patients without a history of penicillin allergy, since clinical experience and in vitro studies have suggested that vancomycin is a less effective antistaphylococcal antibiotic than nafcillin or oxacillin [3,34].

Clindamycin is not an acceptable alternative for staphylococcal endocarditis because relapse is common [3].

Uncomplicated right sided Selected patients with native valve right-sided endocarditis due to S. aureus with no evidence of renal failure, extrapulmonary metastatic infections, or simultaneous left-sided valvular infection may be successfully treated with two-week regimens utilizing the combination of nafcillin and gentamicin [3,35]. Regimens that substitute vancomycin or teicoplanin for nafcillin (eg, for penicillin allergic patients) are not considered to be reliably effective if only two weeks of therapy are given [32]. Although one randomized study showed that cloxacillin alone for two weeks was equivalent to the combination of cloxacillin plus gentamicin in the treatment of right-sided IE in injection drug users, we do not routinely advise short-course monotherapy for such patients [36].

Short-course regimens utilizing combination therapy are also not suitable for patients with simultaneous infection of the left-side heart valves, isolates that demonstrate high-level gentamicin resistance (MIC >500 mcg/mL), metastatic infections outside of the lungs, or in patients who fail to defervesce in a normal time period.

Methicillin resistant Native valve endocarditis due to either methicillin-resistant S. aureus (MRSA) or coagulase-negative staphylococci should be treated with vancomycin for six weeks (table 6) [3,16,37]. However, there have been a number of reports of vancomycin treatment failure in serious infections due to MRSA even when isolates are proven to be susceptible using current microbiological testing methods [38-40].

Some guidelines [7] recommend the addition of rifampin (300 to 600 mg orally every 12 hours) to vancomycin in the treatment of native valve endocarditis due to methicillin-resistant but vancomycin-susceptible S. aureus. However, the evidence for this approach is not conclusive and the risk of rifampin-induced drug interactions or hepatic toxicity should be weighed before utilizing such therapy.

Gentamicin should NOT be combined with vancomycin for treatment of MRSA native valve IE [37]. (See 'Methicillin susceptible' above.)

Daptomycin is an acceptable alternative to vancomycin [37,41]. In a randomized trial of 246 patients with S. aureus bacteremia (SAB) with or without endocarditis, daptomycin (6 mg/kg intravenously [IV] per day) was not inferior to standard therapy for SAB or right-sided endocarditis [29]. Daptomycin resistance (MIC 2 mcg/mL) developed in six patients. Another study demonstrated that clearance of bacteremia among 29 patients with gram-positive left-sided IE treated with daptomycin was significantly faster than among 149 patients treated with conventional therapy (p6 days) may be required to detect their growth. This delayed growth made these organisms synonymous with culture-negative endocarditis. Since the introduction of automated blood culture systems, the HACEK organisms are easily isolated when incubated for five days. (See "Epidemiology, microbiology, and diagnosis of culture-negative endocarditis".)

Although most HACEK organisms were ampicillin sensitive in the past, this is no longer true, as many species in this group have acquired the ability to produce beta-lactamase. However, virtually all of these organisms, even strains that produce beta-lactamase, are highly susceptible to third-generation cephalosporins, such as ceftriaxone (2 g once daily, given intravenously or intramuscularly).

The American Heart Association (AHA) recommends treatment with ceftriaxone, ampicillin-sulbactam, or ciprofloxacin for four weeks (table 7) [3]. The British Society for Antimicrobial Chemotherapy (BSAC), in contrast, recommends combination therapy with either ampicillin (if susceptible) or ceftriaxone for four weeks with gentamicin added for the initial two weeks of treatment (table 7) [7]. The European Society for Cardiology (ESC) recommends either ceftriaxone alone or ampicillin plus gentamicin; both regimens are given for three to four weeks (table 7) [16].

OTHER GRAM-NEGATIVE ORGANISMS Native valve endocarditis due to other gram-negative bacilli, such as E. coli, Pseudomonas, or mucoid strains of Klebsiella or Serratia, is extremely rare. The most common predisposing factor is an implanted endovascular device, and most cases occur in the setting of recent healthcare contact [49].

The choice of antimicrobial therapy is dependent on the antimicrobial susceptibility of the causative organism. Third-generation cephalosporins, such as ceftriaxone (2 g once daily intravenously [IV]), or a fluoroquinolone, such as ciprofloxacin (400 mg every 12 hours IV), are acceptable regimens if the organism is susceptible in vitro. Combination antimicrobial therapy with an antipseudomonal penicillin and an aminoglycoside is an effective regimen for patients with infections due to Pseudomonas [3].

CULTURE-NEGATIVE ENDOCARDITIS Blood culture-negative infective endocarditis (IE) is defined as endocarditis without etiology following inoculation of three blood samples in a standard blood culture system (eg, negative cultures after seven days).

Cultures are negative in infective endocarditis for three major reasons:

Previous administration of antimicrobial agentsInadequate microbiological techniquesInfection with highly fastidious bacteria or nonbacterial pathogens (eg, fungi)The most common causative agents of blood culture-negative IE are fastidious organisms (eg, zoonotic agents and fungi) and Streptococcus spp in patients who have received previous antibiotic treatment. (See "Epidemiology, microbiology, and diagnosis of culture-negative endocarditis".)

Coxiella burnetii and Bartonella spp are relatively commonly agents of culture-negative endocarditis, although the frequency varies in different geographic locations. Treatment of these cases is discussed separately. (See "Q fever endocarditis" and "Endocarditis caused by Bartonella".)

Empiric treatment of patients with culture-negative endocarditis should cover both gram-positive and gram-negative organisms. The American Heart Association (AHA) recommends treatment with either ampicillin-sulbactam plus gentamicin OR vancomycin plus gentamicin plus ciprofloxacin for four to six weeks (table 8) [3]. The British Cardiac Society (BCS) and the European Society for Cardiology (ESC), in contrast, recommend combination therapy with vancomycin for six (BCS) or four to six (ESC) weeks with gentamicin added for the initial two weeks of treatment (table 8) [16,50]. When diagnostic tests (eg, polymerase chain reaction or serology) identify the etiologic agent, therapy should be directed to the specific microorganism. (See "Clinical manifestations and diagnosis of infective endocarditis".)

Complications and outcome of infective endocarditisAuthorsDenis Spelman, MBBS, FRACP, FRCPA, MPHDaniel J Sexton, MDSection EditorsStephen B Calderwood, MDGabriel S Aldea, MDScott E Kasner, MDDeputy EditorsElinor L Baron, MD, DTMHSusan B Yeon, MD, JD, FACCDisclosures: Denis Spelman, MBBS, FRACP, FRCPA, MPH Nothing to disclose. Daniel J Sexton, MD Grant/Research/Clinical Trail Support: Cubist [C. difficile infection (Fidaxomycin)]. Consultant/Advisory Boards: Johnson & Johnson [Pelvic mesh-related infection]; Sterilis [Medical waste disposal systems]; Magnolia Medical Technologies [Intravenous devices]. Other Financial Interest: National Football League [Infection control program]. Equity Ownership/Stock Options: Magnolia Medical Technologies [Intravenous devices]. Stephen B Calderwood, MD Patent Holder: Vaccine Technologies Inc. [Vaccines (Cholera vaccines)]. Equity Ownership/Stock Options: Pulmatrix [Inhaled antimicrobials]; PharmAthene [Anthrax (Anti-protective antigen monoclonal antibody)]. Gabriel S Aldea, MD Nothing to disclose. Scott E Kasner, MD Grant/Research/Clinical Trial Support: WL Gore and Associates [PFO, stroke (HELEX, GSO devices)]; AstraZeneca [Stroke (Ticagrelor)]. Consultant/Advisory Boards: Medtronic [Stroke, atrial fibrillation (CoreValve, REVEAL)]; Merck [Stroke]; Pfizer [Stroke]; Novartis [Stroke]; GSK [Stroke]; AbbVie [Stroke]; Daiichi Sankyo [Stroke]; Boehringer Ingelheim [Stroke]. Elinor L Baron, MD, DTMH Nothing to disclose. Susan B Yeon, MD, JD, FACC Nothing to disclose. Contributor disclosures are reviewed for conflicts of interest by the editorial group. When found, these are addressed by vetting through a multi-level review process, and through requirements for references to be provided to support the content. Appropriately referenced content is required of all authors and must conform to UpToDate standards of evidence. Conflict of interest policyAll topics are updated as new evidence becomes available and our peer review process is complete.Literature review current through: Apr 2015. | This topic last updated: Apr 28, 2015.INTRODUCTION Infective endocarditis (IE) is associated with a broad array of complications. The likelihood of complication(s) depends on several factors including the infecting pathogen, duration of illness prior to therapy, and underlying comorbidities [1]. Complications can occur before, during, and after completion of therapy.

It can be difficult to assess the true incidence of complications since case series are frequently retrospective. In one review including 223 episodes of IE, 57 percent of patients had one complication, 26 percent had two, 8 percent had three or more, and 6 percent had six or more complications [2].

Issues related to complications and outcome of IE will be reviewed here. Other issues related to IE are discussed separately. (See related topics.)

COMPLICATIONS OF ENDOCARDITIS Complications of infective endocarditis (IE) include cardiac, neurologic, renal, and musculoskeletal complications, as well as complications related to systemic infection (including embolization, metastatic infection, and mycotic aneurysm). More than one complication can occur simultaneously.

Complications can also be considered based on pathogenesis (ie, embolic [such as cerebral infarct], local spread of infection [such as heart valve destruction], metastatic infection [such as vertebral osteomyelitis], and immune-mediated damage [such as glomerulonephritis]).

IE due to Staphylococcus aureus is associated with complications more frequently than other pathogens (stroke 21 versus 14 percent, systemic emboli 27 versus 18 percent, persistent bacteremia 17 versus 5 percent, and in-hospital mortality 22 versus 14 percent) [3].

Cardiac complications Cardiac complications are the most common complications in IE; they occur in up to half of patients [4].

Heart failure Heart failure is the most common cause of death due to IE in the modern era. The correlation between heart failure and mortality depends on several factors including the valve involved (aortic valve involvement is more likely to cause heart failure than mitral valve involvement), the pathogen involved (S. aureus infection increases the likelihood of heart failure), and whether the patient undergoes surgery for valve repair or replacement. Heart failure is the most common reason for cardiac surgery in patients with IE; indications for surgery in patients with cardiac complications are discussed further separately. (See "Surgery for native valve endocarditis".)

The usual cause of heart failure in patients with IE is valvular insufficiency resulting from infection-induced valvular damage. Rarely, embolized fragments of valvular vegetations or vegetation-induced stenosis of the coronary ostia can cause acute myocardial infarction and subsequent heart failure [4].

Perivalvular abscess The reported incidence of perivalvular abscess among patients with IE is 30 to 40 percent [5-7]. The aortic valve and its adjacent annulus are more susceptible to abscess formation and associated complications than the mitral valve and annulus [5-7]. This was illustrated in an autopsy study including 95 patients with native valve endocarditis; annular extension of infection was more common in patients with aortic valve compared with mitral valve endocarditis (41 versus 6 percent) [5].

Perivalvular abscesses can extend into adjacent cardiac conduction tissues, leading to heart block. Involvement of the conducting system is most common in the setting of aortic valve infection, especially when there is involvement of the valve ring between the right and non-coronary cusp; this anatomic site overlies the intraventricular septum that contains the proximal ventricular conduction system. Rarely, perivalvular infection can result in extrinsic coronary compression and can cause acute coronary syndrome [8].

Perivalvular abscess is associated with increased risk of systemic embolization and death. In one study including 73 patients with IE, the embolization rate was approximately twice as high among patients with perivalvular abscess (64 versus 30 percent) [6]. Another study including 118 patients with IE noted higher mortality among patients with perivalvular abscess (23 versus 14 percent) [7]. In addition, moderate or severe regurgitation is associated with higher mortality rate than normal valvular function [9].

Perivalvular abscess should be suspected in the setting of conduction abnormalities on electrocardiogram (ECG) and/or persistent bacteremia or fever despite appropriate antimicrobial therapy [10]. Data are conflicting regarding correlation between vegetation size and risk for perivalvular abscess. Large vegetation size had been implicated as a risk factor for perivalvular abscess in some series, although subsequent studies have shown no correlation [6,11]. Patients with IE involving congenital bicuspid aortic valves appear to be more prone to perivalvular complications than those with IE involving tricuspid aortic valves [12]. Injection drug use may be another risk factor for perivalvular abscess [6].

Transesophageal echocardiography (TEE) is more sensitive for detection of myocardial abscess than transthoracic echocardiography (TTE). In one study including 43 patients with perivalvular abscess documented at surgery or autopsy, the sensitivity, specificity, and positive and negative predictive values of TEE were 87, 95, 91, and 92 percent, respectively [7]. The sensitivity of TTE was much lower (28 percent), although the specificity was 99 percent. While TEE is more sensitive than TTE for detecting an abscess, even TEE may miss a significant number of abscesses in some populations (eg, mitral annular abscesses in patients with large mitral annular calcification). (See "Role of echocardiography in infective endocarditis", section on 'Perivalvular abscess or fistula'.)

Other cardiac complications Pericarditis (suppurative or nonsuppurative) can cause chest pain or cardiac tamponade [13,14]. (See "Purulent pericarditis" and "Clinical presentation and diagnostic evaluation of acute pericarditis".)

Intracardiac fistula (eg, aorta-atrial or aorta-ventricular) may develop due to extension of infection from the valve to adjacent myocardium. Rarely, this can lead to development of an aneurysm, aortic dissection, or myocardial perforation [15,16]. In one study including 4681 episodes of IE, the incidence of fistulous intracardiac complications was 1.6 percent [17].

Metastatic infection Forms of metastatic infection include embolization, metastatic abscess, and mycotic aneurysm.

Septic embolization General issues related to embolization in patients with IE will be reviewed here; issues related to embolization in patients with IE who undergo surgery are discussed separately, as are issues related to thrombotic therapy. (See "Surgery for native valve endocarditis", section on 'Embolization' and "Surgery for prosthetic valve endocarditis" and "Antithrombotic therapy in patients with infective endocarditis".)

Embolization with clinical sequelae has been described in 13 to 44 percent of patients with IE; in most cases, embolization occurs prior to clinical presentation but can occur after initiation of antimicrobial therapy [18-20]. Systemic embolization most commonly occurs in left-sided IE; pulmonic embolization most commonly occurs in right-sided IE. Systemic embolization can also occur in patients with right-sided IE and a patent foramen ovale [21].

Emboli can occlude or damage virtually any vessel in the systemic or pulmonary arterial circulation. As a result, embolization can cause:

StrokeParalysis (due to embolic infarction of either the brain or spinal cord)Blindness (due to embolism or due to endophthalmitis as a result of bacteremic seeding)Ischemia of the extremitiesSplenic or renal infarctionPulmonary embolismAcute myocardial infarctionEndocarditis should be considered as a possible etiology in patients who present with signs or symptoms of systemic arterial embolization. Most patients with acute stroke do not have endocarditis, although the likelihood of IE is increased in relatively young patients and in patients with both cerebral and systemic arterial embolization [22,23].

Risk factors Risk factors for embolization include presence of left-sided vegetation, large vegetation size, microbiology, presence of antiphospholipid antibodies, age, diabetes, atrial fibrillation, and embolization prior to initiation of antibiotics [24-31]:

Left-sided vegetation Embolization with clinical sequelae occurs more frequently in patients with left-sided than right-sided vegetations [25], and more frequently with mitral vegetations than aortic vegetations [26]. In one review including 281 patients with suspected IE, the incidence of embolic events was greater with mitral than aortic valve vegetations (25 versus 10 percent) [26]. The risk was highest among patients with a vegetation on the anterior mitral leaflet (37 percent), suggesting that the mechanical effects of broad and abrupt leaflet excursion may contribute to the risk of embolization [25].Large vegetation size One study including 384 patients with IE noted that vegetation size >10 mm and severe mobility of the vegetation were predictors of new embolic events, even after adjustment for pathogen type, and vegetation size >15 mm was a predictor of one-year mortality (adjusted relative risk 1.8, CI 1.10-2.82) [30]. In another study including 59 patients with IE due to S. aureus, the risk of embolization was greater in patients who had visible vegetations by both TTE and TEE compared with patients who had vegetations visualized only by TEE [27]. Another study noted that absence of valvular abnormalities on TTE may be associated with reduced incidence of complications [28]. (See "Role of echocardiography in infective endocarditis", section on 'Echocardiographic estimation of outcome'.)Microbiology The likelihood of symptomatic embolization is increased in the setting of IE due to fungal pathogens or S. aureus. In a review including 270 patients with fungal endocarditis, peripheral arterial embolization occurred in 45 percent of cases [32]. The most common sites were the cerebral circulation (17 percent) and femoral artery (16 percent). Another study including 384 patients with IE noted that emboli were more frequently observed in the setting of infection due to S. aureus and Streptococcus bovis infection [30].Presence of antiphospholipid antibodies In one series including 91 patients with IE, the presence of antiphospholipid antibodies was associated with an increased risk of embolization (62 versus 23 percent); this may be due to increased endothelial cell activation, generation of thrombin, and defective fibrinolysis [29].Effect of antibiotic therapy The risk of embolization declines after institution of appropriate antimicrobial therapy, and serious embolic events weeks after such therapy is instituted are rare [18,33,34]. In one study including 1437 patients with left-sided IE receiving appropriate therapy, the incidence of stroke fell from 4.8 to 1.7 per 1000 patient days between the first and second week of therapy [34]. These findings suggest that surgery may not be necessary for prevention of embolic stroke in the early weeks following initiation of appropriate therapy if there are no other surgical indications (such as a large mobile vegetation or congestive heart failure due to valvular regurgitation).

Metastatic abscess Development of metastatic abscess occurs as a sequela of septic embolization; this may occur in the spleen, kidneys, brain, and/or soft tissues (such as the psoas muscle).

Patients with splenic abscess do not always have significant abdominal pain or splenomegaly on physical examination; in some cases, the only clue(s) may be persistent fever and/or recurrent bacteremia after completion of antimicrobial therapy [35]. Splenic abscess frequently requires splenectomy for cure. In one series including 27 patients with splenic abscesses, mortality among 17 patients who underwent splenectomy was 18 percent; mortality among 10 patients who did not undergo splenectomy was 100 percent [36]. (See "Approach to the adult patient with splenomegaly and other splenic disorders", section on 'Splenic abscess'.)

Solitary brain abscess(es) or discrete microabscess(es) can also occur; purulent meningitis has been observed in some cases. The presence of meningitis due to S. aureus should suggest the possibility of concomitant S. aureus endocarditis. In one series including 33 patients with S. aureus meningitis, IE was diagnosed in 21 percent of cases [37]. Abscess drainage is required to control local infection and to prevent ongoing bacteremia. (See "Treatment and prognosis of bacterial brain abscess".)

Mycotic aneurysm Mycotic aneurysm can develop in the cerebral or systemic circulation in the setting of IE, usually at points of vessel bifurcation. Although some authors use the term "mycotic" to describe infected aneurysm regardless of etiology, we limit the use of this term to those aneurysms that develop when material originating in the heart causes arterial wall infection and, subsequently, dilation [38]. (See "Overview of infected (mycotic) arterial aneurysm".)

Neurologic complications Manifestations of neurologic complications include:

Embolic strokeBrain abscess or cerebritisPurulent or aseptic meningitisAcute encephalopathyMeningoencephalitisCerebral hemorrhage (due to stroke or a ruptured mycotic aneurysm)Seizures (secondary to abscess or embolic infarction)Symptomatic cerebrovascular complications occur in up to 35 percent of patients [20,21,39-44]. Silent cerebrovascular complications (including ischemia and microhemorrhage) may occur in up to 80 percent of patients [20,45-47].

Neurologic complications may be the presenting symptom in patients with IE, and the possibility of IE should be considered in patients who present with stroke, meningitis, or brain abscess. In one series including 68 patients with stroke and IE, for example, two-thirds presented with stroke prior to diagnosis of endocarditis [48]. Unexplained fever accompanying a stroke in a patient with valvular disease can be an important clue for IE. (See "Clinical manifestations and diagnosis of infective endocarditis".)

Among patients with IE, the incidence of cerebral emboli detected by magnetic resonance imaging (MRI) is markedly higher than the incidence of emboli detected based on clinical signs and symptoms [20,47,49]. In one study including 65 patients with IE, clinical findings consistent with embolism were observed in 20 percent of cases; among patients with no symptoms, emboli were detected on MRI in 46 percent of cases [49]. In another study including 60 patients with IE, clinical signs and symptoms of cerebral embolism were observed in 35 percent of cases; clinically silent emboli were detected on MRI in an additional 30 percent of patients [20].

Patient outcomes after a neurologic complication of IE are variable. In one study including 214 patients who underwent cardiac surgery for IE, 70 percent of patients with a preoperative stroke had full neurologic recovery; outcomes were worse in patients with stroke complicated by meningitis, abscess, or intracerebral hemorrhage [43]. Patient mortality in another series including 68 patients with stroke and IE was 50 percent at one year [48]. One study of 91 patients with S. aureus IE and neurologic manifestations noted mortality of 74 percent [41].

Renal complications Renal complications of IE include renal infarction or abscess following septic embolization, glomerulonephritis (due to deposition of immunoglobulins and complement in the glomerular membrane), and drug-induced acute interstitial nephritis. Acute renal failure, defined as a serum creatinine of 2 mg/dL (177 mcmol/L), has been described in up to one-third of patients [50]. Immune complex-mediated renal disease is uncommon in the antibiotic era, especially in patients whose infection is detected and treated early. Chronic renal failure due to immune-complex-mediated glomerulonephritis was a common cause of death in patients with IE in the preantibiotic era; in the modern era, it is rare. (See "Renal disease in the setting of infective endocarditis or an infected ventriculoatrial shunt".)

Renal complications can also occur as a result of toxicity associated with administration of therapeutic drugs; this is particularly important in the setting of coadministration of vancomycin with aminoglycosides and among patients at the extremes of age. (See 'Complications related to therapy' below.)

Musculoskeletal complications Musculoskeletal complications of IE include vertebral osteomyelitis and septic arthritis. Back pain in patients with IE should prompt consideration of vertebral osteomyelitis, particularly in the setting of S. aureus infection [51]. (See "Vertebral osteomyelitis and discitis in adults".)

Clues to the presence of septic arthritis in the setting of IE include involvement of multiple joints and involvement of the axial skeleton (eg, sacroiliac, pubic, or manubriosternal joints). Acute septic arthritis involving one or more joints may be the first clue to the presence of IE in a small percentage of patients, particularly for cases in which an organism with known propensity to cause IE (such as S. aureus, S. viridans, or non-group A beta-hemolytic streptococci) grows from a joint aspirate.

Complications related to therapy Patients with infective endocarditis can develop a number of the complications associated with prolonged parenteral antimicrobial therapy or surgery:

Aminoglycoside-induced ototoxicity or nephrotoxicity, particularly in patients who receive simultaneous treatment with vancomycin and aminoglycosides (see "Pathogenesis and prevention of aminoglycoside nephrotoxicity and ototoxicity")Secondary bacteremia due to central vascular lines (see "Epidemiology, pathogenesis, and microbiology of intravascular catheter infections")Mediastinitis or early postoperative prosthetic valve endocarditis (see "Postoperative mediastinitis after cardiac surgery")Intravenous catheter-associated thrombosis (see "Catheter-related upper extremity venous thrombosis")Drug fever (see "Drug fever")Allergic or idiosyncratic reactions to various antimicrobial agents. Some agents (such as beta-lactams) that are generally tolerable for short courses may not be associated with significant reactions (such as pyelonephritis and acute interstitial nephritis) in the setting of long-term therapy.Bleeding due to disturbances in coagulation caused by anticoagulants (eg, in prosthetic valve endocarditis) (see "Antithrombotic therapy in patients with infective endocarditis")OUTCOME Among patients with infective endocarditis (IE), in-hospital mortality rate is 18 to 23 percent; the six-month mortality rate is 22 to 27 percent [52-56]. The outcomes in patients with neurologic complications are described above. (See 'Neurologic complications' above.)

Some data suggest that the following characteristics appear to confer increased risk of mortality in patients with infective endocarditis:

Microbiology (for example, mortality is higher in the setting of S. aureus infection than streptococcal infection) [53,55,57]Heart failure [54,56]Diabetes mellitus [53]Embolization [53,57]Perivalvular abscess [7,9,56]Larger vegetation size [30,57]Female gender [30]Low serum albumin [52]Persistent bacteremia [56]Abnormal mental status [55]Poor surgical candidacy [55]The association between cardiac surgery, heart failure, and mortality risk is uncertain. In one prospective report, neither cardiac surgery nor heart failure was independently associated with in-hospital mortality [53]. However, among patients with moderate to severe heart failure, other studies have noted an association between cardiac surgery and lower mortality rate compared with medical therapy alone. In one study including approximately 260 patients, the presence of three risk factors (heart failure, periannular complications, and S. aureus infection) during the first 72 hours of hospitalization was predictive of need for urgent surgery or in-hospital mortality [58]. Subsequent prospective validation of the model noted that one risk factor conferred approximately 60 percent risk of adverse outcome; three risk factors conferred nearly 100 percent risk of adverse outcome.

Issues related to the association between cardiac surgery and clinical outcome are discussed further separately. (See "Surgery for native valve endocarditis", section on 'Efficacy'.)