Table 284.2 summarizes the most common laboratory findings in children with IE. Blood cultures are the single most important diagnostic tool for establishing the diagnosis of IE. Because bacteremia usually is continuous and low-grade, the timing and site of collection do not affect the yield. In approximately 66% of all cases of IE, the blood cultures grow the infecting organism and, in 90%, the results of the first two blood cultures are positive. If a patient has received antibiotics before the culture is tried, the chance of obtaining a positive culture may be reduced from between 95% and 100% to 64%. If pretreatment has occurred, placing the blood sample in hypertonic media may enhance the chance of isolating the organism. Patients with fungal endocarditis may have only intermittently positive blood culture results, and the organism may take a week or longer to grow in culture.

Ideally, three to five sets of blood cultures should be obtained within the first 24 hours. The commonly practiced simultaneous drawing of blood for culture with fever is not supported by any studies. Bacteremia in endocarditis usually is continuous and thus can be drawn at any time. In children in whom drawing large volumes of blood would be contraindicated, blood should be cultured in aerobic conditions because endocarditis caused by an anaerobic organism is an extremely rare event.
Specimens should be taken from different sites and should contain 3 to 5 mL of blood. Thioglycolate broth should be used, and the bottles should be kept for 3 weeks to detect slow-growing organisms. Nutritionally variant streptococci should be suspected if a gram-positive coccus is isolated in broth but fails to grow in subculture. The organism then should be subcultured onto media enriched with pyridoxal phosphate or L-cysteine.

Blood culture results may be negative in 10% to 15% of cases of suspected endocarditis because of prior administration of antibiotics; endocarditis caused by rickettsiae, chlamydiae, or viruses; slow-growing organisms (e.g., Candida spp., Haemophilus spp., Brucella spp.) or nutritionally variant streptococci; infections caused by anaerobic organisms; NBTV endocarditis; mural endocarditis; right-sided endocarditis; fungal endocarditis (especially Aspergillus spp.); or an incorrect diagnosis. Additional sites, including urine, sputum, cerebrospinal fluid, synovial fluid, bone marrow, and lymph nodes, may be infected concomitantly, and cultures of these additional sites should be included if blood culture results fail to demonstrate an infecting agent.

Necessary adjunctive laboratory tests include measurement of the erythrocyte sedimentation rate, which is elevated in as many as 90% of patients and correlates with the hypergammaglobulinemia found in this disease. An artifactually low erythrocyte sedimentation rate may be found with renal disease, severe CHF, or polycythemia. The erythrocyte sedimentation rate should decrease toward normal if therapy has been successful, and serial measurements may be helpful in monitoring therapy. A positive rheumatoid factor has been found in one-fourth to one-half of pediatric patients with IE and may be supportive evidence for the diagnosis in cases of culture-negative endocarditis. Immune complex-mediated glomerulonephritis, although not a common finding, may result in hypocomplementemia. Anemia, usually caused by a chronic disease state, is found in 40% of patients. Because children with cyanotic CHD frequently develop IE, the finding of a normal or slightly high hemoglobin level should suggest the possibility of anemia. Although leukocytosis is not an uncommon occurrence, leukopenia may occur in acute cases with overwhelming sepsis. Microemboli and consequent microinfarcts in the kidney produce hematuria with or without proteinuria, casts, and bacteremia in 25% to 50% of patients.

Circulating immune complexes as measured by the Raji cell radioimmune assay or the ¹²³I-C1q-binding assay are present in most adults with IE, although they may be conspicuously absent in acute disease. Immune complexes may be found in septicemic patients and in as many as 10% of normal adult controls. Although few studies have reported on circulating immune complex levels in children with IE, elevated levels have been found in some patients.

Serologic testing for specific organisms may be helpful in cases of culture-negative endocarditis. Antibodies against teichoic acids, which are major components of the cell wall in S. aureus, may be present in 85% of adults with IE, although the false-positive rate may be as high as 10%. A teichoic acid antibody titer of greater than 1:1 in a bacteremic adult patient correlates with deep-seated infections, including but not restricted to endocarditis. However, teichoic acid antibody levels do not differentiate the types of deep-seated infection. As with circulating immune complexes, serial measurements of teichoic acid antibody levels may correlate with successful therapy. Unfortunately, no data exist concerning teichoic acid antibody levels in children with IE.

The role of echocardiography in helping to establish the diagnosis of IE has grown considerably because the technology has improved vastly since single-crystal M-mode techniques were introduced. M-mode echocardiography uses a single, narrow ultrasound beam, and the reflected ultrasound echoes are recorded on moving paper. Two-dimensional (2D) echocardiography uses multiple echo beams and provides a cross-sectional moving image of the heart, which can be recorded from several angles. The 2D technique usually is superior to the M-mode technique in investigating vegetative lesions. Some areas of the heart, such as the pulmonary valve, are visualized better using 2D imaging, and the increased sensitivity of this technique in detecting pulmonary valve endocarditis has been documented. The value of 2D echocardiography in prosthetic valve endocarditis also is superior to that of M-mode echocardiography. An echo-free space found in more than one tomographic plane has aided in establishing the diagnosis of perivalvular abscess, which may complicate 30% of the native valve endocarditis cases and even more cases of prosthetic valve endocarditis. Adjunctive use of the Doppler technique to diagnose prosthetic valve regurgitation before it becomes apparent clinically may aid in earlier establishment of the diagnosis. The sensitivity of echocardiography in detecting vegetative lesions in suspected endocarditis in adults ranges from 13% to 83%, with a greater sensitivity exhibited in the more recently published series using the 2D technique. Several studies have concluded that patients with a “positive echo” were twice as likely to develop serious complications (usually emboli, more commonly cerebral than peripheral) and that patients were at higher risk for death or development of severe CHF if the vegetation were larger than 1 cm². However, the risk of embolization does not appear to correlate with the size of the vegetation.

Transesophageal echocardiography (TEE) has become important in evaluating patients for vegetations. TEE takes advantage of the proximity of the heart, especially the left atrium to the esophagus. This technique also eliminates the inability to image transthoracically in some patients who do not have good “echo windows” from that approach. The quality of the image is better, rendering diagnosis more certain. Although this technique can be accomplished using light sedation, our experience has been that heavy sedation administered by an anesthesiologist is preferable.

One study in adults with endocarditis using the transesophageal approach yielded the following conclusions: When combined with a high clinical suspicion of endocarditis, TEE offers a high sensitivity for properly diagnosing IE, and TEE is significantly better than is transthoracic echocardiography (TTE) in imaging vegetations in these patients. Our clinical experience has confirmed these conclusions, and we now routinely use TEE if the TTE result is equivocal or negative for a patient who we suspect has the disease.

In a study using M-mode and 2D techniques for 15 children with IE, vegetations were detected in ten (66%) of the patients. Of the ten, three had systemic emboli and two died, but none of the patients without an echocardiography-proven vegetation had an embolic complication. This finding suggested that echocardiography may help identify a high-risk population that could benefit from surgical intervention. Some studies have suggested that the size and mobility of a vegetation may be useful in differentiating bacterial from fungal endocarditis. In our experience, however, this claim probably is not valid because large vegetations are seen frequently in bacterial or fungal endocarditis.

Several recent studies have addressed the role for TEE versus TTE in evaluating patients for endocarditis. Generalizing the results, these investigators found that TEE should be reserved for patients in whom TTE has been technically unfeasible, such as in patients who are very obese, are overly muscular, or have concomitant pulmonary disease, all of which result in poor “acoustic windows.” Moreover, as a “screening tool,” any type of echocardiographic study should be reserved for those patients in whom a high index of suspicion exists as
outlined by either the clinical criteria of von Reyn or the so-called “Duke criteria.” In the absence of a moderate to high suspicion of risk, echocardiographic studies are not cost-efficient. TEE also has the potential risks involved with general anesthesia and endotracheal intubation in order to accomplish the study.

A negative echocardiographic study result does not rule out the presence of vegetations. The resolution on most equipment limits the detection of vegetations to those larger than 2 to 3 mm. Poor technique also may hinder evaluation. Rheumatic heart disease with preexisting valve disease, mitral valve prolapse (MVP) with thickened leaflets, marantic vegetations, Löffler endocarditis, Chiari networks in the right atrium, and valve ring abscesses pose interpretive problems to the echocardiographer. After a vegetation is detected, it may show no significant change during therapy, and a recurrence of disease cannot be diagnosed unless a noticeable increase in size occurs or a new vegetation appears. Attempts to estimate serially the size of a vegetation are fraught with technical and interpretive errors and probably are of little value. However, continued growth of a vegetation coexistent with persistent bacteremia or evidence of further endocardial infiltration may indicate a treatment failure or the need for surgical intervention. One study attempted to define risk factors involved with an echocardiographically demonstrable vegetation in adult patients with IE. Significant complications included peripheral and central nervous system (CNS) embolization, failure to respond to therapy, CHF, need for surgery, and death. Using univariate analysis in patients with native left-sided valve endocarditis, the investigators found that vegetation’s size, extent, mobility, and consistency all were significant predictors of complications. Multivariate analysis demonstrated that size, extent, and mobility could be used to score a patient into a high-risk group. An echocardiographic examination performed at the time of hospital discharge of a patient after an apparently successful course of antibiotic therapy can be used as a baseline for further evaluation.

Numerous electrocardiographic abnormalities may be found throughout the course of IE. Ventricular ectopy in patients with hemodynamic compromise may be life-threatening. Atrial fibrillation in adults and children may be secondary to atrioventricular valve regurgitation. Extension of abscess formation or an inflammatory response may cause direct injury to the conduction system. Complete right bundle branch block, left anterior or posterior fascicular block, and complete atrioventricular block have been reported. The onset of complete atrioventricular block may be an indication for surgical intervention, especially if it occurs during the course of appropriate antibiotic therapy. Abscess formation in the perivalvular aortic region may cause direct injury to the atrioventricular node because of its proximity to that structure, which may result in sudden death unless temporary and eventually permanent pacing is instituted.