Infective Endocarditis

Marwa A. Sabe

Brian P. Griffin

I. INTRODUCTION

A. Infective endocarditis (IE) is an infection of the cardiac endothelium, macroscopically seen as vegetations. Despite modern medical and surgical therapy, IE is a serious and life-threatening condition. Mortality rates are 20% to 30% for both native and prosthetic valve endocarditis (PVE) and may be as high as 70% in certain high-risk patients. The clinical diagnosis is based on multiple elements, and IE is best man aged via multidisciplinary collaboration among cardiologists, cardiothoracic surgeons, and infectious disease specialists.

B. The incidence of IE has remained constant over the last 30 years, accounting for 1 case per 1,000 hospital admissions. An estimated 10,000 to 15,000 new cases of IE are diagnosed each year in the United States, and the incidence has increased in the elderly and in illicit injection drug users. There has also been an increase in the number of acute cases, prosthetic valve infections, and cases due to gram-negative, rickettsial, chlamydial, fungal, and fastidious organisms.

C. Risk factors associated with infection include underlying cardiac structural abnor malities, immunosuppressed status, underlying conditions that predispose patients to pacemaker-related infections, prolonged surgery, reoperation, catheter-related bacteremia, and sternal wound infection.

II. CLINICAL PRESENTATION

A. Signs and symptoms.

The clinical manifestations of IE are highly variable, rang ing from subtle symptoms to fulminant congestive heart failure (CHF) with severe valvular regurgitation. Acute IE presents as marked toxicity and progresses to val vular destruction and metastatic infection over several days to weeks. Subacute IE evolves over several weeks to months with mild or modest toxicity and rarely causes metastatic infection. The rate of progression depends upon the virulence of the caus ative organism, the age and underlying health of the patient, and the nature and extent of the underlying valvular disease.

1. The hallmarks of IE are fever and a new murmur (> 85%); however, fever may be absent in patients who are elderly, uremic, or immunosuppressed. Murmurs may be absent with right-sided or mural infection or intracardiac device infections.

2. The patient often has nonspecific symptoms of fatigue, weight loss, malaise, chills, night sweats, and/or musculoskeletal aches.

B. Physical findings.

A new murmur remains an important finding.

1. CHF occurs in up to 55% of cases and tends to be more common in those with aortic valve disease (75%) than in those with mitral (50%) or tricuspid (20%) valve involvement.

2. Neurologic findings may include clinically apparent cerebral emboli (20%), encephalopathy (10%), mycotic aneurysm leak (< 5%), meningitis, or brain abscess (< 5%).

3. Additional physical findings reflecting embolic or immune complex phenomena include mucosal petechiae (20% to 40%), splinter hemorrhages (subungual dark linear streaks: 10% to 30%), Osler’s nodes (painful, tender erythematous nodules on the pads of fingers or toes: 10% to 25%), Janeway lesions (erythematous, macular, nontender lesions on the fingers, palms, or soles: < 5%), clubbing (10% to 20%), arterial embolism (peripherally or centrally), splenomegaly (30% to 50%), and Roth’s spots (retinal hemorrhages: < 5%). These classic physical findings are neither sensitive nor specific for the diagnosis of IE, and their frequency is continuing to diminish due to a decrease in Streptococcus viridans IE and an increase in Staphylococcus aureus IE.

4. A formal fundoscopic examination should be routine in all patients with suspected or documented IE. It may reveal chorioretinitis or endophthalmitis.

5. Systemic embolization occurs in 25% to 50% of cases of IE and may mimic acute coronary syndrome (coronary artery emboli), peritonitis (embolization to the spleen, kidney, or bowel), acute stroke (cerebral emboli), and pulmonary embolism (from right-sided IE) or cause a cold extremity with reduced or absent pulse. Septic emboli also cause Janeway lesions.

III. ETIOLOGY

Table 19.1 presents the various etiologic factors.

A. Seventy percent to 75% of patients with IE have preexisting cardiac abnormali ties. Mitral valve prolapse with regurgitation is the leading condition underlying IE in adults. Rheumatic heart disease as a substrate for IE is decreasing, with con genital heart disease underlying 10% to 20% of IE cases.

B. The source of infection can only be identified occasionally (e.g., dental procedures, an infected vascular catheter, or an infected skin lesion). In many patients, there is no history of an antecedent localized infection.

C. Native valve endocarditis

1. The most common microorganisms that cause native valve IE in adults are streptococcal and staphylococcal organisms (80%). Other important causes
include Streptococcus bovis, Enterococcus, and the HACEK (Haemophilus, Actinobacillus, Cardiobacterium, Eikenella, and Kingella) group organisms. The HACEK group, accounting for roughly 3% of cases, includes fastidious gramnegative organisms that are normal flora in the upper respiratory tract. S. bovis IE is often associated with colonic polyps and colon cancer; as such, a colonoscopy is recommended for these patients.

TABLE 19.1 Frequency of Organisms Causing Infective Endocarditis

Organism	NVE (%)	IDU (%)	Early PVE (%)	Late PVE (%)
Streptococci	60	15-25	5	35
Viridans Streptococcus	30-40	5-10	<5	25
S. bovis	10	<5	<5	<5
Enterococci	10	10	<5	<5
Staphylococci	25	50	50	30
Coagulase positive	23	50	20	10
Coagulase negative	<5	<5	30	20
Gram-negative (aerobes)	<5	5	20	10
Fungi	<5	<5	10	5
Culture negative	5-10	<5	<5	<5
IDU, intravenous drug use; NVE, native valve endocarditis; PVE, prosthetic valve endocarditis. Adapted from Fuster V, Alexander RW, O’Rourke RA, eds. Hurst’s the Heart. 11th ed. New York, NY: McGraw Hill; 2004.

2. Right-sided IE in injection drug users is usually due to S. aureus (60%), with a predilection for normal as well as abnormal cardiac valves. Despite the virulence of this organism, the disease tends to be less severe (mortality rates of 2% to 6%) than with left-sided IE. The valve most commonly affected in injection drug users is the tricuspid valve (60% to 70% of cases), followed by the mitral (30% to 40%) and the aortic valves (5% to 10%). More than one valve is involved in 20% of these patients. Septic pulmonary emboli occur in up to 75% of injection drug users with tricuspid IE.

3. IE from Pseudomonas aeruginosa is both destructive and poorly responsive to antibiotic therapy and thus often necessitates surgical intervention.

4. Enterococcal IE is increasing in prevalence. The diagnosis must be considered in patients who have undergone recent genitourinary or obstetric procedures; these patients may not have underlying heart disease.

5. Other members of the Enterobacteriaceae (Escherichia coli, Salmonella, Klebsiella, Enterobacter, Proteus, Serratia, Citrobacter, Shigella, and Yersinia) are occasionally implicated in IE.

6. Streptococcus pneumoniae accounts for 1% to 3% of native valve IE, and it may present as part of the “Osler triad,” which also includes pneumococcal pneumonia and meningitis. Alcoholics are typically affected, and the mortality rate is high (30% to 50%).

7. The most common congenital heart anomalies predisposing to IE are bicuspid aortic valve, patent ductus arteriosus, ventricular septal defect, coarctation of the aorta, and tetralogy of Fallot. There is no evidence that secundum atrial septal defects increase the risk of IE.

8. Staphylococcus lugdunensis is a rare but destructive cause of IE. This organism is a coagulase-negative Staphyloccus; however, it differs from other coagulase-negative staphylococci in its aggressive nature and predilection for native valves. S. lugdunensis IE portends a high complication and mortality rate without surgical intervention.

D. PVE accounts for about 10% to 20% of all cases of IE. The greatest risk of infection is in the first 6 months after valve implantation and appears to be similar in mechanical and bioprosthetic valves. Recent studies have suggested that infection occurs with similar frequencies at the mitral and aortic positions.

1. PVE occurring within 2 months of surgery (early PVE) is commonly associated with intraoperative contamination and nosocomial infection and this usually implicates coagulase-negative staphylococci (30% of cases). The second most common pathogen in early PVE is S. aureus (20% of cases).

2. The microbiology of PVE with an onset of more than 2 months after surgery (late PVE) reflects the pathogens of native valve IE and is most commonly caused by streptococcal species, S. aureus, and Enterococcus. Coagulase-negative staphylococci cause < 20% of infections in this period. Fungi account for 10% to 15% of late PVE cases and are associated with a higher mortality rate. Among 270 cases of fungal endocarditis reported from 1965 to 1995, 135 (50%) occurred on prosthetic valves. These patients generally lacked an identifiable portal of entry for fungemia. Establishing the diagnosis of fungal PVE can be difficult due to the low yield from blood cultures. Despite aggressive antifungal therapy, these patients remain at risk for the development of PVE months or years later. Corynebacterium species and other coryneform bacteria, often called
diphtheroids, are also an important cause of PVE during the first year after surgery (5%). Although they are often blood culture contaminants, diphtheroids in multiple cultures should not be ignored.

E. Pacemaker/defibrillator endocarditis is increasing in frequency in clinical practice with the burgeoning number of devices being implanted. The incidence of endocar ditis following device therapy ranges from 0.2% to 7%. The infection may involve the generator or defibrillator pocket, the electrodes, and valvular or nonvalvular endocardium.

1. Pacemaker/defibrillator endocarditis occurring within 1 to 2 months of surgery is likely caused by direct intraoperative microbial seeding. Late infection in the pocket produces a thinning of the overlying tissue and ultimately device erosion. The infection may eventually involve the electrodes and ultimately the endocardium. Hematogenous dissemination from distant sites of infection appears to be relatively rare, with the exception of S. aureus bacteremia.

2. The majority of infections in PVE are caused by staphylococci: S. aureus and coagulase-negative staphylococci. More than 90% of early infections are caused by coagulase-negative staphylococci, whereas late infections are caused by both S. aureus (50%) and coagulase-negative staphylococci (50%). Infection by gramnegative bacilli, enterococci, or fungi is rare.

F. The incidence of culture-negative endocarditis may be as high as 10%. Blood cul ture—negative endocarditis is defined as endocarditis without positive cultures after inoculation of three blood samples. Cultures can be negative in IE when there is infection with a fastidious bacteria or fungus, the microbiological techniques are inadequate, or there had been administration of antibiotic therapy prior to obtaining blood cultures. The latter reason is the most common cause of culture-negative IE, and the most common causative agents are Streptococcus or fastidious organisms such as fungi, HACEK organisms, anaerobes, Legionella, Chlamydia psittaci, Coxiella, Brucella, Bartonella, Tropheryma whipplei, and nutritionally deficient streptococci. Bartonella henselae infection is a rare cause of subacute IE, which is associated with exposure to cats. Coxiella burnetii causes Q fever and often infects previously dam aged valves or prosthetic valves. T. whipplei is the cause of Whipple’s disease, and this organism can be identified using periodic acid-Schiff staining of macrophages or polymerase chain reaction (PCR). Nonbacterial endocarditis (Libman-Sachs, marantic, and antiphospholipid syndrome) should also be considered in cases of culture-negative IE.

G. Fungal endocarditis (Candida and Aspergillus) usually occurs in association with prosthetic valves, indwelling intravascular hardware, immunosuppression, or injection drug use. The most common cause is Candida species, but other causes include Histoplasma and Aspergillus. Fungal IE usually presents with large veg etations that extend into the perivalvular apparatus and embolize into large vessels and therefore requires surgical intervention.

IV. PATHOPHYSIOLOGY.

The first step in the pathogenesis of vegetation is the formation of a nonbacterial thrombotic endocarditis (NBTE), which usually results from endothelial injury followed by focal adherence of platelets and fibrin. Microorganisms circulating in the bloodstream in turn infect this sterile platelet—fibrin nidus.

A. Vegetations classically occur along the line of closure of the valve leaflet. The endothelium may be injured by regurgitant jets, leading to vegetation formation on the atrial surface of incompetent atrioventricular valves or the ventricular sur face of incompetent semilunar valves. The foreign body, such as an intracardiac device, is not endothelialized initially and acts as a formation site for platelet—fibrin thrombi.

B. Bacteremia is the event that converts NBTE to IE when host defenses fail. The for eign material also impairs host defenses, rendering them more difficult to treat.

C. Vegetations often further impair valvular coaptation or cause perforation or chordal rupture, leading to worsening of regurgitation and CHF. Furthermore, the vegeta tions may dislodge, causing peripheral septic—nonseptic embolization.

D. The infection may extend to the surrounding structures, such as the valve ring, the cardiac conduction system, the adjacent myocardium, or the mitral—aortic intravalvular fibrosa. Consequently, conduction defects, abscesses, diverticula, aneurysms, or fistula may develop. Infections involving prosthetic valves commonly invade paravalvular tissue, resulting in abscess formation or valve dehiscence.

V. LABORATORY EXAMINATION

A. Blood tests

1. Laboratory findings often reflect nonspecific acute inflammatory response, mani fest as a modest leukocytosis, a normochromic normocytic anemia, and a slightly increased or decreased platelet count. Other laboratory abnormalities may include an elevated erythrocyte sedimentation rate, C-reactive protein, rheumatoid factor, and/or a hypergammaglobulinemia. IE may also cause false-positive Venereal Disease Research Laboratory (VDRL) test and Lyme serologic test.

2. Decreased complement and an elevated blood urea nitrogen or creatinine may implicate renal dysfunction from an immune complex glomerulonephritis or drug toxicity.

3. Blood cultures are critical in the diagnosis and management of IE. However, if a patient is acutely ill, therapy should not be delayed for more than 2 to 3 hours, as a fulminant infection may be rapidly fatal. In recent reports, cultures were negative in 2% to 7% of cases with established IE, despite the best modern methods.

a. If the clinical condition allows, three sets of cultures should be drawn at three different venipuncture sites before empiric antimicrobial therapy is started. Each set should include two flasks, one containing an aerobic medium and the other an anaerobic medium, into which at least 20 cm³ (per tube) of blood should be placed. The HACEK group bacteria are cultured routinely. Fungal cultures should be included when fungal infection is suspected, such as in immunocompromised hosts.

b. Intravascular infection leads to constant bacteremia originating from vegeta tions. Therefore, it is unnecessary to await the arrival of a fever spike or chills to obtain blood cultures.

c. The laboratory should be alerted if a culture-negative IE or a fastidious infectious agent is suspected, as it may be necessary to enhance the cul ture medium or prolong the incubation period. For example, the HACEK group (see Section III.C.1) needs prolonged incubation of up to 21 days. The most common culture-negative IE organisms are C. burnetii, Barton ella sp., T. whipplei, HACEK group bacteria, Brucella sp., Legionella sp., Mycoplasma sp., Mycobacterium, and fungi. Serology for Brucella, Legionella, Coxiella, or Psittacosis may be revealing. Fastidious organisms can also be identified using PCR in valvular specimens. This technique does not require a culture medium; however, it does require excised valvular tissue. PCR has been shown to have a sensitivity of 41% and a specificity of 100% in recent studies, and thus it may provide important information about the causes of IE that could not be identified by culture. This information may help guide future empiric treatment plans in certain patient populations.

d. Special attention should be paid to cultures positive for coagulase-negative staphylocci. S. lugdunensis is a coagulase-negative Staphylococcus that rarely causes IE. Unlike other coagulase-negative staphylococci it often affects native valves, is destructive, frequently causes abscesses, and is associated with high mortality without surgical intervention. Thus, in the setting of
high suspicion for IE, cultures positive for coagulase-negative staphyloccci should not be disregarded as a contaminant and should be further speciated.

B. Histologic evaluation.

Histopathology of resected valvular tissues remains the gold standard for the diagnosis of IE. It may demonstrate valvular inflammation, vegetations, and/or specific organisms. Detection of an etiologic agent in the veg etation using special stains or immunohistology can guide the choice of antimicro bial treatment. This is particularly useful in culture-negative IE, such as Q fever, Bartonella spp., or T. whipplei (Whipple’s disease bacillus). Good communication among cardiologists, surgeons, pathologists, and microbiologists helps ensure accurate diagnosis.

C. Urinalysis.

Microhematuria with or without proteinuria may be seen.

D. Electrocardiography.

All patients with suspected IE should undergo baseline and follow-up electrocardiogram (ECG).

1. ECG may reveal conduction disturbances reflecting intramyocardial extension of infection, ranging from a prolonged PR interval to complete heart block (especially with PVE). A new atrioventricular block carries a 77% positive predictive value for abscess formation with 42% sensitivity.

2. Myocardial infarction due to embolization of vegetations occurs rarely.

E. Chest x-ray may reveal CHF or pleural effusions. Right-sided IE may cause non specific infiltrates due to multiple septic pulmonary emboli.

VI. DIAGNOSTIC IMAGING TECHNIQUES

A. Echocardiography has a key role in both diagnosis and management of IE.

The primary objective is to identify, localize, and characterize valvular vegetations and their effects on cardiac function. Vegetations may occur at intracardiac locations other than valves, such as the site of impact of a high-velocity jet or shunt. A limitation of echocardiography is that vegetations cannot always be distinguished from other noninfectious masses.

1. All patients in whom IE is suspected should undergo baseline transthoracic echocardiography (TTE) to define underlying cardiac abnormalities, to determine the size and location of vegetations, and to explore the possibility of complications (e.g., aortic annular ring abscess). TTE has a low sensitivity for vegetations in IE (29% to 63%) but has close to 100% specificity. However, the finding of morphologically and functionally normal valves on TTE decreases the likelihood of IE. In one series, 96% of patients with normal valves on TTE also had a negative transesophageal echocardiography (TEE).

2. TEE has increased the diagnostic accuracy of IE. If IE is strongly suspected and the TTE is negative, then TEE should be performed because it is more sensitive in detecting vegetations, especially if TTE imaging is difficult. TEE is particularly useful for assessing posterior structures, abscesses, fistulae, perivalvular leaks, small vegetations, right-sided heart structures, masses on intracardiac devices, leaflet perforations, and prosthetic valves. The ability to detect paravalvular abscesses, fistulae, and paraprosthetic leaks has a major impact on management strategy. Intraoperative TEE can be used to evaluate the success of surgical interventions and the need for potential modification of reparative cardiac surgical procedures. A postoperative TTE should also be done as a baseline measure of cardiac anatomy/function for long-term follow-up. Although in most cases a TTE should be the first diagnostic test of choice, in certain circumstances TEE may be the optimal initial test to rule out IE. These include cases that involve S. aureus bacteremia, prosthetic valves, prior IE, limited echo windows, and bacteremia due to an organism that is known to commonly cause IE.

a. A negative result on TEE indicates a low likelihood of IE (provided adequate images are available). However, it does not completely rule out the diagnosis. The negative predictive value is > 90%, but false negatives may occur early
in endocarditis or if vegetations are small. Repeat TEE should be considered if clinical suspicion is high. Of note, a negative TEE should never override strong clinical evidence of endocarditis in the diagnosis of PVE.

b. Myocardial abscesses are more reliably detected with TEE (87% sensitive) than with TTE (28% sensitive). Detection of a perivalvular abscess is essen tial, as an abscess is a serious complication and a strong indication for surgi cal intervention.

c. In the setting of PVE, TEE is superior (82% sensitive) to TTE (36% sensi tive) in the detection of vegetations due to acoustic shadowing of prosthetic valves, especially in the mitral and aortic positions. TEE should be per formed if PVE or pacemaker endocarditis is suspected but is not evident on TTE.

3. Fungal endocarditis tends to cause larger vegetations than bacterial infections, whereas in Q fever vegetations are often absent. Care should be taken to differentiate bacterial vegetations from myxomas, papillary fibroelastomas, rheumatoid nodules, inflammation involving degenerative valvular lesions, Lambl’s excrescences, and nonbacterial endocarditis. It is essential to interpret images in conjunction with clinical data.

4. One meta-analysis showed that the risk of embolization in patients with large vegetations (> 10 mm) was nearly three times higher than in patients with no detectable vegetations or small vegetations. Prolapsing vegetations and involvement of extravalvular structures increase the overall risk of heart failure, embolization, and need for valve replacement. Vegetations that increase in size, despite appropriate therapy, are also more likely to be associated with adverse events requiring surgery.

5. TEE is indicated in patients with suspected pacemaker or defibrillator endocarditis. The sensitivity of TTE for detecting valvular or lead vegetations is 30%, compared with 90% with TEE.

B. Cardiac catheterization.

Left heart catheterization with selective coronary angi ography is indicated prior to surgical intervention if there is a suspicion of obstruc tive coronary disease. The abnormal rocking motion of a dehisced prosthetic valve may be noted on fluoroscopy. Care should be taken to avoid unnecessary coronary angiography or cardiac catheterization in aortic valve endocarditis because of the risk of embolization of vegetations.

C. Central nervous system (CNS) imaging.

Computerized tomography (CT), mag netic resonance imaging (MRI), or cerebral angiography should be considered in any patient who has sustained a CNS complication, such as an embolic infarct, intracra nial bleed, or mycotic aneurysm, or in the patient with persistent headaches.

D. Body imaging.

CT or MRI may be useful in the detection of metastatic infection. The value of CT may increase in the future as spatial resolution improves. MRI does not currently have a significant role in assessing cardiac manifestations of IE, owing to intrinsic problems related to temporal resolution.

VII. DUKE CRITERIA.

Given the complexity of IE, the diagnosis requires a high index of suspicion. The Duke schema is currently the most sensitive and specific diagnostic set of criteria available. It is particularly useful in diagnosing endocarditis in patients with S. aureus bacteremia, those with right-sided endocarditis, and those with negative blood culture results. However, these criteria have not been validated in PVE.

A. The criteria are divided into definite (pathologic or clinical), possible, and rejected diagnostic groups.

B. For a definite pathologic diagnosis, either (A or B) of the pathologic findings listed in Table 19.2A is sufficient.

C. For a definite clinical diagnosis (Table 19.2B), two major criteria, or one major and three minor, or five minor criteria are needed.

TABLE 19.2A Duke Criteria: Definite Pathologic Diagnosis

A. Microorganisms, as demonstrated by culture or histology in vegetation
	Vegetation that has embolized
	Intracardiac abscess
B. Pathologic lesions
	Vegetation or intracardiac abscess present, confirmed by histology showing active endocarditis
Adapted from Baddour LM, Wilson WR, Bayer AS, et al. Infective endocarditis: diagnosis, antimicrobial therapy, and management of complications. Circulation. 2005;111:394-434.