Epidemiology

The incidence of infective endocarditis in adults has been difficult to determine because the methods of study and criteria for diagnosis vary among series. Accurate figures on the incidence of infective endocarditis in children are difficult to obtain. The most common method of reporting the incidence in pediatric series expresses the number of cases of infective endocarditis as the numerator and the total number of hospital admissions during the analyzed period as the denominator. Zakrzewski and Keith reported an incidence of endocarditis of 1 in 4500 pediatric admissions at the Hospital for Sick Children in Toronto from 1952 to 1962, whereas Van Hare and colleagues at Case Western Reserve in Cleveland, Ohio, found an incidence of 1 in 1280 in the period from 1972 to 1982. In a large series from Boston Children’s Hospital spanning the period between 1933 and 1972, the incidence before 1963 was 1 in 4500 pediatric admissions, whereas that for 1963 to 1972 was 1 in 1800 admissions. A study from a children’s hospital in Australia reported an incidence of 1 in 4500 hospital admissions between 1971 and 1983. One Japanese center reported an annual incidence of 0.9 cases per 1000 children seen at the cardiology clinic. In a report from Canada the cumulative incidence of IE occurring in children with congenital heart disease was 6.1 per 1000 through 18 years of age. Although differences in referral patterns at these centers may have introduced bias into these figures, the incidence of infective endocarditis in children appears to be rising. This rise may be explained by the increased survival rate of children with all forms of cardiovascular disease and an increase in the percentage of cases that occur after cardiac surgery (especially implantation of foreign material ) and are related to intravascular catheters. Early surgical correction of many types of congenital heart diseases, along with effective appropriate perioperative antibiotic prophylaxis regimens, ultimately may lower the incidence of postoperative infective endocarditis. However, the use of invasive therapeutic modalities, especially intravenous catheters and pacemakers, has led to an increased incidence of health care–associated endocarditis. In general children with predisposing cardiac conditions who develop infective endocarditis while hospitalized have longer hospitalizations and higher mortality than patients with community-associated endocarditis.

The average age of children with infective endocarditis is increasing, a phenomenon that may reflect the longer life expectancy created by improved therapy for children at risk. From 1930 to 1950, the mean age for children with infective endocarditis was close to 5 years. Between 1960 and the present, it increased to 8.5 and then to 13 years. The number of reports of infective endocarditis in children younger than 2 years of age had been small but has increased significantly since the late 1980s. The clinical course of infective endocarditis in these young children often is atypical, and some cases are diagnosed at autopsy. Before the 1950s, this disease was a rare event in neonates, with only eight autopsy cases reported. Several reports suggest a rapidly increasing rate associated with the development of intensive supportive care in neonates. Symchych and colleagues found a 3% incidence of bacterial endocarditis among all neonatal autopsies. Endocarditis in neonates frequently occurs on the tricuspid valve when associated with an indwelling central venous catheter. Congenital heart defects also predispose neonates to the development of infectious endocarditis.

Any form of structural cardiac disease may predispose to infective endocarditis, especially disorders associated with turbulence of blood flow. In autopsy and clinical series, children with ventricular septal defect, tetralogy of Fallot, left-sided valvular disease, and systemic-pulmonary arterial communication were at highest risk, whereas those with pulmonary stenosis, coarctation of the aorta, and secundum atrial septal defect were at low risk. Hypertrophic obstructive cardiomyopathy rarely is associated with infective endocarditis. Isolated pulmonic or tricuspid valve endocarditis can occur in “otherwise normal” children and adolescents with sepsis or focal bacterial infection, but usually it is associated with congenital heart disease, intravenous catheters, or intravenous drug abuse. A bicuspid aortic valve is recognized as an important risk factor for the development of infective endocarditis, especially in elderly men. The underlying heart diseases in 266 pediatric cases of infective endocarditis are listed in Table 26.1 . In a large Canadian study, independent risk factors for infective endocarditis among children with congenital heart disease were cyanotic congenital heart disease, endocardial cushion defects, left-sided heart lesions, age less than 3 years, and cardiac surgery within 6 months.

A cooperative study on the natural history of aortic stenosis, pulmonary stenosis, and ventricular septal defect reported data from a controlled pediatric population collected over a period of 4 to 15 years. In patients not undergoing surgical correction, the risk of acquiring endocarditis by 30 years of age in those with ventricular septal defects was 9.7% versus 1.4% for aortic stenosis and 0.9% for pulmonic stenosis. Aortic valvotomy in children with aortic stenosis actually increases the relative risk, whereas successful repair of ventricular septal defect significantly decreases long-term susceptibility to infective endocarditis. Similarly endocarditis is an extremely rare occurrence after ligation of patent ductus arteriosus has been performed. At present, palliative systemic-to-pulmonary shunting is the surgical procedure most often complicated by infective endocarditis. In a review of 115 patients with tetralogy of Fallot, Kaplan and colleagues reported an 8% incidence of infective endocarditis after placement of a Pott shunt. Transcatheter placement of prosthetic pulmonary valves (TPV) (Melody valve) has been associated with an incidence of subsequent development of infective endocarditis of 2.4% per patient-year and 0.88% per patient-year for TPV-specific endocarditis.

The increasing use of prosthetic valves and valved conduit repairs in children with complex heart disease may lead to a larger number of cases of infective endocarditis in the future. Most medical centers report an incidence of prosthetic valve endocarditis of 2% to 4% after surgery, with the aortic and mitral valves affected most frequently. Prosthetic material is also implanted for right ventricular outflow reconstruction. In such cases conduit endocarditis is more common for bovine jugular vein grafts compared to cryopreserved homografts. In a report from Canada reviewing conduits placed in almost 300 patients with a median follow-up of 3.4 years, conduit endocarditis occurred in 9.4% (23/244) of bovine jugular vein grafts implanted versus 0.7% (1/135) of the cryopreserved homografts ( P < .001).

Older studies arbitrarily divided prosthetic valve endocarditis into two categories—early and late—based on whether the infection occurred within 60 days of valve placement or later. The rationale for categorizing by time was based on apparent differences in bacteriologic, pathogenetic, and prognostic associations. So-called early cases most often were caused by coagulase-negative staphylococci (CONS), gram-negative bacilli, and fungi, whereas oral and enterococcal streptococci, along with staphylococci, predominated in late cases. These older reports suggested that early cases were acquired by contamination of an intraoperative valve or were secondary to postoperative extracardiac infections, whereas late cases were acquired by the same mechanisms as native-valve endocarditis. Nosocomial bacteremia that develops at any time after the patient has undergone valve placement is a significant risk factor for development of endocarditis. Finally the mortality rate was thought to be higher in early versus late infection.

However, more recent studies have blurred this arbitrary time distinction between early and late prosthetic valve endocarditis. The risk probably is highest in the first 6 to 12 months and decreases to its lowest point beyond 1 year after valve replacement. CONS are the dominant organisms both before and after the 60th postoperative day. Clinical and epidemiologic data also suggest that prosthetic valve infection caused by staphylococci within the first year after placement probably is acquired at the time of surgery. Identified risk factors for the development of prosthetic valve endocarditis in adults include native valve endocarditis, black race, male sex, a mechanical (vs. biologic) prosthesis, and prolonged cardiopulmonary bypass time ; no comparable information is available for children.

Mitral valve endocarditis occurs frequently on an anatomically normal valve in patients with other predisposing factors. An association between mitral valve prolapse and infective endocarditis has been recognized in adults and children. This heart lesion is detected with increasing frequency in adolescent girls and may be only one component of a developmental syndrome. In adults, 40% to 50% of cases of infective endocarditis associated with isolated insufficient mitral valves occur in patients with mitral prolapse. In some series of native valve endocarditis, mitral valve prolapse has been the most common underlying lesion. The reported incidence of infective endocarditis in patients with mitral valve prolapse has varied markedly among studies, from low rates of 14 per 100,000 per year to 5 of 58 patients monitored prospectively for 9 to 22 years. A retrospective epidemiologic analysis involving matched cases and controls yielded an odds ratio of 8.2, indicative of a substantially higher risk for development of endocarditis in patients with mitral valve prolapse than in normal controls.

The risk of developing infective endocarditis is not uniform for all patients with mitral valve prolapse. The risk is increased in patients with a preexisting systolic murmur (but not for those with an isolated click and no murmur), echocardiographically demonstrated regurgitation, and valvular redundancy. The signs and symptoms of endocarditis associated with mitral valve prolapse may be more subtle than those of other types of left-sided endocarditis. However, significant complications are relatively common occurrences and sometimes require valve replacement during the acute illness or during convalescence.

Fungal endocarditis is a rare disorder in children but should be suspected in certain clinical and epidemiologic settings. It is more likely to occur after cardiac surgery and rarely occurs on native heart valves. It occurs more commonly in neonates treated in intensive care settings than in older children. Other predisposing factors include (1) the presence of an indwelling vascular catheter, (2) prolonged use of antibiotics, (3) intrinsic (immunodeficiency diseases, malignancy, malnutrition) or extrinsic (corticosteroids, cytotoxic drugs) immunosuppression, (4) bowel surgery resulting in transient fungemia, (5) intravenous drug use, and (6) preexisting or concomitant bacterial endocarditis.

Many conditions other than structural heart disease predispose children to the development of infective endocarditis. The most important is the presence of an indwelling central venous catheter, especially in patients who are seriously ill or immunocompromised. The catheter acts as a foreign body and presumably causes microscopic damage by abrading endocardial and valve surfaces; such damage results in nonbacterial thrombotic vegetation. Infection of intracardiac pacemaker wires also can lead to endocarditis. Infection acquired during the placement procedure and infection of the pacemaker pouch are most common. Infective endocarditis, usually of the tricuspid valve, has developed in children with ventriculoatrial shunts placed for the treatment of hydrocephalus. In patients with arteriovenous fistulas created for hemodialysis, bacterial vegetations may develop in the fistula and on heart valves. Rarely, penetrating wounds or foreign bodies can initiate endocarditis. Piercings of various body parts also have been associated with endocarditis. One important group of patients with an increased risk for development of infective endocarditis is intravenous drug users. A predilection for involvement of the tricuspid valve, followed by the mitral and aortic valves, has been noted. Radiographic evidence of septic pulmonary emboli and signs of tricuspid insufficiency dominate the clinical findings. Within this group of patients, increased rates of infective endocarditis and mortality are associated with infection by human immunodeficiency virus (HIV), particularly as CD4 cell counts fall to less than 200/mm ³ .

Several recent studies have found an apparent shift in the epidemiology of pediatric infective endocarditis toward a higher proportion of children without preexisting heart disease, which accounted for 35% to 58% of all the infective endocarditis cases. In these patients, S. aureus was the most common causative organism, and delay in diagnosis was common.

Although the incidence of infective endocarditis in children may be rising, the prognosis has improved dramatically during the past several decades. Current mortality rates usually are close to 10%. Most survivors remain hemodynamically stable at long-term follow-up. However, patients who experience infective endocarditis appear to be at higher risk for developing recurrent endocarditis than are those with similar cardiac abnormalities who have not had previous endocarditis. The patient’s functional class before treatment appears to be most predictive of long-term functional status. In one study, 22% of children who survived infective endocarditis required surgery related to the infection, including vegetectomy, evacuation of a hematoma, atrioventricular valve replacement, and placement or replacement of a graft or intracardiac shunt.

Pathophysiology

Clinical observations, autopsy studies, and work with experimental animal models have demonstrated that the occurrence of several independent events is required for the development of subacute infectious endocarditis. The endocardial surface usually is disrupted by stress or injury commonly caused by the turbulence of blood. This surface damage results in the deposition of fibrin and platelets, which form nonbacterial thrombotic vegetations. If bacteria adhere to these deposits, infective endocarditis will result. The surface of the infected vegetation becomes protected by a cover of fibrin and platelets. A tremendous proliferation of organisms (as many as 10 ⁹ CFU/g) may ensue. The protective sheath isolates the organisms from the action of host neutrophils and antibiotics. The clinical manifestations and complications of infective endocarditis are related to both the hemodynamic changes caused by local infection and the occurrence of embolization and metastatic infection.

In experimental animals, the valvular surface must be damaged, usually by an intravenous catheter, to produce infective endocarditis. The first step in the pathogenesis of subacute infective endocarditis in humans is the development of hemodynamic factors that favor endocardial damage. In an autopsy study of 1024 patients with infective endocarditis, Lepeschkin showed that the location of the endocardial lesions correlated with the impact of pressure; this finding makes a strong argument for the role of mechanical stress as a critical factor in the evolution of the lesions. When associated with valvular insufficiency, infective endocarditis usually occurs on the atrial surface of the mitral valve and the ventricular surface of the aortic valve. Injection of a bacterial aerosol into the air stream passing through a Venturi tube demonstrates how high pressure drives an infected fluid into a low-pressure sink. This process establishes maximal deposition of bacteria in the low-pressure sink immediately beyond the orifice. Mitral insufficiency creates a Venturi effect when blood is driven from the high-pressure left ventricle into a low-pressure atrium; maximal deposition occurs around the mitral annulus on the atrial side. Similarly, with aortic valve insufficiency, the high-pressure source is the aorta, and the low-pressure sink is the left ventricle, which leads to deposition on the ventricular surface of the valve.

Lesions also are created more directly by a jet stream causing endocardial damage. For example, in a small, restrictive ventricular septal defect with a left-to-right shunt, a Venturi effect leads to the development of lesions on the right ventricular septal side of the defect, whereas secondary lesions created by the jet effect are located on the right ventricular wall opposite the defect. Heart defects with a surface area sufficiently large to prevent a significant pressure gradient and those in which smaller volumes minimize the gradient do not create the jet and Venturi effects. This difference helps to explain the rarity of endocarditis in patients with atrial septal defects and the increased risk of infection complicating small, but not large, ventricular septal defects.

Once endocardial damage has occurred, collagen is exposed and platelet and fibrin deposition ensues in a manner analogous to formation of the primary plug of normal hemostasis after vascular injury. The sterile platelet-fibrin thrombus that is formed subsequently is referred to as a nonbacterial thrombotic vegetation. Formation of the vegetation reflects two pathogenic mechanisms: hypercoagulability and endothelial damage. To establish experimental infective endocarditis without initial formation of the vegetation is nearly impossible. Microscopic examination demonstrates that this lesion is the one to which microorganisms attach during the early stages of experimental endocarditis. Nonbacterial thrombotic vegetations have been found in both adults and children with malignancy, chronic wasting diseases, uremia, connective tissue diseases, and congenital heart disease and after the placement of intracardiac catheters, and they have been associated with embolism and infarction in distant organs.

Once a nonbacterial thrombotic vegetation has been established, transient bacteremia or fungemia may result in colonization of the lesion. Transient bacteremias are common occurrences, especially with traumatization of a mucosal surface. The incidence of bacteremia in adults and children after various procedures is listed in Table 26.2 . The bacteremia usually is of low grade and is proportional to the amount of trauma produced by the procedure and the number of organisms inhabiting the surface. In addition, “silent” bacteremia probably occurs frequently. Many persons have circulating antibodies to their own oral flora, as well as an increase in peripheral T cells sensitized to the flora of their dental plaque. Some children with congenital heart disease may be at increased risk for having gingival colonization and subsequent development of bacteremia with organisms associated with infectious endocarditis, such as the HACEK (Haemophilus spp., Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens, Kingella kingae) microbes.

The ability of microorganisms to adhere to the platelet-fibrin thrombus is a critical factor in the development of infective endocarditis. In a canine model, S. aureus and the viridans streptococci, which frequently cause infective endocarditis, adhere more readily to normal aortic leaflets than do organisms uncommon in endocarditis. Within isolates of S. aureus, strains devoid of microencapsulation are less capable of inducing endocarditis in an experimental model than are encapsulated strains. Specific products released by these organisms, including dextran, mannan, teichoic acid, and slime, may enhance their ability to colonize the vegetation. The amount of dextran produced by various viridans streptococci in broth correlates with both their adherence and their ability to produce endocarditis in the rabbit model. Candida albicans is readily adherent and produces infective endocarditis in rabbits more easily than does C. krusei, a nonadherent yeast rarely implicated in human infective endocarditis. In addition, endocarditis-producing strains of streptococci and staphylococci are more potent stimulators of platelet aggregation than are other bacteria that do not produce infective endocarditis. This action may accelerate the formation of an infected vegetation or may increase the removal of organisms from the circulation. The importance of adherence by organisms has been studied by preincubating organisms with many classes of antibiotics. After incubation at subinhibitory concentrations, adhesion of streptococcal species to fibrin-platelet matrices and damaged canine valves is decreased. Antibiotics may prevent development of infective endocarditis by both bacterial killing and inhibition of adherence to the vegetation.

Host tissue factors undoubtedly play an important role in adherence of bacteria to the developing thrombus. Activation of the coagulation system ensues once bacteria become adherent to a nonbacterial thrombus. Some organisms that produce endocarditis may be able to initiate procoagulant activity through microbial enzymes. Activation of the intrinsic coagulation pathway is triggered by exposed connective tissue components and platelet aggregation. However, activation of the extrinsic coagulation pathway probably is the major stimulus for growth of vegetations. Elements of the extracellular matrix, including fibronectin, laminin, and collagen, have been shown to facilitate the adherence of bacteria on fibrin-platelet matrices. Fibronectin may be the host receptor for organisms within the nonbacterial thrombotic vegetation. Laminin-binding proteins have been found on the cell walls of organisms recovered from patients with endocarditis.

The platelet-organism interaction is complex and not understood completely. Streptococcus sanguinis produces two cell surface antigens that promote platelet aggregation: a class I antigen promotes adhesion of S. sanguinis to platelets, whereas coexpression of a class II antigen promotes platelet adhesion or aggregation. The induced platelet aggregation appears to be an important determinant of further development of vegetation and progression of disease in experimental endocarditis. In addition, production of streptococcal exopolysaccharide inversely correlates with platelet adhesion while inhibiting aggregation, thus indicating that surface molecules may enhance endocarditis at only certain pathogenic steps. Platelets also may be involved in host defense within the vegetation. After exposure to thrombin, platelets may release microbicidal proteins with bactericidal activity against some gram-positive cocci; resistance to these proteins may be a virulence factor for S. aureus in the development of endocarditis.

As bacterial colonization of a nonbacterial thrombotic vegetation progresses, it enlarges by further bacterial proliferation and platelet-fibrin deposition ( Fig. 26.1 ). Kissane described three histologic zones: (1) necrotic endocardium; (2) a broad zone of bacterial colonies, pyknotic nuclear debris, and fibrin; and (3) a thin coating on the surface of fibrin and leukocytes. The location of bacterial colonies below the surface and the minimal infiltration by phagocytic cells create an environment of impaired host resistance that results in extreme bacterial proliferation. The structure of the vegetation diminishes the penetration of antibiotics into the bacterial layer. In addition, the metabolic activity of bacteria within this lesion is slowed, thus rendering antibiotics less effective. The formation of vegetations and erosion of heart valves may cause valvular incompetence and thereby may result in cardiac failure.

Subacute endocarditis of the mitral valve with vegetation and rupture of the papillary muscle caused by Staphylococcus aureus.

(Courtesy of Dr. Edith P. Hawkins, Texas Children’s Hospital, Houston.)

Immunopathologic factors may have important roles in both the development and sequelae of infective endocarditis. The susceptibility of a gram-negative bacillus to complement-mediated bactericidal activity is critical to its potential to create endocarditis. Gram-positive cocci are a more frequent cause of infective endocarditis than are gram-negative bacilli. Gram-positive organisms are resistant to this bactericidal activity; phagocytosis is required for killing.

The frequent presence of hypergammaglobulinemia, splenomegaly, and monocytes in the blood of patients with infective endocarditis indicates stimulation of the humoral and cellular immune systems. Macroglobulins, cryoglobulins, and agglutinating, opsonic, and complement-fixing antibodies have been associated with infective endocarditis. Studies in animals preimmunized with heat-killed streptococci before aortic valve trauma and infection are induced suggest that circulating antibody has a protective role. However, antibody to S. aureus or Staphylococcus epidermidis does not prevent the development of endocarditis in immunized animals, perhaps because this antibody does not enhance opsonophagocytosis. The continuous antigenic challenge created by intravascular organisms leads to increased production of specific antibody (including opsonic, agglutinating, and complement-fixing antibodies), cryoglobulins, macroglobulins, and antibodies to bacterial heat shock protein, as well as to the subsequent formation of circulating immune complexes. These complexes are found with increased frequency in patients with a long duration of illness, hypocomplementemia, extravalvular manifestations, and right-sided disease. Quantitative levels of circulating immune complexes may be helpful in distinguishing endocarditic from nonendocarditic sepsis and in monitoring anti-infective therapy. Effective treatment usually leads to a prompt decrease in these levels, whereas relapses may be characterized by rising titers. The diffuse glomerulonephritis occasionally noted with infective endocarditis is caused by subepithelial deposition of immune complexes and complement. Immune complexes can be demonstrated in some diffuse purpuric lesions seen with endocarditis. Bacterial antigens have been found within these complexes.

Further evidence of stimulation of the immune system in infective endocarditis is the development of rheumatoid factor in approximately 50% of adults with disease lasting longer than 6 weeks. Titers of rheumatoid factor correlate with hypergammaglobulinemia and, as with immune complex levels, decrease with therapy and increase during relapse. The role of rheumatoid factor in the disease process is unknown, but it may be involved by blocking immunoglobulin G opsonic activity, stimulating phagocytosis, or accelerating microvascular damage. Antinuclear, antiendocardial, antisarcolemmal, and antimyolemmal antibodies also have been identified in patients with infective endocarditis; their role in pathogenesis is unclear.

The pathologic changes that occur in the heart in association with infective endocarditis are secondary to local extension of the infection. The vegetations vary from a millimeter to several centimeters; frequently they are singular, but they may be multiple. Valvular stenosis may result from large lesions. Vegetations secondary to certain organisms, especially Candida, Haemophilus, and S. aureus in acute cases, often are large and friable, with a propensity for embolization. Ulcerative lesions may occur and may lead to perforation of the valve and subsequent congestive heart failure. Other local complications include rupture of the chordae tendineae or papillary muscle (see Fig. 26.1 ), valve ring abscess with subsequent fistula formation and pericardial empyema, aneurysms of the sinus of Valsalva or ventricle, myocarditis, and myocardial infarction. Persistent fever occurring during appropriate medical therapy for infective endocarditis may reflect a persistent vegetation, especially with right-sided disease, or extension of infection into a valve ring and adjacent structures. In such cases, surgery frequently is required.

The pathologic changes in distant organs usually are secondary to embolization with subsequent infarction or metastatic infection. In many cases of infective endocarditis, the causative organism is of low pathogenicity; infections caused by septic emboli often are low grade because of the reduced propensity of these organisms to invade tissue. However, the emboli in acute S. aureus endocarditis frequently cause severe metastatic infections and overwhelming sepsis. Emboli from right-sided heart lesions lodge in the lungs and cause pulmonary infarcts and abscesses, which usually are small and multiple. Left-sided lesions may embolize to any organ but most commonly affect the brain, kidney, spleen, and skin. Cerebral emboli have been detected in 30% of cases in adults and children and have caused infarction, abscess, mycotic aneurysm, subarachnoid hemorrhage, meningitis, and acute hemiplegia of childhood. Kidney abscess is a rare occurrence, but infarcts are noted in most patients at autopsy. Amyloidosis involving primarily the kidneys is a rare complication of chronic infective endocarditis. Splenic abscess also is a rare event but can be a fatal complication if undetected. The most common manifestation of embolization to the skin is petechiae. Janeway lesions are septic emboli consisting of bacteria, neutrophils, necrosis, and subcutaneous hemorrhage. Osler nodes are areas of thrombosis and necrosis. They may be related to both immune complex deposition and septic emboli.

Clinical Manifestations

The signs and symptoms of infective endocarditis are determined by the extent of local cardiac disease, the continuous bacteremia, and the degree of involvement of distant organs as a result of embolization, metastatic infection, and circulating immune complexes. Consequently the clinical findings are highly variable and mimic those of many other diseases. Unexplained embolic phenomena in any organ should suggest the diagnosis of endocarditis, especially in children with known heart disease. Patients with acute bacterial endocarditis initially may be seen with florid sepsis; the endocarditis is diagnosed at autopsy. The indolent manifestations of subacute endocarditis may evolve for weeks or months before medical care is sought. Endocarditis frequently occurs in children with preexisting heart disease, so subtle changes in cardiac function may be difficult to detect early in the course. The frequency of the major clinical manifestations of bacterial endocarditis in infants and children is listed in Table 26.3 .

Fever is the most common symptom of infective endocarditis, but it is absent in 10% of cases. It usually is of low grade and has no specific pattern. Chills may accompany the fever, but they rarely are seen in children. Persistent fever during antimicrobial therapy is an uncommon occurrence. Prolonged (≥2 weeks) fever is associated with certain etiologic agents ( S. aureus, gram-negative bacilli, fungi), with culture-negative endocarditis, and with complications such as embolization of major vessels, intracardiac or peripheral abscess, tissue infarction, a need for cardiac surgery, and a higher mortality rate. Nonspecific symptoms such as malaise, anorexia, weight loss, and fatigue are common findings. Arthralgia occurs in 24% of patients. The arthralgia frequently is multiple and most commonly affects the large joints. Although adults initially may have synovitis, this finding is rare in children. Osteoarticular infection in association with infective endocarditis in adults occurs almost exclusively in intravenous drug users. It is seen very rarely in children except those with disseminated S. aureus infection. Gastrointestinal complaints are noted in 16% of cases and include nausea, vomiting, and abdominal pain. Chest pain occurs in approximately 10% of older children and generally is mild and nonspecific. Although chest pain usually is related to diffuse myalgias, it may be secondary to pulmonary complications or cardiac lesions, especially if the tricuspid valve is involved.

Heart murmurs occur in more than 90% of children with infective endocarditis, but most patients have underlying heart disease with existing murmurs. The appearance of a new murmur or appreciation of a significant change in a previous one occurs in only 25% of cases. Significant blood flow turbulence caused by compromised valvular function must have occurred for a murmur to be detected or to change. The frequent absence of changes in the cardiac examination early in the disease contributes to the long average delay in establishing the diagnosis, especially in children with preexisting heart disease. Congestive heart failure occurs in 30% of children with infective endocarditis and is especially common in those in whom a new murmur of valvular insufficiency develops. Endocarditis should be suspected in any child who has rheumatic or congenital heart disease and unexplained deterioration in cardiac function. Although valvular regurgitation is the most common hemodynamic complication of endocarditis, significant obstruction of a valve or shunt requiring rapid surgery rarely occurs.

Neurologic signs and symptoms are reported in approximately 20% of children with endocarditis. These signs and symptoms may dominate the clinical findings, especially in patients with endocarditis caused by S. aureus. Neurologic abnormalities also are common in children with endocarditis caused by Kingella kingae . The sudden development of cerebral lesions in an infant or child should suggest this diagnosis. The manifestations are those that commonly accompany a cerebral infarct or abscess—namely, acute hemiplegia of childhood, seizures, ataxia, aphasia, sensory loss, focal neurologic deficits, and alterations in mental status. They may be the initial features of endocarditis or may occur years after the infection has been eradicated. Mycotic aneurysms of the cerebral vessels occur rarely in cases of pediatric endocarditis. They usually are single, small, and peripheral but may lead to subarachnoid hemorrhage. Whereas computed tomographic scanning of the brain is useful for delineating central nervous system involvement in patients with infective endocarditis, magnetic resonance imaging may be more sensitive for detecting small infarctions and changes secondary to cerebral edema. Other neurologic manifestations associated with endocarditis include cranial nerve palsies, neuropathy, visual changes, choreoathetosis, seizures, and toxic encephalopathy.

Splenomegaly, a common manifestation of endocarditis in children, occurs in 55% of cases. It is found frequently in patients with long-standing disease and other evidence of immune system activation. The spleen generally is nontender and may be associated with mild hepatomegaly. Splenic infarction and abscess are rare events but should be suspected in patients with left upper quadrant abdominal pain that radiates to the left shoulder, a pleural friction rub, or left pleural effusion.

Skin manifestations occur less commonly in children than in adults. Clubbing is found in 10% to 20% of children with endocarditis but frequently is related to underlying heart disease. Petechiae are noted in approximately one-third of patients, especially those with long-standing disease. These lesions are found most commonly on the extremities, oral mucosa, and conjunctivae. Splinter hemorrhages are linear red or brown streaks seen in the nail beds. They are present in only 5% of children with endocarditis and are associated with other conditions. Three other types of lesions are more specific for infective endocarditis but occur in only 5% to 7% of patients: Osler nodes, which are small (2 to 10 mm), painful nodular lesions found in the pads of the fingers or toes ; Janeway lesions, which usually are painless hemorrhagic macular plaques that frequently occur on the palms and soles ; and Roth spots, which are small, pale retinal lesions associated with areas of hemorrhage located near the optic disk.

Other than fever and, perhaps, splenomegaly, no single sign or symptom occurs in more than 50% of children with endocarditis. That no classic clinical manifestation exists for this disease is obvious because the chance that even three or more signs will be present is extremely low. The appearance of any one of these clinical features in a child with predisposing heart disease should raise suspicion of infective endocarditis and should lead to an appropriate diagnostic evaluation.

The clinical findings of infective endocarditis in infants and neonates are less specific than are those in older children. The onset more often is acute and related to overwhelming infection. Infants with heart defects undergo corrective and palliative surgery at a younger age than in the past. Infants in whom postoperative endocarditis does develop probably will have clinical findings more similar to those in older children.

Infective endocarditis is an uncommon occurrence in neonates and frequently is associated with indwelling vascular catheters. It may affect the tricuspid valve and have a fairly “silent” clinical manifestation. Persistent bacteremia or fungemia should lead to a search for a cardiac focus of infection. Deterioration in pulmonary function, coagulopathies, thrombocytopenia, and low-grade murmurs often develop in neonates. Skin abscesses and hepatomegaly also are common findings.

Reported series of infective endocarditis in children with prosthetic valves are scarce. In early stages of disease, fever may be the only finding because the other signs of endocarditis are masked by the medical and surgical complications occurring in the immediate postoperative period. Late infections generally produce clinical findings similar to those in native valve endocarditis. Clinical evidence of systemic embolization occurs in as many as 40% of patients. Neurologic complications carry a particularly poor prognosis for survival. A new or changing murmur often indicates valvular insufficiency caused by a paravalvular leak. Florid cardiac failure is the major manifestation if local infection or an abscess creates valve instability and acute, severe regurgitation.

The signs and symptoms of infective endocarditis in intravenous drug users may be similar, but these patients have several more distinctive features of their illness. Two-thirds of these patients have no predisposing heart disease. The valve most commonly affected is the tricuspid, which leads to a predominance of pulmonary signs and symptoms resulting from pleural effusion, pulmonary infarction, and lung abscesses. Signs of tricuspid insufficiency (gallop rhythm, pulsatile liver, regurgitant murmur) are found in one-third of cases. Many patients have extracardiac sites of infection that are helpful in establishing the diagnosis.

Laboratory Findings

The most important diagnostic procedure is the blood culture. Because many bacteria that usually are not pathogenic cause infective endocarditis, scrupulous aseptic technique must be used to distinguish causative agents from contaminants. The yield of organisms is not increased by obtaining blood from arterial puncture or cardiac catheterization. The bacteremia usually is of low grade and continuous. The first two cultures yield the organism 90% of the time; in two-thirds of cases, all blood cultures are positive. Therefore, isolated positive cultures generally are not significant. Previous outpatient antibiotic therapy may change the yield significantly. In one study, culture positivity in cases of proven endocarditis was 64% in patients who received antibiotics before blood was drawn for culture versus 100% in patients without exposure to antibiotics.

When Candida endocarditis is suspected, several additional points should be considered. Isolation of Candida spp. may require incubation for 1 week or longer. All blood cultures from a patient with Candida endocarditis may not be positive, in contrast to the usual situation with bacterial endocarditis; several positive cultures may be interspersed among negative cultures. In patients with fungal endocarditis, Candida is isolated commonly from other infected sites, such as urine, sputum, synovial fluid, cerebrospinal fluid, lymph nodes, and bone marrow.

Three to five samples of blood for culture should be obtained from different sites within the first 24 hours in children with suspected endocarditis. Although difficult to obtain in smaller children, 3 to 5 mL of blood per culture is desirable for optimal yield. The samples should be injected into thioglycolate and trypticase soy (or brain-heart infusion) broth and held for at least 3 weeks to detect slow-growing organisms. However, one study has shown that the method of detection and not the time of incubation is critical to detect fastidious organisms. Baron et al. reported evidence that the Bactec9240 system can detect the HACEK organisms within 5 days. If gram-positive cocci grow in the broth but fail to grow on subculture, nutritionally variant streptococci should be suspected and subculture should be performed on media with either l -cysteine or pyridoxal phosphate.

Negative blood cultures are noted in 10% to 15% of patients with clinically diagnosed endocarditis. However, when patients have not received antibiotic therapy previously and blood for culture is obtained properly, these cases account for less than 5% of the total. Potential reasons for negative cultures include the following: (1) right-sided endocarditis; (2) previous administration of antibiotics; (3) fungal (especially Aspergillus ) endocarditis; (4) endocarditis caused by Bartonella spp., rickettsiae, chlamydiae, or viruses; (5) mural endocarditis; (6) slow growth of organisms ( Candida, Haemophilus, Brucella, nutritionally variant streptococci); (7) anaerobic infection; and (8) nonbacterial thrombotic endocarditis or an incorrect diagnosis. In some instances, intraleukocytic organisms may be seen in layered peripheral blood, even when cultures are negative. If surgical resection of vegetations or valve replacement is performed, a cause may be demonstrated by appropriate histologic examination and stains for bacteria and fungi. Molecular testing (universal bacterial, fungal, or mycobacterial (polymerase chain reaction [PCR]) for organisms in the valvular tissue may reveal the causative organism and likely will be more widely available in the future. Organisms also may be isolated from extracardiac sites (bone marrow, urine).

Many nonspecific laboratory findings are abnormal in patients with infective endocarditis ( Table 26.4 ). The total white blood cell count rarely is helpful, but peripheral eosinophilia may be seen with Loeffler endocarditis. Leukocytosis occurs in a few patients, but leukopenia is a rare finding in the absence of acute endocarditis with overwhelming sepsis. The erythrocyte sedimentation rate is elevated in 80% to 90% of cases. However, frequently it is normal or low when congestive heart failure or renal failure is present. Serum C-reactive protein levels usually are elevated initially and return to normal during the course of successful therapy. An increase during therapy may result from treatment failure, but it can also be caused by drug allergy or intercurrent infection. Rheumatoid factor rarely has been measured in a series of pediatric patients, but when measurements have been made, they have been positive in 25% to 50% of children with endocarditis. A positive test may be a diagnostic aid in cases of culture-negative endocarditis when other causes are excluded. Serial measurements may provide evidence of efficacy of therapy, although a fall in the titer of rheumatoid factor may lag behind the clinical and bacteriologic response. Hypocomplementemia is seen in association with glomerulonephritis. Anemia is present in approximately 40% of patients, especially those with long-standing disease. Although hemolysis may occur in the areas of turbulence in the heart, more often it is anemia of chronic disease. Because many patients with cyanotic heart disease normally have a compensatory polycythemia, a serial drop in hematocrit is of more significance than is a single measurement. Hematuria and proteinuria, present in 25% to 50% of cases, usually are secondary to microemboli in the kidneys and may be accompanied by “pyuria,” casts, and bacteriuria.

Circulating immune complexes are present in most adults with subacute endocarditis, as measured by Raji cell radioimmunoassay or the ¹²⁵ I-Clq binding assay. These immune complexes frequently are absent in acute endocarditis. Low levels of immune complexes have been found in 32% of adults with septicemia but not endocarditis, in 10% of normal controls, and in 40% of noninfected intravenous drug users. However, levels higher than 100 µg/mL are correlated highly with the presence of endocarditis. Serial measurement of immune complex levels may aid in monitoring therapeutic efficacy. Systematic investigation of immune complexes has been reported infrequently in children with endocarditis. When immune complexes have been sought, most patients, including two of three children with culture-negative endocarditis, have had significant levels.

When infective endocarditis is suspected but blood cultures remain negative, serologic testing for specific organisms may prove helpful. Antibodies to teichoic acid, major components of the S. aureus cell wall, are present in more than 85% of adults with staphylococcal endocarditis, but the false-positive rate is as high as 10%. False-negative results correlate with a short (<2 weeks) duration of illness. Specific information about the accuracy of this test in children is lacking, and the tests are not readily available. Serologic testing is available or under investigation for many other organisms that cause infective endocarditis, including Bartonella, Brucella, Candida, Aspergillus, Histoplasma, Cryptococcus, Chlamydia, and Coxiella. In general, the usefulness of these tests in children with endocarditis is unproven. Some patients with nonspirochetal bacterial endocarditis who reside in locales endemic for Lyme disease have significantly elevated levels of antibodies reactive to Borrelia burgdorferi. Diagnostic confusion may occur because the signs and symptoms of infective endocarditis and Lyme disease can be quite similar. Other techniques, such as broad-range PCR, have been used to identify Bartonella and other causative agents of endocarditis.

Radiographic techniques have not been a great aid in establishing the diagnosis of infective endocarditis. The findings on plain chest radiographs are nonspecific, but evidence of complications, such as septic pulmonary emboli or congestive heart failure, may be helpful. Computed tomography may help in establishing the diagnosis of an infected shunt. Immunoscintigraphy using technetium-labeled antigranulocyte antibodies has been reported as being useful in adults when the echocardiographic findings were equivocal.

The electrocardiogram also is useful in the evaluation of patients with endocarditis because it detects arrhythmias and conduction disturbances that complicate the disease. Ventricular ectopy may be related to myocardial ischemia, myocarditis, or myocardial abscess. New conduction defects imply extension of infection beyond the valve ring into the myocardium. Any degree of atrioventricular block, a new left bundle branch block, or a new right bundle branch block with a left anterior hemiblock may represent extension of infection from the aortic valve into the ventricular septum. Junctional tachycardia, Wenckebach atrioventricular block, or complete heart block may be produced by extension of the infection from the mitral valve anulus into the atrioventricular node or proximal His bundle. In general, an unstable conduction block is more likely to develop in patients with aortic valve endocarditis than in those with mitral infection.

Echocardiography has become a valuable adjunct to the diagnosis and treatment of endocarditis in children. Color Doppler imaging is a sensitive modality for detection of valvular insufficiency, and the results may influence surgical and medical treatment decisions. Echocardiography can be performed by the traditional transthoracic approach or the transesophageal approach. The sensitivity and specificity of transthoracic echocardiography continue to be defined, with positive results obtained in 36% to 100% of children in various series of pediatric patients.

In general, two-dimensional echocardiography is more sensitive than is the M-mode technique, especially in cases of right-sided endocarditis, and it is superior in diagnosing complications of the destructive process. The smallest vegetation detectable is approximately 2 mm, but the acoustic impedance of the mass relative to the surrounding structures is a more important factor than is size in identifying the vegetation. Echocardiography has identified vegetations in culture-negative cases. Its accuracy in prosthetic valve endocarditis is diminished by the difficulty in resolution around the prosthetic device. Serial evaluation of valvular vegetations generally does not assist in assessing the efficacy of antibiotic therapy because diminution or disappearance of vegetations may take place long after successful medical treatment has been completed.

The use of transthoracic echocardiography to predict the clinical course and need for operative intervention in patients with endocarditis is controversial. A synopsis of many reports that have assessed the role of transthoracic echocardiography in the diagnosis and management of infective endocarditis suggests the following: (1) because of variable sensitivity among studies for detection of vegetations, a negative study does not rule out endocarditis, especially when foreign material is present within the heart; (2) false-positive studies are quite rare (the specificity is high); (3) the reliability of transthoracic echocardiography depends on the experience of the examiner and the technical adequacy of the study; (4) transthoracic echocardiography is valuable in assessing local complications of endocarditis on native valves; and (5) in most but not all studies, patients with a vegetation identified by transthoracic echocardiography have an increased risk for the development of systemic emboli and congestive heart failure.

Although some investigators contend that the presence of a vegetation should hasten early surgery, most suggest that a positive echocardiogram is adjunctive evidence that should be considered along with other clinical parameters when considering surgical intervention. One study suggested that the relative risk for having embolic events associated with echocardiographically visualized lesions is microorganism dependent, with a significant attributable risk seen, for instance, in patients with viridans streptococcal infection. The absence of a vegetation on transthoracic echocardiography may define a subset of patients at low risk for the development of embolic complications.

Transesophageal echocardiography (TEE) has been studied extensively in adults with infective endocarditis. It uses a 5-MHz phased-array transducer with Doppler and color flow encoding capabilities mounted on the tip of a flexible endoscope. Biplane TEE is considered the standard technique and is superior to transthoracic echocardiography because of improved spatial resolution, lack of acoustic interference from the lungs and chest wall, and closer proximity to posterior structures, such as the mitral valve and left atrium. Multiplane TEE facilitates and abbreviates the examination procedure and may be more accurate in providing the dimensions of a vegetation associated with infective endocarditis.

TEE generally is well tolerated by children, even with the use of an adult probe (when the child’s weight is more than 7 kg), and rarely is associated with bacteremia. TEE usually is more sensitive than is transthoracic echocardiography in the detection of intracardiac vegetations and is positive in 70% to 95% of adults with strongly suspected endocarditis. One recent study showed that TEE significantly increased the detection of vegetations in bigger (>60 kg) children but did not improve on the results of transthoracic echocardiography in smaller children. It is significantly more sensitive in the detection of vegetations and complications in infected prosthetic valves. TEE is particularly useful for detecting an aortic root abscess or involvement of the sinus of Valsalva in adults, and it should be considered in children with aortic valve endocarditis and changing aortic root dimensions on a standard transthoracic echocardiogram. It appears to be less helpful for detection of vegetations in right-sided endocarditis. Although a negative transesophageal echocardiographic study does not exclude endocarditis, the procedure should be considered for patients with suspected endocarditis and a negative transthoracic echocardiogram, when the transthoracic echocardiographic windows are suboptimal, and when perivalvular extension of infection is suspected.

To aid in establishing the diagnosis of infective endocarditis, various sets of clinical criteria have been suggested. The most widely used diagnostic criteria were proposed by investigators from Duke, and they have been modified subsequently ( Boxes 26.1 and 26.2 ). These criteria have been validated in large series of infective endocarditis in adults and children. In two pediatric series of clinically defined endocarditis, no cases were rejected by the Duke criteria, whereas 25% and 19% were rejected by older criteria. However, one study found that 12% of pediatric endocarditis cases were not classified as “definite” by the modified Duke criteria. In addition, the presence of an indwelling venous catheter causing prolonged bacteremia may cause an overestimation of the rate of infective endocarditis using the Duke criteria.

Microbiology

Many different microorganisms are capable of causing infective endocarditis in humans. A list of the organisms isolated from patients in major pediatric series is presented in Table 26.5 . Gram-positive cocci are the etiologic agents in 90% of cases in which an organism is isolated. Streptococci remain the bacteria isolated most frequently, although the percentage of cases caused by staphylococci and fungi has been increasing during the past two decades. In a preliminary study, Gupta et al. used the Nationwide Inpatient Sample to study the incidence, pathogens, and outcomes of infective endocarditis in children admitted to hospitals in the United States from 2000 to 2010. Streptococcus spp. were most common (40.1%), followed by S. aureus in 36.6% of the 3840 patients reported. Polymicrobial infective endocarditis, especially in nosocomial settings, also appears to be increasing in incidence. The characteristics of selected organisms and the type of disease that they produce are considered next.

Treatment

In the preantibiotic era, infective endocarditis was a uniformly fatal disease. With the current improved methods of diagnosis and therapy, 80% to 90% of children with this disease can be expected to survive. Mortality rates are higher for acute staphylococcal infection, fungal endocarditis, and prosthetic valve endocarditis, although the tendency toward earlier surgical intervention for these entities may improve survival rates. The cornerstone of successful therapy is selection of antibiotics with specific activity against the causative organism. Better analysis of pharmacodynamic variables, such as bactericidal activity and the postantibiotic effects of various drugs, may assist in the selection of optimal therapeutic regimens. Although persistent infection occasionally complicates treated endocarditis, deterioration in cardiac function is the major cause of morbidity and mortality.

Several general principles provide the basis for the current recommendations for treatment of endocarditis. Parenteral administration of antibiotics is preferred because erratic absorption of oral antibiotics, especially in infants, can lead to therapeutic failure. Although patient selection criteria for the use of outpatient parenteral antibiotic therapy for endocarditis in adults have been suggested, no data have been published about this practice for children. The 2015 AHA Pediatric endocarditis treatment guidelines state that home parenteral therapy can be considered after initial treatment in the hospital in selected patients who are stable, afebrile, have negative blood cultures, and are at low risk for a complication (not fungal endocarditis or young age). Prolonged treatment, usually 4 to 6 weeks or longer, is necessary to sterilize the vegetations and to prevent relapse. Bacteriostatic antibiotics are not effective and lead to frequent relapses or failure to eradicate the infection, or both. Antibiotic combinations may produce a rapid bactericidal effect through synergistic mechanisms of action. When synergy exists, smaller doses of each drug may be used, thereby reducing toxic side effects.

Blood should be drawn for culture for several days to evaluate the effect of the antibiotics. Negative follow-up cultures do not guarantee the success of therapy, but persistent positive cultures usually require that a change or addition to the antibiotic regimen be made. Observation of the patient’s clinical course is extremely important. When fever is present initially, the temperature often returns to normal within a few days after therapy is started. However, fever can persist for weeks in patients whose eventual outcome is good. Such patients must be monitored closely for cardiac arrhythmias and congestive heart failure, which may require intensive care observation and electrocardiographic monitoring. Evidence of major embolic phenomena must be sought diligently by physical examination.

Several laboratory tests may aid in monitoring therapy. In all cases of bacterial endocarditis, the minimal inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) ideally are determined for the antibiotics being used because disk susceptibility testing may not be as reliable and is not quantitative. The role of monitoring the inhibitory and bactericidal activity of the patient’s serum is highly controversial. Standardization of this test is poor, with laboratories using variations in inoculum size, in composition of the broth, in timing of samples (at expected peak or trough antibiotic concentrations in serum), in methods of dilution, and in determining the bactericidal end point. In the rabbit endocarditis model, peak serum bactericidal titers greater than 1 : 8 correlate with therapeutic success. A retrospective review of 17 reports of serum bactericidal activity in patients with endocarditis failed to show any correlation between titers greater than 1 : 8 and therapeutic success. A prospective study suggested adjusting antibiotic doses to achieve peak titers of 1 : 64 or greater and trough titers of 1 : 32 or greater. At present, no generally accepted recommendation can be made.

Little information is available concerning optimal antibiotic therapy for infective endocarditis in children; most treatment regimens are adapted from studies of adults with endocarditis. In general, these regimens have been equally successful (and generally less toxic) in children. Recommended doses of the antibiotics commonly used are listed in Table 26.6 .

After performing the initial evaluation of a patient with suspected infective endocarditis, the physician must make a clinical judgment about when to initiate therapy. If the findings are strongly indicative of the diagnosis or the child is very ill, treatment should be started as soon as blood has been drawn for culture. Initial empiric therapy depends on the clinical setting in which the tentative diagnosis is made. If the infection is subacute, a combination of penicillin G and an aminoglycoside usually is recommended for its activity against viridans streptococci, enterococci, and most gram-negative organisms. If S. aureus endocarditis is a strong consideration (acute manifestation, narcotic addicts), vancomycin and a penicillinase-resistant penicillin should be added to this regimen. Patients who recently have undergone cardiac surgery, especially prosthetic valve placement, are treated best with an aminoglycoside and vancomycin to “cover” for health care–associated infection caused by MRSA or CONS; some physicians add penicillin G to this regimen to improve activity against streptococci. When culture and susceptibility data are known, antibiotic therapy can be changed as needed.

Most strains of viridans streptococci, S. pyogenes, and nonenterococcal group D streptococci are exquisitely susceptible to penicillin, with an MIC of less than 0.2 µg/mL. However, 15% to 20% of viridans streptococci have an MIC of 0.2 µg/mL or greater and are defined arbitrarily as relatively resistant. In addition, some strains (particularly S. mutans and S. mitior ) demonstrate tolerance; that is, an MIC to penicillin of less than 0.1 µg/mL but an MBC that is more than 10-fold higher (1.25 to 50 µg/mL). Most strains of nutritionally dependent streptococci are tolerant to penicillin. Clinical failure may occur in endocarditis caused by these tolerant organisms when penicillin alone is used for treatment. However, except for nutritionally dependent streptococci, therapy for tolerant viridans streptococci generally should be the same as for susceptible strains.

Although most experts recommend that patients with endocarditis caused by relatively resistant streptococci be treated with high doses of penicillin combined with 2 to 4 weeks of an aminoglycoside, some authorities consider that penicillin alone usually is adequate therapy. Synergy in vitro between penicillin or vancomycin and streptomycin or gentamicin can be demonstrated against virtually all penicillin-susceptible streptococci. This observation correlates with a faster rate of eradication of bacteria from cardiac vegetations in the rabbit endocarditis model when synergistic combinations of antibiotics are used. However, streptomycin is not synergistic for strains with high-level streptomycin resistance; gentamicin is the preferred second drug for these rare isolates. In pediatric patients, gentamicin usually is substituted for streptomycin because of its lower toxicity.

Several regimens have been examined in adults with penicillin-susceptible viridans streptococcal native valve endocarditis ( Table 26.7 ). A 2-week course of penicillin alone leads to an unacceptable relapse rate. However, a 2-week course of intramuscular procaine penicillin and streptomycin cured 99% of adults with penicillin-susceptible streptococcal endocarditis in one report. These results are similar to those obtained with β-lactams alone for 4 weeks or with penicillin for 4 weeks combined with streptomycin for the first 2 weeks. Gentamicin may be substituted for streptomycin. The 2-week penicillin-gentamicin regimen is the least expensive and is the preferred therapy in uncomplicated cases of penicillin-susceptible streptococcal endocarditis in young adults. In general, the regimen of 4 weeks of penicillin alone is preferred for patients in renal failure or at high risk for developing aminoglycoside-induced ototoxicity. Vancomycin or ceftriaxone administered for 4 weeks can be used in patients with penicillin-susceptible viridans streptococcal endocarditis who have a penicillin allergy. A 4-week regimen of penicillin plus an initial 2 weeks of gentamicin is recommended in children with infection caused by relatively penicillin-resistant organisms ( Table 26.8 ). Most nutritionally deficient streptococci are tolerant to penicillin and should be treated as for enterococci ( Table 26.8 ). In patients with streptococcal infection of prosthetic valves or other prosthetic materials, a 6-week regimen of penicillin usually supplemented with an aminoglycoside is recommended ( Tables 26.7 and 26.8 ). None of the regimens discussed has been evaluated specifically in children with endocarditis.

Stay updated, free articles. Join our Telegram channel

Join