Diagnostic approach to infective endocarditis

http://www.uptodateonline.com/online/content/topic.do?topicKey=endocard/6293&selectedTitle=10~150&source=search_result

Diagnostic approach to infective endocarditis

Author
Daniel J Sexton, MD
Section Editor
Catherine M Otto, MD
Deputy Editor
Elinor L Baron, MD, DTMH



Last literature review version 16.2: May 2008 | This topic last updated: March 21, 2008 (More)


INTRODUCTION — The diagnosis of infective endocarditis (IE) is usually based upon a constellation of clinical findings rather than a single definitive test result. The diagnosis is usually obvious when a patient has the characteristic findings of IE:

• Numerous positive blood cultures in the presence of a well recognized predisposing cardiac lesion
• Evidence of endocardial involvement


However, some patients with IE do not have positive blood cultures (ie, culture-negative endocarditis), and approximately one-third to one-fourth of patients have no identifiable predisposing cardiac lesion at disease onset. The presence of atypical features may result in misdiagnosis or a correct diagnosis that is greatly delayed. (See "Culture-negative endocarditis" below).

The general approach to the diagnosis of IE will be reviewed here. Risk factors for IE and antibiotic prophylaxis and treatment of IE are discussed separately. (See "Infective endocarditis: Epidemiology and risk factors" and see "Antimicrobial prophylaxis for bacterial endocarditis" and see "Antimicrobial therapy of native valve endocarditis" and see "Antimicrobial therapy of prosthetic valve endocarditis").

DIAGNOSTIC CRITERIA — The diagnosis of IE is based upon a careful history and physical examination, blood culture and laboratory results, an electrocardiogram (ECG), a chest radiograph, and an echocardiogram.

Debate persists as to the optimal case definition for IE. Practical and logical case definitions are important since underdiagnosis can lead to clinical catastrophe and death, while overdiagnosis can result in weeks of unnecessary antimicrobial therapy with excessive costs and potentially avoidable drug-related side effects. It can, for example, be difficult to distinguish between IE and an alternate source of infection in a bacteremic patient with underlying heart disease.

Several sets of criteria for IE have been described. The most commonly accepted are the Duke criteria (show table 1 and show table 2 and show Calculator). (See "Infective endocarditis: Historical and Duke criteria", section on Duke criteria).

HISTORY — During the initial assessment of patients with suspected endocarditis, a careful history should be performed with special attention given to a history of prior cardiac lesions and historical clues pointing toward a recent source of bacteremia, such as indwelling intravascular catheters or intravenous drug use.

PHYSICAL EXAMINATION — 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 classic clinical stigmata of endocarditis, including evidence of small and large emboli with special attention to the fundi, conjunctivae, skin, and digits. A neurologic evaluation may reveal evidence of focal neurologic impairment; it can also be used as a baseline examination should such abnormalities appear later. (See "Complications and outcome of infective endocarditis").

Associated peripheral cutaneous or mucocutaneous lesions of IE include petechiae, splinter hemorrhages, Janeway lesions, Osler's nodes, and Roth spots. Petechiae are not specific for IE but are its most common skin manifestation. They may be present on the skin, usually on the extremities, or on mucous membranes such as the palate or conjunctivae, the latter usually as hemorrhages best seen with eversion of either upper or lower eyelids. Splinter hemorrhages, also nonspecific for endocarditis, are nonblanching, linear reddish-brown lesions found under the nail bed (show picture 1).

Janeway lesions, Osler's nodes, and Roth spots are more specific (but still not diagnostic) for IE. They are also less common, and Roth spots are rare.

• Janeway lesions are macular, blanching, nonpainful, erythematous lesions on the palms and soles (show picture 2).
• 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 (show picture 3).
• Roth spots are exudative, edematous hemorrhagic lesions of the retina.

In addition to these physical findings, patients with IE may have involvement of other organs due to embolic events (eg, focal neurologic deficits, renal and splenic infarcts) or a systemic immune reaction (eg, glomerulonephritis, arthritis). In right-sided endocarditis, septic pulmonary infarcts may be seen (show picture 4). (See "Complications and outcome of infective endocarditis" and see "Renal disease in infective endocarditis").

LABORATORY STUDIES

Blood cultures

Collection — Blood cultures should be obtained prior to antibiotic therapy. This was illustrated in a prospective study of 348 patients with suspected culture-negative endocarditis; among the 73 patients without an identifiable etiologic agent, 58 (79 percent) received antibiotics before blood cultures were obtained [9] .

A minimum of three blood cultures should be obtained over a time period based upon the severity of the illness. If the tempo of illness is subacute and the patient is not critically ill, it is reasonable and often preferable to delay therapy for one to three days while awaiting the results of blood cultures and other diagnostic tests. However, if the patient is acutely ill, three blood cultures should be obtained over a one hour time span before beginning empiric therapy. Almost all cases of bacterial IE are due to aerobic organisms; thus, culturing for anaerobes is rarely useful. (See "Blood cultures for the detection of bacteremia").

The additional diagnostic yield of more than three cultures is minimal in patients who have not recently received antimicrobial therapy. In a series of 206 cases of endocarditis, for example, the initial blood culture in patients with streptococcal endocarditis was positive in 96 percent and one of the first two blood cultures was positive in 98 percent; in patients with IE caused by bacteria other than streptococci, the first blood culture was positive in 82 percent and one of the first two cultures was positive in 100 percent [1] . Additional blood cultures are occasionally useful in patients who have been treated recently with antibiotics.

The bacteriologic diagnosis of IE is facilitated by the relative constancy, rather than random, release of bacteria from the cardiac vegetations [2] . However, since many patients with bacterial endocarditis have low grade bacteremia (eg, 1 to 10 CFU/mL of blood) [1] , 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. Blood cultures can be taken at any time; they do not need to be obtained with the appearance of chills or fever since patients with IE typically have a continuous bacteremia. (See "Blood cultures for the detection of bacteremia").

Organism — Not all microorganisms have the same propensity to cause endocarditis. As an example, organisms such as viridans streptococci and Staphylococcus aureus are more likely to cause endocarditis than are gram-negative rods such as Escherichia coli and Proteus spp. This distinction is important. The Duke Criteria for the diagnosis of endocarditis define the following organisms as "typical causes" of IE (show table 1 and show table 2) [4] :

• Staphylococcus aureus
• Viridans streptococci and Streptococcus bovis
• Enterococci
• HACEK group organisms (show table 2)

It was previously thought that patients who present with community-acquired enterococcal bacteremia are significantly more likely to have endocarditis than patients who develop enterococcal bacteremia while hospitalized for another cause [5] . However, this conclusion has been questioned as a result of a case-control study that compared the clinical and demographic characteristics of all patients with enterococcal endocarditis seen at a single center over eight years with controls randomly chosen from 455 patients with enterococcal bacteremia without endocarditis [6] . Community acquisition of bacteremia was not a risk factor for IE.

The probability of endocarditis varies by species of bacteria. As examples:

• S. sanguis bacteremia is more often indicative of endocarditis than is bacteremia due to S. milleri (also known as S. anginosus).
• Bacteremia with groups A or C streptococci are seldom associated with IE whereas group G streptococcal infection is often indicative of endocarditis [7] . (See "Group C and group G streptococcal infection").
• In a study of enterococcal bacteremia cited above, infection with E. faecalis, compared to other enterococcal species, was associated with IE by both univariate and multivariate analysis [6] .


The 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 "Complications of Staphylococcus aureus bacteremia").

Positive cultures — The interpretation of positive blood cultures in patients with suspected endocarditis is confounded by the fact that false-positive results occasionally occur despite use of the most exacting techniques for collection and processing. When organisms such as Propionibacterium spp., Corynebacterium spp., Bacillus spp., and coagulase-negative staphylococci are recovered from a single blood culture or a minority of blood culture bottles, the result is probably falsely positive. However, since all of these organisms are capable of causing endocarditis, it is important to determine if the bacteremia is persistent. (See "Endocarditis due to coagulase-negative staphylococci").

The definition of persistent bacteremia varies with the likelihood that the organism is a cause of endocarditis [4] :

• For an organism likely to cause endocarditis (eg, S. aureus, viridans streptococci), two positive samples collected more than 12 hours apart
• For an organism that is more commonly a skin contaminant, three or a majority of four or more separate blood cultures are positive and the first and last samples are collected at least one hour apart



Additional tests — The utility of other laboratory tests in the diagnosis of endocarditis is limited, other than the finding of an antiphase I IgG titer >1:800 for Coxiella burnetii [8] .

The following findings may be identified among patients with IE but are relatively nonspecific:

• An elevated erythrocyte sedimentation rate and/or an elevated level of C-reactive protein.
• A normochromic normocytic anemia.
• The 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, cryoglobulins, circulating immune complexes, hypocomplementemia, elevated rheumatoid factor titers, and false positive serologic tests for syphilis all occur in some patients.


An elevated rheumatoid factor titer in patients without a known prior rheumatologic disorder is one of six minor criteria in the Duke diagnostic scheme (show table 2) [4] . (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 (show table 2) [4] . (See "Renal disease in infective endocarditis").


Electrocardiogram — We recommend that a baseline electrocardiogram be performed as part of the initial evaluation of all patients with suspected endocarditis even though this test rarely shows diagnostic findings. The presence or subsequent appearance of changes suggestive of ischemia or infarction on the electrocardiogram may provide useful clues to the presence of emboli to the coronary circulation. In addition, the initial presence or new appearance of heart block or conduction delay may provide an important clue to extension of infection to the valve annulus and adjacent septum. (See "Complications and outcome of infective endocarditis").

Chest radiograph — Chest radiographs occasionally reveal important diagnostic clues. As an example, patients with tricuspid valve endocarditis often present with radiographic evidence of septic pulmonary emboli. In such cases, there may be a few or multiple focal lung infiltrates, which may reveal central cavitation. Rarely, chest radiographs show calcification in a cardiac valve, which may raise suspicions of endocarditis in a febrile patient.

Echocardiography — The 2006 American College of Cardiology/American Heart Association (ACC/AHA) guidelines on the management of patients with valvular heart disease included recommendations for the use of transthoracic and transesophageal echocardiography in patients with suspected or proven native or prosthetic valve endocarditis (show table 3) [10] . These recommendations are generally consistent with the 2005 AHA guidelines on infective endocarditis, which were endorsed by the Infectious Diseases Society of America [11] .

The information that can be obtained by echocardiography includes:

• Evaluation of patients in settings in which endocarditis is suspected (such as persistent bacteremia without a known source or high clinical suspicion with negative cultures)
• Detection and characterization of vegetations on valves and in other sites (as in patients with congenital heart disease)
• Detection of valvular dysfunction and assessment of hemodynamic severity
• Detection of associated abnormalities such as shunts or abscesses
• Re-evaluation of patients in complex settings (such as those with virulent organisms, severe hemodynamic effects, persistent or recurrent fever or bacteremia, or clinical deterioration)


An echocardiogram should be performed in all patients with a moderate or high suspicion of endocarditis (show table 1 and show table 2). In comparison, among patients with a low clinical probability of endocarditis, the diagnostic yield of both transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) is very low, and neither should be performed [12] . (See "Infective endocarditis: Historical and Duke criteria", section on Case definitions).

Role of TTE — Transthoracic echocardiography may provide confirmation of the diagnosis of endocarditis. Detection of a vegetation is a positive test show echocardiogram 1 show echocardiogram 2 show echocardiogram 3. (See "Role of echocardiography in infective endocarditis").

However, TTE has relatively low sensitivity for vegetation in IE (29 to 63 percent in different series) [13-15] . Thus, the absence of vegetation does not preclude the diagnosis and, as described in the next section, TEE is usually warranted. However, 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 [16] .

In contrast to the variable sensitivity, TTE has a specificity (the likelihood of a negative result in a patient without disease) approaching 100 percent, indicating very few false positive studies [13] .

Role of TEE — Transesophageal echocardiography has a higher spatial resolution than TTE and is much more sensitive for the detection of endocarditis show echocardiogram 4 show echocardiogram 5. TEE is especially useful for the detection of vegetations, diagnosis of prosthetic valve endocarditis, detection of a valve abscess, and assessment of embolic risk.

• Detection of vegetations — Two series with a total of 162 episodes of suspected IE demonstrated the greater sensitivity of TEE compared to TTE (100 versus 63 percent and 94 versus 44 percent, respectively) [13,14] . In a third report of 103 patients with S. aureus bacteremia, TTE revealed valvular abnormalities in 33 patients and vegetations in seven; in contrast, TEE identified vegetations in 22 patients and abscesses in two [15] .
• Diagnosis of IE on prosthetic valves — TEE is especially important for prosthetic valves in the mitral or aortic position, because acoustic shadowing frequently makes the transthoracic approach suboptimal. TEE has superior spatial resolution, and bacteremic patients with these prostheses are at increased risk for endocarditis show echocardiogram 6. Few data are available regarding the superiority of TEE for suspected endocarditis of tricuspid and pulmonic prostheses, but presumably it would be similar. (See "Role of echocardiography in infective endocarditis", section on Prosthetic valve endocarditis).

The greater value of TEE in patients with prosthetic heart valves was demonstrated in a prospective study that compared TTE and TEE in 114 episodes of IE suspected on clinical grounds (80 on native valves and 34 on prosthetic valves) [17] . The results of the two tests were concordant in 55 percent of cases; TEE led to a reclassification in 11 percent of patients with native valves and 34 percent with prosthetic valves.

The negative predictive value of TEE is nearly 100 percent for patients with native valves, but IE can be missed in patients with prosthetic valves. In the latter patients, clinical assessment is especially important.

• Detection of a valve abscess — TEE is much more sensitive than TTE for the detection of valve abscess. This was illustrated in a series of 44 patients with IE complicated by abscess formation (sensitivity 87 versus 28 percent with TTE) [18] .

While TEE is generally much more sensitive than TTE for detection of abscess, even TEE may miss a significant number of abscesses in some populations. This was illustrated in a report of 44 patients with endocarditis by Duke criteria who had valvular abscesses identified intraoperatively. Endocarditis was detected by TEE in only 48 percent of cases; prosthetic heart valves were present in 32 percent of patients. Indications for surgery included severe valvular dysfunction (with or without heart failure), embolization during antibiotic therapy, or concommitant pacemaker infection [19] . Abscesses not visualized by TEE were located at the posterior mitral annulus in 61 percent of cases; the majority of these were associated with a large calcification, which may have interfered with detection.

• Assessment of embolic risk — In general, larger vegetation size is associated with increased risk of embolization. This issue is discussed separately. (See "Role of echocardiography in infective endocarditis" section on Echocardiographic estimation of outcome).

Recommended use — We generally perform a TTE as the first diagnostic test in most patients with suspected IE. However, it is reasonable to begin with TEE in selected settings:

• Limited transthoracic windows (eg, due to obesity, chest wall deformity, or mechanical ventilation)
• Prosthetic valves, especially prosthetic aortic or mitral valves in which shadowing may make visualization difficult by TTE
• A prior valvular abnormality (including previous endocarditis)
• S. aureus bacteremia [15]
• Bacteremia due to an organism known to be a common cause of IE such as viridans streptococci


For patients with a normal TTE (both morphology and function), the likelihood of IE is very low [16] . A subsequent TEE is not necessary unless one or more of the following is present:

• A high clinical suspicion of IE (persistently positive blood cultures and/or multiple minor criteria for endocarditis) (show table 1 and show table 2)
• A technically limited TTE study

Some patients with abnormal on findings on TTE may require further evaluation by TEE. These include patients with significant valvular regurgitation in whom surgery is contemplated and patients who have one or more of the following risk factors for paravalvular abscess:

• Conduction delay by ECG that is not known to be old
• Persistent fever despite appropriate antimicrobial therapy
• Aortic valve endocarditis


(See "Complications and outcome of infective endocarditis", section on Paravalvular abscesses).

Controversies — There are a number of unresolved controversies related to the use of echocardiography in the diagnosis and management of patients with IE:

• Some experts recommend that all patients with suspected IE, especially those with staphylococcal bacteremia should have a TEE as the initial study [20] . However, we regard the steps outlined above as sufficient to exclude IE without TEE in many cases, and do not require that all patients undergo this invasive procedure.
• Some experts believe that patients with staphylococcal bacteremia associated with a condition (such as vertebral osteomyelitis) that will require a protracted course of antimicrobial therapy do not require echocardiography to rule out associated IE, especially if there are no hemodynamic signs or symptoms of IE. However, we recommend that echocardiography be performed in such patients, since a positive result will influence the type and intensity of follow-up examinations. (See "Complications and outcome of infective endocarditis").
• On theoretical grounds, some physicians delay echocardiography for several days after the onset of bacteremia, because both TTE and TEE can be falsely negative if vegetations are small (and, occasionally, if previously present vegetations have embolized). In addition, even TEE can miss a paravalvular abscess, especially if the study is done in the first few days of illness [21] .


The 2005 AHA guidelines on the diagnosis and management of IE, which were endorsed by the Infectious Diseases Society of America, recommend that echocardiography be performed as soon as possible after the diagnosis of IE is suspected [11] .

Histologic examination — Histologic demonstration of microorganisms, vegetations, or active endocarditis in cardiac valve tissue obtained at surgery is included in the Duke criteria and is considered to be a criterion of confirmed infective endocarditis (show 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) [22] .

IE was characterized by microorganisms, vegetations, and significant neutrophil-rich inflammatory infiltrates with extensive neovascularization. In contrast, tissue adjoining valves from noninfectious complications showed 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.

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 IE (eg, stroke or other manifestations of emboli). Culture-negative IE should also be considered in patients with a vegetation on echocardiogram with no clear microbiologic diagnosis. Issues related to culture-negative IE are discussed separately. (See "Culture-negative endocarditis").

TREATMENT — Standard antimicrobial therapy for infective endocarditis is generally administered to patients characterized as definite or probable by the Duke criteria. Patients in whom the diagnosis is "rejected" by these criteria are not usually treated with prolonged antimicrobial therapy. (See "Antimicrobial therapy of native valve endocarditis").

SUMMARY

• The diagnosis of infective endocarditis (IE) is usually based upon a constellation of history, clinical findings, laboratory studies (particularly blood cultures), and echocardiography. (See "Introduction" above).
• Physical examination findings supporting a diagnosis of IE include new regurgitant murmurs or heart failure, evidence of embolic events (eg, focal neurologic impairment, glomerulonephritis, renal and splenic infarcts, and septic pulmonary infarcts), and peripheral cutaneous or mucocutaneous lesions (eg, petechiae, conjunctival or splinter hemorrhages, Janeway lesions, Osler's nodes, and Roth spots). (See "Physical examination" above).

As part of the initial evaluation of all patients with suspected endocarditis the following studies should be performed:

- At least three blood cultures from separate sites over a time period ranging from a few hours to one to two days depending upon the severity of illness and urgency of the need for treatment. (See "Blood cultures" above).

- An electrocardiogram to evaluate for the presence of changes suggestive of ischemia or infarction or the presence or new appearance of heart block or conduction delay. (See "Electrocardiogram" above).

- Echocardiography. (See "Echocardiography" above).


Cultures remain negative in 2 to 5 percent of patients with endocarditis, which is referred to as culture-negative endocarditis. (See "Culture-negative endocarditis").