Infective Endocarditis
Jeffrey A. Dixson
Jeffrey Gaca
Todd L. Kiefer
INTRODUCTION
Since its early descriptions by Osler in 1885, infective endocarditis (IE) has been recognized as a “vexing” problem, challenging to recognize, and associated with high morbidity and mortality. Beginning in the 1940s, two major advances—the development of penicillin, and in subsequent decades, advances in cardiac surgical therapy—have substantially reduced mortality from endocarditis. Early reports of endocarditis therapy for susceptible organisms using various forms and routes of penicillin with heparin for 10 to 62 days of therapy resulted in 70% cure rates. In 1965, Wallace et al. offered one of the first descriptions of surgery for antibiotic refractory endocarditis, ushering in the modern era of surgical management for complicated endocarditis.
Shifting Epidemiology
Several clinically important shifts in IE epidemiology warrant attention. In developed countries, the proportion of IE cases related to rheumatic heart disease has decreased, whereas cases related to indwelling vascular access, cardiac implantable electronic devices (CIEDs), prosthetic valves, other prosthetic implants, and health care contact have increased.1 Prosthetic valve infections now account for up to 20% of IE cases, and CIED infections comprise 4% to 8% of cases. As many as 25% of cases in one series had recent health care exposure.2 Patients with comorbidities requiring indwelling catheters, immunosuppression, or frequent health care contact (specifically human immunodeficiency virus [HIV], cancer, hemodialysis, and older patients) are at increased risk for IE as a result of immunosuppression and frequent transient bacteremia. Patients who use intravenous (IV) drugs similarly represent a growing proportion of endocarditis cases, particularly in the United States with the ongoing opioid epidemic. A recent analysis of claims data in Pennsylvania showed a 20% increase in the overall incidence of endocarditis between 2013 and 2017, but a 238% increase in endocarditis related to intravenous drug use (IVDU).3
Concurrent with the shift in risk factors toward an older population with more comorbidities and indwelling prosthetic implants has been a shift in microbiology toward more virulent organisms. Several studies have found an increase in the percentage of cases due to Staphylococcus aureus, both in the United States and in Europe, making it the most common microorganism isolated.1,4 In a nationwide inpatient cohort in the United States, a steady increase in the percentage of endocarditis cases due to S. aureus was reported: from 33% in 2000 to 40% in 2011.1 The same study reported increases in Streptococcus, fungal, and gram-negative endocarditis over the study period, with an increase in the percentage of cases due to Streptococcus after 2007 IE guidelines restricted the populations for whom endocarditis prophylaxis was recommended. Others have questioned the finding of increased Streptococcus endocarditis after guidelines were updated.5
Incidence
Overall, the incidence of IE is at least stable, if not increasing, and is generally reported between 3 and 10 cases per 100,000 population.6 Among certain subgroups, however, the incidence is higher. The incidence in areas with high IVDU has been reported at 16 per 100,0001 and among older individuals 20 per 100,000.7 In a U.S. hospital database, rates of hospitalization for endocarditis increased from 11 per 100,000 to 15 per 100,000 between 2001 and 2012.1 Increasing incidence is generally attributed to high rates of IVDU, increasing use of implantable cardiac devices and prosthetic valves, higher rates of indwelling or repeated vascular access for chemotherapy and hemodialysis, and an aging population with more immunocompromising comorbidities.1 A male predominance is consistently observed, with an estimated ratio of 1.2 to 2.7:1 in comparison to females.
With respect to the sites of valve involvement, the most common valve affected is the aortic valve (38% of cases).8 This is followed closely by the mitral valve, which is involved in 34% of cases. Isolated tricuspid valve endocarditis is less common, but is linked to IVDU, chronic hemodialysis, or the presence of pacemaker or implantable cardioverter defibrillator (ICD) leads.9
PATHOGENESIS
IE develops when a platelet-fibrin matrix deposits at the site of endothelial disruption in the heart or great vessels, providing a suitable milieu for bacteria or fungal attachment and reproduction. Complications result from primary tissue destruction, local extension of the infection, embolization of vegetative material, or immune complex activation and deposition. A variety of microorganism-specific factors interact with the host substrate to determine the clinical manifestations of the infection. Understanding these processes helps explain several clinical entities associated with IE.
Intact valvular endothelium effectively resists bacterial adhesion and colonization despite experimental intravascular inoculation.10 Endothelial disruption, therefore, is a necessary precursor to IE, and may result from a variety of mechanisms: systemic disease (eg, rheumatic carditis), turbulent blood flow (eg, regurgitant valve lesions), mechanical injury (eg, from catheters or electrodes), or repeated insult from foreign material (eg, from particles injected during IV drug use).11 Vegetations classically form in the line of closure of valve leaflets, on the “low-flow” side of the valves (ie, atrial side of atrioventricular valves and ventricular side of the semilunar valves).12 Endothelial damage promotes sterile platelet-fibrin deposition and interstitial edema at the damaged site, thereby providing the substrate for capable bacteria to colonize an otherwise sterile thrombus.11
Although adhesion to platelet-fibrin matrix on damaged endothelium appears to be held in common between microorganisms causing IE, the mechanisms vary considerably between species. S. aureus may bind directly to damaged endothelial cells via species-specific clumping factors and coagulase interactions with fibrinogen. S. aureus also binds and activates platelets via the von Willebrand receptor. Other organisms bind to components of damaged endothelium or platelet-fibrin matrix, such as fibronectin, laminin, or collagen.13,14 Streptococcal species may use dextran and other virulence factors to adhere to the platelet-fibrin matrix.15
Once microorganisms establish residence in the sterile thrombus, complex interactions determine the subsequent progression of the vegetation. Bacteria proliferate and promote additional platelet aggregation and activation. Species such as S. aureus resist the inherent bactericidal activity of platelet microbicidal proteins and interact with monocytes to promote release of tissue factor and the procoagulant cascade that facilitate bacterial and vegetation proliferation, respectively.11 Deep inside a vegetation, some bacteria exhibit reduced metabolic activity that promotes survival against some antibiotics. S. aureus—and methicillin-resistant Staphylococcus aureus (MRSA) in particular—produce a biofilm, although the clinical relevance of this has been questioned.11 IE related to CIEDs has been more directly linked to biofilm formation and contributes to the clinical challenges managing these infections.
Transient bacteremia is a necessary, but not sufficient, event to promote endocarditis. Although bacteremia is common after mild mucosal disruption following dental, gynecologic, urologic, or gastrointestinal procedures,16 this rarely leads to the development of IE. The minimal magnitude of bacteremia necessary to cause IE is unknown, but it is likely that values above 104 colony-forming units (CFUs) per milliliter of blood are required.16,17 In addition, bacteria-specific factors help determine whether transient bacteremia establishes IE. Gram-negative pathogens are usually eliminated by complement-mediated bactericidal activity in the blood, protecting against IE from gastrointestinal sources of bacteremia.11 Organisms well suited to adhere to valve leaflets, on the other hand, such as Staphylococcus, Streptococcus species, and Pseudomonas, are more likely to cause IE.
The immunologic response of the host accounts for several clinical manifestations of IE. In response to infection, a variety of circulating antibodies are produced, including rheumatoid factor in up to half of patients with IE for more than 6 weeks, and antinuclear antibodies that may contribute to musculoskeletal manifestations, fever, and pleurisy.18,19 Circulating immune complexes can be found in high titers in patients with IE, and are implicated in the development of IE-associated glomerulonephritis, tenosynovitis, Osler nodes, splinter hemorrhages, Roth spots, and mycotic aneurysms of blood vessels.20
Complications are the rule rather than the exception in IE, and can affect nearly any organ via a variety of mechanisms. Large emboli can occlude vessels in any downstream organ. Most notably, cerebral emboli are clinically apparent in 20% to 30% of patients with IE and present in the majority of asymptomatic patients undergoing magnetic resonance imaging (MRI), a risk that declines precipitously with appropriate antibiotic therapy.21,22,23 Splenic infarcts are common, and myocardial infarction can occur with aortic valve vegetations, in particular.24 If the emboli are infected, bacteria can directly invade adjacent tissues, leading to abscess formation (commonly in the kidney and spleen) or mycotic aneurysm (commonly at the branch points of cerebral vessels). Bacterial seeding in the absence of emboli can similarly lead to tissue infection and abscess formation. Within the heart, infections can lead to perforation of a leaflet, rupture of chordae tendinae, erosion of the interventricular septum, valve ring abscess with fistulae, myocardial infarction, and conduction system block. Right-sided endocarditis commonly leads to septic pulmonary emboli, which in turn are associated with pneumonia, abscesses, pleural effusions, and empyema. In the skin, immune-mediated perivascular inflammation can cause Osler nodes, and septic emboli lead to Janeway lesions. Ocular manifestations include immune-mediated micro hemorrhages (Roth spots) or, less commonly, direct bacterial seeding causing endophthalmitis.
Microbiology
Microbiology of IE can be divided broadly between native and prosthetic valve infections, with native valve endocarditis (NVE) further characterized according to predisposing heart conditions and comorbidities. Staphylococcus, Streptococcus, or Enterococcus species account for the vast majority ( 80%-90%) of native valve infections.6 S. aureus now accounts for up to 30% of cases in high-income countries and affects patients across risk groups, including those with no identifiable predisposing valve lesions.6 Coagulase-negative staphylococci (eg, S. epidermidis and S. lugdunensis) are ubiquitous skin commensals and are associated with health care-acquired NVE. In low-income countries, Streptococcus viridans species are the most common causes of IE and are common commensal organisms of the oral, gastrointestinal, and genitourinary tracts. Enterococci account for about 10% of cases overall, but particularly affect elderly and immunocompromised hosts.6 Gram-negative (eg, Acinetobacter and Pseudomonas species) and HACEK organisms (Haemophilus, Actinobacillus, Cardiobacterium, Eikenella, and Kingella species) that are slow-growing colonizers of the
oropharynx are rare offenders, accounting for fewer than 3% of cases.6 Other rare causes of IE with particular exposures include zoonotic infections such as Brucella (livestock), Bartonella (cats), Chlamydia (birds), and fungal infections (esp. Candida and Aspergillus) after cardiac surgery and in immunocompromised hosts.6 Viral endocarditis is essentially nonexistent.12 Endocarditis associated with IVDU (IVDU-IE) can be either right- or left-sided, and the vast majority (>80%) of IVDU-associated tricuspid valve infections are caused by S. aureus.8
oropharynx are rare offenders, accounting for fewer than 3% of cases.6 Other rare causes of IE with particular exposures include zoonotic infections such as Brucella (livestock), Bartonella (cats), Chlamydia (birds), and fungal infections (esp. Candida and Aspergillus) after cardiac surgery and in immunocompromised hosts.6 Viral endocarditis is essentially nonexistent.12 Endocarditis associated with IVDU (IVDU-IE) can be either right- or left-sided, and the vast majority (>80%) of IVDU-associated tricuspid valve infections are caused by S. aureus.8
The microbiology of prosthetic valve endocarditis (PVE) is characterized by the time interval since valve surgery. Early infections (<2 months from the time of surgery) are usually due to nosocomial infections with coagulase-negative Staphylococci or S. aureus.6 Late infections (>12 months after surgery) have a microbiology similar to that of NVE, with gram-positive Staphylococci, Streptococci, and Enterococci species predominating. Mid-term IE (2-12 months after surgery) is a mixture of the nosocomial infections and the usual native valve organisms. Although rare, fungal infections carry a high mortality and predominantly affect prosthetic valves.6 Intracardiac device-related IE (CIED-IE) occurs in up to 2% of implanted devices, and can occur at the device pocket, on the leads, or in adjacent areas of disturbed endocardium.6 Staphylococcal skin commensal organisms predominate as pathogens in CIED-IE, including coagulase-negative species and S. aureus.25
Risk Factors
IE develops as a result of transient bacteremia in the setting of endothelial damage and platelet-fibrin matrix deposition. Therefore, conditions that increase transient bacteremia or are associated with endothelial disruption predispose to endocarditis.
Historically, the most common predisposing cardiac condition was rheumatic heart disease. Although this remains true in developing countries,26 rheumatic disease accounts for less than 5% of IE cases in developed countries in the modern era.27
In developed countries, the dramatic increase in implantable cardiac devices—including prosthetic valves, pacemakers, and defibrillators—now makes prosthetic device implantation one of the most important risk factors for IE. A recent report of more than 2000 IE cases from 25 countries found that 20% had a prosthetic valve and 7% had an implantable cardiac device.27 Mitral valve prolapse (MVP) is the most common predisposing native valve abnormality, accounting for between 8% and 30% of NVE cases,12 and has been reported to carry an odds ratio (OR) for IE of 8.2 (95% CI 2.4-38.4)28 in comparison to individuals without MVP. As the population ages, degenerative valve lesions also become important risk factors. In autopsy studies, mitral annular calcification has been found more commonly among those with endocarditis than among the general population, suggesting this may represent an important substrate for the development of IE.29 Finally, as more children with congenital heart disease reach adulthood—often after multiple surgeries and some with implantable devices—they are at increased lifetime risk for developing IE.30
Factors that increase the risk of transient bacteremia or alter the immune response also increase the risk of IE. In the United States, the proliferation of IVDU has contributed particularly to the development of tricuspid valve endocarditis. This is likely the result of increases in transient bacteremia and endothelial damage caused by injection of solid particle contaminants.11 Increased frequency of health care contact—itself associated with more indwelling vascular access catheters, repeated procedures, and exposure to virulent organisms—is also an increasingly important risk factor for IE. A recent, prospective, multinational series of NVE found that health care exposure accounted for up to 30% of cases.31 Some of the associated risk of health care exposure may also be related to the underlying disease state and altered immune response, as can be the case for HIV, malignancy, and patients on hemodialysis.32 Finally, although endocarditis was formerly a disease of young patients with rheumatic heart disease, the highest risk in modern cohorts appears to be among older patients (aged 58-77 years) with other predisposing comorbidities.32
Clinical Outcomes
Despite advances in diagnosis and management, mortality remains high. In-hospital mortality in the contemporary era is 20%,33 although it varies substantially by age, risk factors, and microorganism. For example, right-sided endocarditis is generally reported to have less than 10% early mortality;34 adults with congenital heart disease were recently reported to have a similarly low early mortality of 6.9%.35 In contrast, 30-day mortality was 27.5% in patients on dialysis and 27.7% among older (58-77 years old) patients in a single-center study.36 In the International Collaboration on Endocarditis Prospective Cohort Study, the strongest risk factor for in-hospital mortality is paravalvular complications (OR 2.25, 95% CI: 1.64-3.09), followed by pulmonary edema (OR 1.79, 1.39-2.3), S. aureus infection (OR 1.54, 1.14-2.08), coagulase-negative Staphylococcal infection (OR 1.50, 1.07-2.10), prosthetic valve involvement (OR 1.47, 1.13-1.90), mitral valve vegetation (OR 1.34, 1.06-1.68), and increasing age (OR 1.30, 1.17-1.46 per 10-year interval).27 S. viridans infection (OR 0.52, 0.33-0.81) and surgery (OR 0.61, 0.44-0.83) were associated with lower mortality.27
Although short-term mortality from endocarditis remains around 20%, longer term (up to 1 year) mortality is substantially higher at 30% to 40%.37 This too, however, varies substantially by age and comorbidities, suggesting that some of the late mortality reflects the substrate more than the disease. For instance, a small study of octogenarians with IE reported 1-year mortality of 37.3% compared to 13% among those younger than 65 years of age,38 and 6-month mortality among patients on dialysis was 38.5% versus 29.2% in the cohort overall in a single tertiary-care teaching hospital.36
DIAGNOSIS
Because the clinical manifestations of IE are most commonly vague and nonspecific and many patients lack the classic textbook findings, combined clinical, microbiologic, imaging, and
pathologic criteria as defined in the modified Duke criteria remain the cornerstone of diagnosis (Table 14.1). After an initial set of investigations, a diagnosis of IE is characterized as definite, possible, or rejected.39 For a definite diagnosis, two major, one major, and three minor, or five minor clinical criteria must be met. Alternatively, pathologic criteria may be met by culture or histologic examination of in situ valve tissue, embolized vegetations, or an intracardiac abscess. The diagnosis is rejected when there is a clear alternative diagnosis, the IE syndrome resolves with less than or equal to 4 days of antibiotic therapy, there is no pathologic evidence of IE after surgery or autopsy despite less than or equal to 4 days of antibiotic therapy, or criteria for possible IE are not met. Possible endocarditis is defined as the presence of one major plus one minor criterion or three minor criteria.
pathologic criteria as defined in the modified Duke criteria remain the cornerstone of diagnosis (Table 14.1). After an initial set of investigations, a diagnosis of IE is characterized as definite, possible, or rejected.39 For a definite diagnosis, two major, one major, and three minor, or five minor clinical criteria must be met. Alternatively, pathologic criteria may be met by culture or histologic examination of in situ valve tissue, embolized vegetations, or an intracardiac abscess. The diagnosis is rejected when there is a clear alternative diagnosis, the IE syndrome resolves with less than or equal to 4 days of antibiotic therapy, there is no pathologic evidence of IE after surgery or autopsy despite less than or equal to 4 days of antibiotic therapy, or criteria for possible IE are not met. Possible endocarditis is defined as the presence of one major plus one minor criterion or three minor criteria.
TABLE 14.1 Summary of Modified Duke Criteria for Infective Endocarditis | |||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |||||||||||||||||||||||||||||||||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
