Infective Endocarditis:

The incidence of infective endocarditis (IE) has remained unchanged over the past several decades, with only a slight decrease in associated mortality. Factors that affect the incidence of IE include the increase in device-related infections that has resulted from the greater number of invasive procedures and indwelling catheters. In addition, diagnosis and treatment have been complicated by the involvement of organisms that are difficult to culture or that are resistant to conventional antimicrobial therapy. Several new drugs may effectively treat IE caused by resistant strains of bacteria.

In this article, we update developments in IE by addressing a number of questions physicians commonly raise concerning prophylaxis, diagnosis, and management.

1. Is there evidence that supports prophylaxis for IE?

No randomized controlled trials in patients with structural heart disease have definitively established that antibiotic prophylaxis prevents endocarditis after procedures that put patients at risk for bacteremia. Few infections occur after medical or dental procedures.¹ Tooth extractions are associated with an elevated risk of bacteremia, but a causal relationship between tooth extraction and endocarditis has not been established.²

A recent randomized, double-blind, placebo-controlled trial evaluated the impact of amoxicillin prophylaxis on bacteremia during nasotracheal intubation and dental procedures in children with no structural heart disease.³ The highest incidence of positive blood cultures (46%) (mostly for Gram-positive cocci) occurred 1.5 minutes after the last tooth extraction, and the overall incidence of positive blood cultures was significantly lower among patients who received amoxicillin (33% vs 84%). However, similar rates of bacteremia are seen after routine activities such as tooth brushing, flossing, and chewing food, and no study has demonstrated the prevention of IE by prophylactic antibiotics during invasive procedures in adults.⁴ In a case-control study of 273 patients with IE, dental treatment was no more frequent among case patients than among controls.⁵ IE prophylaxis in children with cardiac lesions who undergo urinary catheterization has not proved cost-effective.⁶

Because of the absence of strong evidence, recommendations for antibiotic prophylaxis are changing rapidly. However, prophylaxis continues to be strongly recommended in patients with intravascular hardware, including prosthetic valves, automatic implantable cardioverter defibrillators (AICDs), pacemakers, and left ventricular assist devices (Table). In addition, prophylaxis is advised for patients with a history of IE and those with complex congenital heart disease, such as single ventricle states, transposition of the great arteries, and tetralogy of Fallot.

2. What is the role of molecular diagnostics in culture-negative IE?

Conventional microbiologic techniques may not detect causative microorganisms, whether because of previous antibiotic therapy, suboptimal collection of specimens, or infection attributable to fastidious, nonculturable agents. Polymerase chain reaction (PCR) assays allow rapid detection in such cases, using blood samples and excised valves from patients with suspected IE.

In a recent study, PCR assays, along with traditional cultures and Gram staining, were performed on excised heart valves from patients with (n = 51) and without (n = 16) suspected IE.⁷ The sensitivity of bacterial broad-range PCR was 41.2%, compared with 7.8% and 11.8% for culture and Gram staining, respectively. Similarly, the positive predictive value (PPV) and negative predictive values (NPV) for PCR were greater (100% and 34.8%, respectively) than those for culture (PPV = 80%, NPV = 24.2%) or Gram staining (PPV = 100%, NPV = 26.2%).

Houpikian and Raoult⁸ were able to identify the culprit agent in 275 of 348 patients with culture-negative endocarditis using serologic tests, blood culture on shell vial, and culture and direct PCR of valve specimens. The most frequently identified pathogens were Coxiella burnetii (48%) and Bartonella (28%). The prevalence of these organisms varies widely among geographic regions; the findings of this study reflect the high prevalence of Bartonella and Coxiella infections in southern France, where the study was performed. Other studies from the United States and Europe have found streptococci to be a common cause of culture-negative IE.^9,10

The type of specimen chosen for molecular diagnosis depends on the stage of disease. In early stages, blood cultures may be used; embolic tissue has been examined to identify causative organisms in more progressive stages. Heart valves and vegetations are used for molecular diagnosis in more advanced stages, either during cardiac surgery or postmortem examination.^7,10

Like any diagnostic modality, molecular detection of pathogens has limitations. Inadequate removal of inhibitory factors during specimen preparation, or difficulties in primer annealing because of nonoptimal temperature or concentration of components, can lead to false-negative results. False-positive results occur with primers that are homologous to sequences other than the target gene and also as a result of contamination of the clinical specimen.¹¹

DNA amplification does not provide information about susceptibility to antibiotics; therefore, culture-based methods are still important.

3. What is the risk of IE in patients with Staphylococcus aureus bacteremia?

S aureus is a leading cause of bacteremia and endocarditis; the use of indwelling catheters has led to an increased incidence of endocarditis. A prospective observational study that evaluated risk factors for complicated S aureus bacteremia reported a 12% prevalence of IE in 724 consecutive patients admitted with 1 or more positive blood cultures.¹² This is significantly lower than in a previous study at our institution, in which the prevalence of IE was almost 25%.¹³ This disparity is probably a result of selection bias, because only patients undergoing both transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) evaluation were included in the latter study.

For 12 weeks, Fowler and associates¹² monitored patients who had 1 or more blood cultures that were positive for S aureus. Risk factors for complicated S aureus bacteremia were identified as positive follow-up blood cultures 48 to 96 hours after admission (odds ratio [OR], 5.58); persistent fevers 72 hours after admission (OR, 2.23); community acquisition (OR, 3.10); and skin findings suggestive of acute systemic infection (OR, 2.04). A risk scoring system based on these 4 criteria predicted a complication rate of 16% when no risk factors were present and 90% when all risk factors were present. In this study, the outcomes for patients with methicillin-resistant S aureus (MRSA) bacteremia were similar to those for patients with methicillin-sensitive S aureus (MSSA) bacteremia, although different results have been observed in other studies.

Abraham and coworkers¹⁴ compared the incidence of IE in patients with MSSA and MRSA bacteremia. MSSA was associated with significantly higher rates of IE, although most of the patients with MSSA had community-acquired bacteremia, while most MRSA endocarditis was acquired nosocomially. Higher rates of IE among patients with MRSA bacteremia were noted in cases of nosocomially acquired bacteremia. The emergence of community-acquired MRSA, with its unique exotoxin gene profile, points to the changing profile of staphylococcal infection and is the focus of ongoing research.

In an earlier study at our institution, Fowler and colleagues¹³ reported that using TEE as well as TTE to evaluate patients with S aureus bacteremia increased the sensitivity of IE detection from 32% to 100%. In addition, higher rates of severe sepsis, strokes, and multiorgan failure, as well as increased mortality from sepsis, were observed in patients in whom S aureus bacteremia was complicated by IE.

It is essential that patients who have S aureus bacteremia undergo aggressive evaluation, including follow-up blood cultures 48 hours after admission, as well as TEE, to rule out IE even without physical examination findings or metastatic foci.

4. What are the treatment options for vancomycin-intermediate S aureus (VISA) and vancomycin-resistant S aureus (VRSA) endocarditis?

The first glycoprotein-intermediate S aureus isolate was reported in Japan in 1997; since then, several reports of VISA infections in the United States have been published. The primary cause of reduced susceptibility to vancomycin is the presence of a thickened cell wall with several additional peptidoglycan layers. Vancomycin binds to the many D-alanine-D-alanine residues and is thus unable to reach the surface of the cytoplasmic membrane.

Risk factors for VISA infection include a previous MRSA infection treated with prolonged courses of vancomycin. A history of dialysis and the presence of indwelling catheters have also been associated with an increased incidence of VISA infections in some studies.¹⁵

Successful treatment of VISA endocarditis has been reported with combinations of various antibiotics, such as linezolid and amikacin¹⁶ or vancomycin and linezolid.¹⁷ Woods and colleagues¹⁸ treated a patient who had VISA IE with a 67-day course of linezolid alone. The patient died of respiratory failure, but autopsy findings revealed no cardiac vegetations.

5. What are the latest therapies for the treatment of endocarditis caused by Gram-positive organisms?

Newer antistaphylococcal agents, such as daptomycin, dalbavancin, and telavancin, show great promise in the treatment of VISA infections, although more studies of these agents in the treatment of endocarditis are needed.

Daptomycin. Daptomycin is a rapidly acting bactericidal cyclic lipopeptide derived from the fermentation of Streptomyces roseosporus. Destruction of cell membrane function results in bactericidal activity against a variety of Gram-positive organisms, including MRSA, VISA, VRSA, and vancomycin-resistant enterococci. Advantages include once-daily dosing, favorable side-effect profile, and low potential for resistance. In a recent study involving 1092 patients with complicated skin and soft-tissue infections, daptomycin was as effective as vancomycin and penicillinase-resistant penicillins.¹⁹ Successful treatment of MSSA, vancomycin- resistant enterococci, and MRSA IE with daptomycin was reported in a small trial.²⁰

However, there have also been reports of clinical failure following prolonged treatment with daptomycin.^21,22 In a recent study that compared daptomycin with vancomycin or nafcillin for treatment of complicated staphylococcal bacteremia, Fowler and colleagues²³ reported similar failure rates in both groups (daptomycin, 55% vs vancomycin/ nafcillin, 58%). However, reasons for failure differed between the 2 groups: in the daptomycin group, 6 patients had persistent or recurrent bacteremia, and in 5 of these patients there was an increase in daptomycin minimum inhibitory concentration from less than 1 µg/mL to 2 µg/mL. (Most apparently had undrained pus.) The main reason for failure with vancomycin was renal failure.

Tigecycline. Tigecycline is a semisynthetic glycylcycline with a broad range of activity against Gram-positive and Gram-negative bacteria, including vancomycin-resistant enterococci, MRSA, and many species of multidrug-resistant Gram-negative bacteria. Two double-blind phase 3 trials that evaluated the safety and efficacy of tigecycline in patients with intra-abdominal and skin and soft tissue infections have recently been published.^24,25 No statistically significant difference in the clinical cure rates for intra-abdominal infections was seen in the tigecycline group (80.2%) compared with the imipenem/cilastatin group (81.5%). Similar results were observed when tigecycline was compared with vancomycin and aztreonam for the treatment of skin and soft tissue infections. The most common adverse effects with tigecycline in both studies were nausea and vomiting.

Telavancin. Telavancin is a novel lipoglycopeptide that demonstrates concentration-dependent bactericidal activity against Gram-positive bacteria, including MRSA, VISA, and VRSA strains. A recent randomized, double-blind, placebo-controlled phase 2 clinical trial compared the efficacy and side-effect profile of telavancin with that of standard therapy for treatment of complicated skin and soft tissue infections caused by suspected or confirmed Gram-positive organisms.²⁶ At least 1 dose of study medication was given to 167 patients (84 in the telavancin group and 83 in the standard-therapy group). Although the overall success rates were similar for both groups, patients with MRSA infection demonstrated higher cure rates (82% vs 69%) and microbiologic eradica- tion (84% vs 74%) when treated with telavancin.

Dalbavancin. Dalbavancin is a bactericidal semisynthetic glycopeptide shown to be 16-fold more active than vancomycin against staphylococci in vitro. In addition to an excellent safety profile, this drug offers once-weekly dosing, a unique advantage for the treatment of complicated staphylococcal infections.

Clinical success rates during a phase 2 clinical trial involving 62 patients with skin and soft tissue infections were 94.1% among patients treated with 2 doses of dalbavancin, 61.5% among patients treated with 1 dose of dalbavancin, and 76.2% among patients treated with a standard-of-care regimen.²⁷ Limitations of the study included small sample size and differences in the treatment durations for the 3 groups, with earlier end-of-treatment follow-up assessment visits for the 1-dose dalbavancin group than for either of the other 2 groups. Dalbavancin also appeared to be more effective than vancomycin in the treatment of Gram-positive, catheter-related bloodstream infections, although the number of true pathogens, such as S aureus, was quite small.

Oritavancin. Oritavancin is a glycopeptide antibiotic closely related to vancomycin that demonstrates bactericidal activity against staphylococci. No clinical trials have evaluated its efficacy in bloodstream infections, but oritavancin was found to be as effective as vancomycin in the treatment of complicated skin and soft tissue infections.²⁸

Although some of the glycopeptides mentioned above show significant potential for treating Gram-positive bloodstream infections, further randomized controlled trials are required to evaluate their efficacy in the treatment of IE.

Staphylococcal monoclonal antibody and vaccines. Tefibazumab, a humanized monoclonal antibody that exhibits a high affinity and specificity for the S aureus MSCRAMM (microbial surface components recognizing adhesive matrix molecules) protein ClfA, has been recently developed and is undergoing further evaluation in phase 2 studies.²⁹

StaphVAX is a staphylococcal vaccine consisting of types 5 and 8 capsular polysaccharides, the strains that account for more than 80% of S aureus infections. A phase 3 clinical trial that evaluated the efficacy of StaphVAX in 1804 hemodialysis- dependent patients showed a 57% reduction in S aureus bacteremia at 10 months compared with placebo. However, a recent confirmatory phase 3 clinical trial in 3600 hemodialysis-dependent patients found no reduction in S aureus types 5 and 8 infections in the StaphVAX group compared with the placebo group.³⁰

6. What is the standard of care for evaluation and treatment of infection related to intravascular devices?

Infection rates for patients with implantable devices range between 1% and 7%. Early infection often develops within the first month after implantation and is a consequence of intraoperative contamination. Late infection occurs after the first month and may develop as a result of mechanical erosion or bacte- remic seeding. Most infections are caused by staphylococci, including S aureus and coagulase-negative staphylococci, and streptococci.

Risk factors for development of AICD- and pacemaker-associated infections include prolonged operation, reoperation, generator replacement, catheter-related bacteremia, and sternal wound infection, as well as diabetes mellitus, malignancy, and immunosuppression.³¹ In a recent prospective study of 45 patients with pacemaker-associated IE, 3 infection scenarios with almost equal distribution were encountered: infection exclusively localized on pacemaker leads, the combination of a pacemaker-lead infection and a valvular infection, and an isolated valvular infection that was apparently independent of pacemaker leads.³²

Complete removal of the device is strongly recommended unless the infection is clearly limited to a single component-usually the generator unit and the immediately adjacent leads. The availability of laser sheath technology has facilitated nonoperative removal of intravascular devices. Chua and associates³³ evaluated therapeutic modalities in 123 patients who had pacemaker-associated infections. One hundred seventeen underwent complete device removal and antibiotic therapy and 1 of these patients relapsed. Three of 6 patients with incomplete device removal relapsed.

Two sets of blood cultures should be obtained in all patients with device infections. Cultures should also be obtained from the device pocket and from materials excised during surgery. If evidence of endocarditis is present on TEE, at least 4 to 6 weeks of intravenous antibiotic therapy are recommended. A shorter course of antibiotics-10 to 14 days-can be prescribed for patients in whom infection is confined to the generator pocket.

The timing of pacemaker/AICD reimplantation is a topic of debate. To minimize the risk of reinfection from ongoing bacteremia, adequate antibiotic prophylaxis must be provided before replacement of the intracardiac device. In the study by Chua and coworkers,³³ the mean interval between explantation and reimplantation of a new device was 7 days. Interestingly, 18% of patients did not require replacement of the device after removal.

Infections involving left ventricular assist devices are particularly challenging, because device removal is difficult unless a donor organ is available. In the first 3 months after implantation, the infection rate is about 33%. Infections during this time involve the drive-line tract, generator pocket, or pump interior; these infections may lead to sepsis. Limited data are available regarding therapeutic strategies and outcomes for patients with infections that involve left ventricular assist devices.

7. What are the indications for surgery in a patient with IE?

The principal indications for valve repair or replacement are congestive heart failure (CHF) resulting from valve dysfunction (primarily regurgitation); perivalvular extension, including new-onset conduction abnormalities; persistent fever for 10 days or more despite appropriate antibiotic therapy; large vegetation size; embolic phenomena; and infection with fungi, Pseudomonas aeruginosa, or S aureus. Although no trials clearly support the need for surgery in patients with prolonged fever or large vegetation size, both of these factors are associated with an increased incidence of complications, thereby necessitating early surgery.³⁴ Newer antimicrobial agents, such as caspofungin and daptomycin, may result in more effective medical management that will decrease the need for surgery; however, randomized trials that substantiate the potential value of these drugs have yet to be performed.^35,36

CHF secondary to valve dysfunction remains the most common indication for surgery. Valve replacement or repair is required in up to 25% to 30% of patients with IE. The decline in mortality associated with IE over the past 3 decades-from 25% to 30% to 10% to 25%-may, in part, result from the more aggressive use of surgical therapy.³⁷ In general, valve repair should be considered an alternative to valve replacement whenever feasible.

Despite an increase in the number of patients who undergo valve surgery for native valve endocarditis, there is a lack of prospective data on the outcomes for medical versus surgical therapy. Both surgeons and physicians are understandably reluctant to perform surgery on patients who have multiple comorbid illnesses or CNS complications, or on elderly patients. As a result, most of these patients are treated with medical therapy alone. This creates a bias in favor of surgical therapy, which is often performed on younger or healthier patients.

Vikram and associates³⁸ evaluated a subset of 218 patients with complicated left-sided native valve endocarditis; the investigators used propensity analyses in an attempt to control for confounding variables associated with surgery. Valve surgery, compared with medical therapy alone, was associated with significantly reduced mortality in patients with moderate to severe heart failure. However, no difference in mortality was observed in patients with mild or no heart failure.

Indications for surgery in patients with prosthetic valve endocarditis are similar to those in patients with native valve endocarditis, with the addition of valvular dehiscence. Repeated surgery is associated with a lower mortality rate than is medical therapy alone; this is especially true for prosthetic valve surgery in patients with S aureus infection. In a retrospective case series of 33 patients in whom S aureus prosthetic valve endocarditis was diagnosed, the mortality rate was about 75% with medical treatment alone, and 25% with medical treatment plus surgery.³⁹ However, elderly patients and those with CNS complications or multiple comorbid conditions were all treated with antibiotic therapy alone, thereby introducing bias in favor of surgical therapy.

Similar results have been obtained in unpublished data from our institution. Surgical therapy in patients with prosthetic valve endocarditis was independently associated with factors such as age, microorganism, intracardiac abscess, and CHF. Multivariate logistic regression demonstrated that in-hospital mortality was predicted by the presence of brain embolization or S aureus infection.⁴⁰ When surgical treatment was included in the model, there was a nonsignificant trend toward improved survival after surgery.

Predictors of mortality after surgery for IE include older age; preoperative renal failure; New York Heart Association heart disease class IV; and aggressive disease of short duration, most commonly associated with S aureus infection.^41,42

8. What is the appropriate timing for surgery?

Early surgery, before the patient's clinical condition deteriorates, is generally believed to improve prognosis. However, this recommendation is based on retrospective data, with results confounded by selection bias. There are currently no randomized trials that compare outcomes in patients treated with early versus delayed surgery. Appropriate timing for surgery is also affected by the culprit agent: S aureus infection requires earlier surgery than infection with more indolent streptococcal species.⁴³

CHF secondary to valve dysfunction carries a worse prognosis for patients treated with medical therapy and increases operative mortality after valve replacement (6% to 11% in patients without CHF vs 17% to 33% in patients with CHF).⁴⁴ Heart failure secondary to acute aortic regurgitation is particularly poorly tolerated, and it requires emergent surgery.

Duration of antibiotic therapy before surgery does not affect operative mortality, and the incidence of recurrent endocarditis is extremely low even after only a few doses of antibiotics. Therefore, delaying surgery to provide a longer course of antibiotic therapy is inappropriate.⁴⁵

9. What is the appropriate timing of surgery following embolic stroke in patients with IE?

Embolic events that affect the CNS occur in approximately 10% to 20% of patients with IE, with a higher frequency in patients who have mitral valve involvement. Neurologic complications include transient ischemic attack, embolic stroke with or without hemorrhage, ruptured mycotic aneurysms, meningitis, and nonfocal encephalopathy. Most embolic strokes involve the distribution of the middle cerebral artery.⁴⁵

A patient with endocarditis in whom neurologic symptoms develop should undergo contrast CT or MRI of the head to identify the nature and extent of the lesion as well as the presence of hemorrhage. Cerebral angiography is recommended for patients with evidence of intracranial hemorrhage to evaluate for possible neurosurgical intervention before cardiac surgery. Intracranial bleeding develops in about 10% of patients with IE and mycotic aneurysms.^46-48

Concerns about postoperative neurologic complications include the immediate risk of intracranial bleeding during cardiopulmonary bypass and the risk associated with short- and long-term anticoagulation. If there is no hemorrhagic infarction, valve replacement is best performed at least 72 hours after a stroke. In stable patients, a delay of 2 weeks is reasonable. Patients with evidence of hemorrhage on CT are at increased risk for intracranial bleeding during surgery and therefore should be operated on at least 2 to 4 weeks after the neurologic event has occurred.³⁷