Infectious Endocarditis and Infections of Indwelling Devices

Epidemiology

The global burden of disease from IE is largely unknown. Much of the world’s population lives in developing countries, where many people do not have routine access to advanced medical care, and usually no local or national infrastructure exists for disease reporting (see Chapter 2 ). Thus the clinical characterization of IE is biased, shaped by the collective experiences at large teaching facilities in countries where patient access is available and disease reporting is done. However, even in many developed countries, including the United States, IE is not included among the diagnoses requiring mandatory reporting to public health agencies that would define a statewide or national disease incidence or burden.

IE is a heterogeneous syndrome that is heavily influenced by the epidemiology of the infection. For example, in developing countries where rheumatic fever is still endemic, younger adults with longstanding rheumatic heart disease frequently present with a subacute clinical course spanning several weeks that involves left-sided native valve infection caused by viridans group streptococci (VGS). In contrast, in large, teaching, tertiary care centers in developed countries, patients with previous health care exposure frequently present with an acute illness that can be measured in days and is caused primarily by Staphylococcus aureus , with numerous anatomic sites of metastatic foci of infection and worse outcomes. Injection drug use (IDU) as a complication of the opioid epidemic currently active in the United States is another factor contributing to the escalating rate of IE due to S. aureus and has been prominent in rural settings. All types of care centers have provided care for these patients who are often seen in primary care settings initially with subsequent transfer to larger institutions with IE expertise.

The incidence of IE is influenced by multiple host factors that modify the risk of infection. Such factors include the underlying anatomic (usually valvular) cardiac conditions that result in turbulent blood flow and endothelial cell disruption (see later, Pathogenesis). In addition, aging of the population in developed countries has resulted in more patients with myxomatous degeneration of the mitral valve, with subsequent prolapse and insufficiency (see Chapter 76 ). At the same time, a dramatic fall in the incidence of rheumatic fever in developed countries has reduced the overall risk of IE in younger persons. Advances in medicine also alter the incidence of IE. For example, reduced use of tunneled catheters and increasing use of arteriovenous fistulas for chronic hemodialysis will reduce the risk of bloodstream infection and complicating IE. Improvement in oral health in developed countries also may affect the incidence of IE, but this notion remains to be proven.

Population-based studies ^, have been used to estimate both the incidence of IE and its clinical characterization, but complete case ascertainment is difficult to secure. For example, in the United States, patients may receive medical care in locations that are not in their place of residence. Thus large medical centers that have unique team expertise in IE management may be unable to obtain complete case ascertainment in a population because of changing referral patterns or second-party coverage. Data generated from a population-based investigation will have limited applicability (generalizability) if the cohort under study is not representative of other populations in demographic or clinical features.

Incidence studies of IE are limited in number and in geographic coverage of populations. Adjusted annual incidence rates reported among more recent surveys from Western Europe and Olmsted County, Minnesota, have varied and have ranged from approximately 3 to 14 cases per 100,000 persons annually. Incidence trends have slowly increased in some countries with an expected increase due to increased IDU in the United States (see later). Historically, a sex predilection has been noted, with males more often affected by IE, and has been, in part, due to IDU, which more frequently is reported among men. This male predominance may be fading; an increasing prevalence of female IDU with IE has been noted in many rural areas in the United States in the current opioid epidemic. In addition, the female incidence had increased with a high level of health care exposure cited as a predisposing condition for the development of IE in a recent analysis. Health care exposure, including both nosocomial and non-nosocomial exposure, has been recognized only recently ^{, ,} as a major contributor to the development of IE. Not only do indwelling central venous catheters and hemodialysis predispose to bloodstream infection, but infection with antimicrobial resistant pathogens is more likely to occur as a consequence of health care–related exposure. The virulence of some of these pathogens, in particular methicillin-resistant S. aureus (MRSA), is notable and is associated with increased mortality in patients with IE.

As alluded to previously in this section, people who inject drugs (PWID) are a unique group at increased risk for IE and the current opioid epidemic has magnified the effect of IDU on IE in the United States. Surgery for drug use–associated IE increased 2.7-fold from 2011 to 2018, with higher rates observed in the East South Central and South Atlantic regions. These patients tend to be young, male, and otherwise healthy, except for having hepatitis C virus infection, which is highly prevalent. The prevalence of human immunodeficiency virus (HIV) infection has been considerably less to date. The contact of these patients with the health care system often is limited to short stays in an emergency department (ED).

Pathogenesis

Two overarching aspects of endocarditis pathogenesis have been identified. Already noted is a primary predilection for development of IE from an underlying valvular or nonvalvular cardiac structural abnormality that results in blood flow turbulence, endothelial disruption, and platelet and fibrin deposition. This lesion, termed nonbacterial thrombotic endocarditis (NBTE), serves as a nidus for subsequent adhesion by bacteria or fungi in the bloodstream. This pathway is thought to account for a majority of cases of IE, most often related to left-sided valvular stenosis or regurgitation. This picture of pathogenesis is mirrored, in many ways, in the animal model of endocarditis that has been used for decades to examine the pathogenesis, treatment, and prevention of IE. The microbiologic and histopathologic findings in infected animals reflect those seen in humans. A second factor is that infection may involve normal valves. Some reservations regarding this pathway of infection seem appropriate, because it is impossible to know if a valve is completely normal, including its endothelial surface, before onset of valve infection. In addition, animals do not develop experimental endocarditis after an intravascular challenge with a relatively large inoculum of virulent organisms, in particular S. aureus , in the absence of a previous disruption of the cardiac endothelial surface. Nevertheless, in vitro endothelial cell cultures studies have demonstrated uptake of organisms by endothelial cells.

The predominance of gram-positive cocci as causing IE deserves additional comment. Advances in molecular biologic techniques have resulted in the ability to define virulence factors that are unique to these organisms. Infectivity studies that have compared “wild-type” parent strains to molecularly “engineered” strains using an experimental IE model have been of critical importance in defining virulence factors among strains of staphylococci, streptococci, and enterococci. Some of these factors serve as “adhesins” and are largely responsible for initial bacterial attachment to an NBTE nidus or to endothelial cells. They also are responsible for the attachment to medical devices, including prosthetic valves and cardiovascular implantable electronic device (CIED) leads. In this regard, biofilm formation occurs with some of these organisms and is important in both native tissue and prosthetic valve infections, in the context of factors responsible for the propagation of IE after initial bacterial attachment.

The findings from these investigations are expected to affect future treatment and prevention of IE. Novel vaccines containing bacterial proteins that function as adhesins and are good immunogens are being examined, for example, and already have proven to be efficacious in the prevention of experimental IE. In this case, the protein (FimA) is expressed by several VGS species in the pathogenesis of IE. In addition, it is conceivable that work focusing on treatment and prevention of dental caries by VGS could have some role in the management and prevention of IE.

Clinical Presentation

Symptoms

The presentation of IE encompasses a broad spectrum of symptoms and is influenced by multiple contributing factors ( Figs. 80.1 and 80.2 ). These factors would include (1) the virulence of the infecting organism and persistence of bacteremia, (2) extent of local tissue destruction of the involved valve(s) and hemodynamic sequelae, (3) perivalvular extension of infection, (4) septic embolization to any organ in the systemic arterial circulation or to the lungs, as in the case of right-sided IE, and (5) the consequences of circulating immune complexes and systemic immunopathologic factors.

The diverse potential symptoms associated with IE are listed in Table 80.1 . The frequency of symptoms has been approximated from numerous clinical series in both the older and more contemporary literature. Fever (>38°C) is the most common presenting symptom, in up to 95% of patients, but may be absent in up to 20% of cases, particularly in elderly persons, the immunocompromised, patients treated with previous empiric antibiotic therapy, or patients with CIED infections. ^, Fever defervescence usually occurs within 5 to 7 days of appropriate antibiotic therapy. Persistence of fever may indicate progressive infection with perivalvular extension such as abscess, septic embolization, an extracardiac site of infection (native or prosthetic), infected indwelling catheters or devices, inadequate antibiotic treatment of a resistant organism, or even an adverse reaction to the antibiotic therapy itself.

Other nonspecific constitutional symptoms of infection, such as chills, sweats, cough, headache, malaise, nausea, myalgias, and arthralgias, are less common accompanying symptoms and may be noted in approximately 20% to 40% of patients. In more protracted subacute cases of IE, symptoms and signs such as anorexia, weight loss, weakness, arthralgias, and abdominal pain may also occur in 5% to 30% of patients, misleading the clinician to pursue incorrect diagnoses such as malignancy, connective tissue disease, or other chronic infection or systemic inflammatory disorders.

Dyspnea is important to recognize because they may indicate a severe hemodynamic lesion, usually left-sided valvular regurgitation. Associated symptoms of orthopnea and paroxysmal nocturnal dyspnea herald the onset of heart failure (HF). Early recognition of HF symptoms is imperative because it is the most common complication of IE, has the greatest impact on prognosis, is the most frequent indication for surgical intervention, and is the most important predictor of poor outcome with surgical therapy for IE. HF complicates the course of approximately 30% to 50% of patients with IE, ^{, , ,} and even with early surgical intervention, still doubles in-hospital mortality to almost 25%.

A variety of chest pain syndromes can accompany IE. Pleuritic chest pain may result from septic pulmonary embolization and infarction complicating tricuspid IE. Much less common is angina pectoris related to embolization of vegetation fragments into the coronary circulation, which complicates IE in approximately 1% of the cases. Musculoskeletal chest symptoms related to systemic infection or superimposed infectious pneumonitis also would be in the differential diagnosis.

Physical Examination (see Chapter 13 )

Potential findings on physical examination are delineated in Table 80.2 . These data are approximated from both older and more recently reported clinical series. ^{, , ,} A definite murmur is audible in at least 80% of patients on presentation, particularly with left-sided IE. In the large ICE-PCS collaboration, the murmur was new in almost 50% of the patients. The same cohort study found that worsening of a preexisting murmur occurred in 20% of cases. The presence of a new heart murmur also is noted more frequently in patients with IE complicated by HF, and an S ₃ gallop and pulmonary rales would further substantiate this diagnosis. Murmurs are detected in less than half of patients with IE complicating an implanted cardiac device and are infrequently heard in patients with right-sided IE. Heart murmurs associated with acute IE complicated by extensive left-sided valvular destruction with acute, severe regurgitation may also be deceptively unimpressive because of rapid equalization of pressures between the chambers that diminishes the substrate for turbulent flow. Precipitous HF, pulmonary edema, and cardiogenic shock are most often associated with severe acute aortic regurgitation associated with IE, less so by severe acute mitral regurgitation. Severe tricuspid regurgitation, even as an acute complication of IE, is much better tolerated.

A central neurologic abnormality can often be identified, and focal deficits consistent with stroke may be detected in 10% to 20% of patients (see Chapter 45 ). ^, In subacute, indolent IE, an acute stroke typically is the event that prompts the patient to seek medical attention. Most frequently, the stroke is cardioembolic in nature but may infrequently result from complications of intracranial cerebrovascular mycotic aneurysm, such as hemorrhagic rupture. Seizures, visual deficits, cranial nerve deficits, subarachnoid hemorrhage, and toxic encephalopathy are other potential neurologic complications of IE. The development of neurologic deterioration during the course of IE is associated with significantly increased mortality.

Abdominal examination may elicit nonspecific findings of tenderness and discomfort, particularly in the left upper quadrant, suggestive of splenic embolization and infarction, particularly if complicated by splenic abscess. The spleen is a common site of septic embolization. This most often is not identified by localized symptoms or findings but is discovered incidentally on computed tomography (CT) or using other imaging techniques. Splenomegaly usually is seen in subacute IE and is reported in approximately 10% of patients in more recent clinical series in which the diagnosis is established earlier in the course of the disease. ^{, ,}

As a result of advances leading to earlier diagnosis and therapy, the classic peripheral manifestations of IE are now infrequently observed. Petechiae are the most common, occurring on the conjunctivae, oral mucosa, or extremities. Janeway lesions are painless hemorrhagic macules with a predilection for the soles or palms and are sequelae of peripheral septic embolization, most often associated with staphylococcal IE. Splinter subungual hemorrhages also are painless, dark-red linear lesions in the proximal nailbed and may coalesce. Brown distal splinter lesions at the tips of the nails are quite common in patients who perform manual labor and are caused by trauma, not infection. Osler nodes are painful, erythematous, nodular lesions usually located in the pads of the fingers and toes and are the result of immune complex deposition and focal vasculitis. Roth spots are retinal hemorrhages with a pale center of coagulated fibrin and also are related to immune complex–mediated vasculitis secondary to IE. An immune complex–mediated diffuse glomerulonephritis rarely may be associated with these findings. Both Osler nodes and Roth spots can be observed with other disorders, such as systemic lupus erythematosus (SLE), leukemia, and nonbacterial endocarditis. Aside from petechiae and conjunctival hemorrhage, these peripheral findings were detected in less than 10% of patients in the recent ICE-PCS cohort. A recent multicenter prospective cohort study of 1804 patients with IE confirmed similar results as pertains to the spectrum of clinical presentations and physical examination findings.

Diagnosis

The protean clinical presentations and manifestations of IE encompass a broad differential diagnosis in the patient presenting with fever without a readily apparent cause. Other primary cardiac diagnoses that may potentially mimic IE include acute rheumatic fever, left atrial (LA) myxoma, antiphospholipid antibody syndrome, and nonbacterial thrombotic or marantic endocarditis. A number of connective tissue disorders, including SLE, reactive arthritis, polymyalgia rheumatica, and vasculitides, may be additional diagnostic considerations in select patients, as well as many other serious syndromes of infectious disease. The index of suspicion for IE incrementally increases in the presence of predisposing cardiac conditions, new or changing murmurs, bloodstream infection, clinical evidence of embolic phenomena, and evolving HF or certain other hemodynamic abnormalities.

In 1994, Durack and associates proposed diagnostic criteria, subsequently known as the Duke criteria, to establish the diagnosis of definite or possible IE, and also to reject the diagnosis of IE. These criteria incorporated direct histopathologic evidence of IE or major clinical criteria, namely, blood culture positivity and evidence of endocardial involvement, supplemented by minor clinical criteria, for the definite diagnosis of IE. Thereafter, multiple clinical series using the Duke criteria in the diagnosis of IE reported the sensitivity to be in the range of 80%, with both specificity and negative predictive value (NPV) exceeding 90%. ^, Recognizing the increasing impact of S. aureus IE, the potential for IE associated with C. burnetii infection, and the evolving role of transesophageal echocardiography (TEE) in the diagnosis of IE, Li and colleagues proposed the modified Duke criteria ( Table 80.3 ). Major clinical criteria include (1) blood culture positivity for bacteria typically associated with IE, or persistently positive cultures for organisms uncommonly associated with IE, or a blood culture or serology clearly positive for C. burnetii and (2) evidence of endocardial involvement by echocardiography demonstrating vegetation, significantly new valvular regurgitation, dehiscence of a prosthetic valve, or findings consistent with perivalvular extension of infection, such as abscess. Minor clinical criteria include (1) predisposing cardiac conditions or IV drug use; (2) persistent fever with temperatures greater than 38°C without an alternative explanation; (3) vascular phenomena such as systemic or pulmonary embolism, mycotic aneurysm, or intracranial or cutaneous hemorrhagic lesions; (4) immunologic phenomena such as Osler nodes, Roth spots, or glomerulonephritis; and (5) positive blood culture status not meeting major criteria or serologic evidence of active infection with an organism that could be associated with IE. By this diagnostic classification, a definite clinical diagnosis of IE is established in the presence of (a) two major criteria or (b) one major and three minor criteria, or (c) five minor criteria. A possible clinical diagnosis of IE is appropriate in the presence of (a) one major and one minor criterion or (b) three minor criteria. The diagnosis of IE is rejected if clinical evaluation (a) does not meet criteria for possible IE or (b) reveals complete resolution of a suspected IE syndrome or absence of anatomic evidence for IE on a course of antibiotic therapy for 4 days or less, or if (c) an alternative diagnosis explaining the initial presentation is confirmed.

Since their publication in 2000, the modified Duke criteria have been validated in subsequent investigations of diagnostic accuracy (confirmed to be high) and also clinical and epidemiologic utility and have been endorsed by guideline documents pertinent to the evaluation and management of the patient with IE. ^, In view of the vast heterogeneity of clinical presentations of IE, the modified Duke criteria must always be used in combination with circumspect clinical judgment.

Electrocardiogram

The 12-lead electrocardiogram (ECG) usually demonstrates nonspecific findings in patients with uncomplicated IE ( Fig. 80.3 ) . Because of the close proximity of the atrioventricular node and proximal intraventricular conduction system to the aortic valve and root, perivalvular extension of infection from this location is the most common cause of new atrioventricular block (AVB) of any degree or bundle branch block (BBB). With perivalvular extension of infection, the incidence of AVB ranges from 10% to 20%, whereas new BBB occurs in approximately 3%. ^, The occurrence of a new conduction abnormality also is a multivariate risk predictor for death associated with IE. In a minority of patients, perivalvular extension complicating aortic valve IE may compromise proximal coronary artery patency, or emboli from aortic valve vegetations may cause damage, resulting in ischemic ECG changes or even ST-segment elevation acute coronary syndromes. Ischemic ECG changes may also manifest secondary to hemodynamic sequelae resulting in coronary arterial demand supply mismatch. Other atrial and ventricular arrhythmias may potentially complicate structural or hemodynamic complications of IE, especially in the context of underlying LV dysfunction. In a recent investigation that included 507 patients with left-sided native valve IE, new-onset atrial fibrillation was independently associated with HF and in-hospital mortality.