Overall Incidence

The true incidence of IE is difficult to ascertain. In a Swedish urban setting, Hogevik et al ⁶ found an incidence of 5.9 episodes per 100,000 person-years from 1984 to 1988. During a similar period in a Philadelphia metropolitan study, the total incidence was calculated to be 9.29 episodes per 100,000 person-years. ⁷ When intravenous drug users were excluded, this incidence fell to 5.02 episodes per 100,000 person-years. In both urban and rural settings in France, the incidence was estimated to be around 2.43 episodes per 100,000 person-years in 1991 ⁸ and increased to 3.1 episodes per 100,000 person-years in 1999 ⁹ with a peak incidence of 14.5 episodes per 100,000 person-years in the elderly. The growing incidence in elderly individuals has been confirmed in the Medicare population in the United States, in which it was 20.4 episodes per 100,000 person-years in 1998 (a 13.7% increase from 1986). ¹⁰ In fact, more than half of all cases of IE in the United States and Europe now occur in patients older than 60 years, and the median age of patients has increased steadily during the past 40 years. ¹¹ Health care–associated IE results from health care–associated bacteremias. They include both nosocomial and nonnosocomial infections, have a high mortality rate, and are frequent in patients who are undergoing hemodialysis and/or who have other debilitating diseases. The typical patient nowadays is therefore less likely to be one with poor dentition and rheumatic disease ¹² and more likely to be elderly and to have undergone a procedure to implant a device such as a prosthetic valve, pacemaker, and/or defibrillator ⁵ or to have a major comorbid condition.

Incidence of Infective Endocarditis and Associated Mortality

Table 25-1 summarizes cardiac conditions and the subsequent estimated incidence of IE per 100,000 patient-years. ¹³ Sex and age also influence the incidence of IE, with males predominating. Male : female ratios have been noted to range from 3.2 : 1 to 9 : 1.^11,14 Of interest, 50% to 70% of children younger than 2 years in whom IE develops have no apparent underlying heart disease, whereas older children usually have a congenital heart condition. ¹⁵ Endocarditis in patients with injection drug use (IDU)—defined as the intravenous injection of recreational drugs such as heroin, cocaine, and amphetamines—also may occur when there is no apparent underlying valvular pathologic lesions. ¹⁶ Despite these exceptions, most patients do have identifiable underlying structural heart disease at the time of their endocarditis diagnosis.^17,18 Earlier reports, before 1967, showed that rheumatic heart disease was the most common cardiac abnormality, being present in 39% of patients, ¹⁹ whereas later series suggest its presence in only about 6%. ¹⁸ Estimates of specific valvular lesion involvement is summarized in Table 25-2.

Despite advances in the diagnosis and management of IE, it remains a disease with unacceptably high morbidity and mortality. More than 50% of patients with IE have some type of serious complication, including HF, stroke, and paravalvular extension, whereas the in-hospital mortality rates (15% to 20%) and 1-year mortality rates (30% to 40%) have changed little over the past 20 years.5,9,20–22 Death is still disturbingly frequent and usually relates to cardiogenic shock, multiorgan failure, or stroke. Surgery is necessary and important for survival in around half the cases. ²³

When valvular disease is considered as a whole, endocarditis still is an uncommon disease process. In the Euro Heart Survey ²⁴ of the incidence of valvular disease in a general population, endocarditis was the major diagnosis in less than 1% of patients who were found to have aortic or mitral stenosis, in only 7.5% of those with aortic regurgitation, and in 3.5% of those who had mitral regurgitation.

Left-sided native valve endocarditis (NVE) remains the most common presentation, accounting for 70% of all cases of IE. Mortality depends on comorbidities but still is at least 15% as a whole. ²⁵ Degenerative mitral valve disease (mitral valve prolapse) is the leading predisposing valve lesion, with the risk particularly high in children and in patients older than 50 years. Patients with degenerative aortic valve disease are also at risk, helping explain the rising age of patients presenting with IE. One review estimated a slightly greater incidence of mitral than aortic involvement, 8% involving both, 4% involving the tricuspid valve, and 3.5% occurring in patients with congenital heart disease. ²⁶ Endocarditis is unusual in patients with isolated pulmonary stenosis, atrial septal defect, mitral stenosis, or hypertrophic cardiomyopathy.

Prosthetic valve endocarditis (PVE) accounts for up to 20% of the patients with endocarditis reported in a recent series from the International Collaboration on Endocarditis-Prospective Cohort Study. ²⁷ Staphylococcus aureus was the most common organism. Having a prosthetic heart valve is the greatest risk for development of IE in every series. It is estimated that IE will develop in 1.4% to 3.1% of all patients with prosthetic valve at 1 year and 3% to 5.7% at 5 years. ²⁸ There are two disparate risk periods for the development of PVE, an early period and a late period, although some writers believe it is more useful to consider early (2 months), middle (2 to 12 months), and late (>12 months) periods as the organisms involved shift gradually rather than abruptly. ²⁸ The early period is generally defined as the first 60 days after heart surgery, and most of the implicated organisms are considered nosocomial. The late period involves organisms more like those involved in NVE. Although there has always been a suggestion that the mechanical valves are more susceptible to IE, by 5 years there appears to be no real difference, and most series do not suggest a difference in the risk by model, position, or type of valve (mechanical or bioprosthetic). ²⁹ Some patient factors have been associated with PVE, including renal dysfunction, young age, prior endocarditis, and perioperative wound infections. ³⁰ Health care–associated prosthetic valve endocarditis is identified in 36.5% of all cases, and most infections (71%) occur in the first year after the valve was implanted. The rate of in-hospital death remains high, at 23%, and its occurrence is associated with older age and the complications related to the surgical intervention. The higher risk of “redo” aortic valve replacement (AVR) for endocarditis is emphasized in a report of 313 patients by Leontyev et al. ³¹ Perioperative mortality was 24.3% in “redo” AVR for IE compared with 6.8% for redo AVR for reasons other than endocarditis.

Right-sided endocarditis is seen in about 5% to 10% of IE surveys16,25 and has a better prognosis than left-sided disease, though the mortality remains high in patients with human immunodeficiency virus (HIV). ³² Right-sided endocarditis typically occurs in patients with illicit IDU (including patients with HIV) and those with structural abnormalities of the right heart due to congenital heart disease, pacemaker or defibrillator implantation, or central venous catheters. The most significant risk factor for right-sided IE is certainly IDU; however, left-sided disease may be more common in some groups of addicts. In one series, left-sided involvement occurred in 57% of patients with IDU compared with 40% with right-sided disease. ³³ The most common infecting organism in patients with IDU is S. aureus, where it has been reported as the offending organism in up to 82%. ³⁴

Prognosis in patients with IDU and IE is generally better than that in overall patients with IE who do not have a history of IDU, because of the lower risk of IE on the right side. ³² The patients with IDU and IE generally are also much younger than other IE populations. Of importance, the presence of HIV infection does not appear to alter the diagnostic use of the Duke criteria or the course of the disease, ³⁵ although patients with a very low CD4⁺ count (<200 cells/mm³) are at greater risk. ⁹

The expanded use of cardiovascular electronic devices has resulted in infections not only on the device leads themselves but also on the tricuspid leaflet. A pocket infection appears to predispose to this form of IE. ³⁶ A study from the Medicare database indicates that although the device implantation rate rose 42% in the 1990s, the IE infection rate rose 124%. ³⁶ One estimate of the rate of these device-related infections suggests it is about 0.55 cases per 1,000 implants. ³⁷ Removal of the device is almost always required for cure.^38,39

Groups at High Risk for Development of Infective Endocarditis, and the Determination of Prognosis

A summary of a variety of noncardiac clinical conditions predisposing to IE and the organisms frequently associated with these conditions is given in Table 25-3.

IDU is clearly a risk factor for IE, and those who use cocaine may have the greatest risk. ⁴⁰ A prior history of endocarditis is an important predisposing factor in recurrent IE in patients with IDU. Recurrent endocarditis occurred in 4.5% of one large cohort of nonaddicts who survived their initial episode. ⁴¹

Pacemaker-associated infections have increased with the increased use of electrophysiologic (EP) devices. A report from the Multicenter Electrophysiologic Device Infection Cohort (MEDIC) registry ⁴² from 2009 through 2011 found that early (<6 months after implantation) device infections were generally related to pocket infections and that later IE was the result of other bacteremias. Staphylococci (coagulase-negative, methicillin-resistant, and methicillin-sensitive) were the most common organisms involved. As mentioned, effective treatment almost always requires removal of leads.

Patients undergoing hemodialysis are the largest subgroup with health care–related IE.43,44 Predisposing factors in this population include intravascular access, calcific valvular disease, and impairment of the immune system. Of patients with health care–associated IE who are not undergoing dialysis, most have underlying predisposing conditions, including diabetes, cancer, and immunosuppressive therapy use. Identifiable underlying cardiac predisposition to IE occurs in less than 50% of this group. ⁵ Most invasive organisms originate from the skin or urinary tract, and the presence of intravenous lines or other invasive procedures is frequently evident. ⁴⁵ Staphylococcus is the predominant offender.

Other noncardiac conditions also can predispose to IE and are often associated with specific infecting organisms. For instance, in the International Collaboration on Endocarditis-Prospective Cohort Study, patients with IE due to S. aureus were significantly more likely to be hemodialysis dependent, to have diabetes, to have a presumed intravascular device source, to receive vancomycin, to be infected with methicillin-resistant S. aureus (MRSA), and/or to have persistent bacteremia. ⁴⁶

Nosocomial endocarditis infections are also becoming more common. They are defined as a diagnosis of IE made more than 72 hours after admission in patients with no evidence of IE on admission or as development of IE within 60 days of a prior hospital admission during which there was risk for bacteremia or IE. ⁴⁷ Nosocomial IE is usually a complication of bacteremia caused by an invasive intravascular procedure or an intravenous catheter–related device infection.^47,48 It accounts for almost 10% of cases of IE in some series.

A number of cases of IE have been reported in patients with HIV infection.49,50 Some valves have been infected with unusual organisms such as Salmonella and Listeria. ⁵⁰ It has been reported that HIV infection is an independent risk factor for IE in injection drug users, ⁵¹ although this finding has not been confirmed in other studies. ⁵⁰

The clinical examination, the organism involved and its response to therapy, and the echocardiographic information can establish the prognosis and guide decision making in the treatment of IE. A number of studies have examined other factors in an effort to understand prognosis in patients with IE. Chu et al ⁵² examined 267 consecutive patients with acute IE to determine factors early in the course of the disease that were independently associated with mortality. After controlling data for severity of illness with Acute Physiology and Chronic Health Evaluation (APACHE) II scoring, they found that the independent predictors of early mortality were the presence of diabetes mellitus (odds ratio [OR], 2.48; 95% confidence interval [CI], 1.24-4.96), S. aureus infection (OR, 2.06; 95% CI, 1.01-4.20), and an embolic event (OR, 2.79; 95% CI, 1.15-6.80). In a similar fashion, Hasbun et al ⁵³ found that five baseline features were independently associated with mortality and developed a scoring system that included the following: mental status, lethargy or disorientation (4 points); Charlson comorbidity scale, 2 or greater (3 points); HF, moderate to severe (3 points); microbiology, S. aureus (6 points), other non-viridans infection (8 points); and therapy, medical therapy only (5 points). On the basis of this point system, patients with a score of 6 points or less only had 6% mortality at 6 months, whereas patients with a score of more than 15 points had 63% mortality. In other studies, the need for hemodialysis has also been found to portend a poor outcome,^43,54 as has the presence of poor ventricular function. ⁵⁵ Another study identified patients with an altered mental state, those with mobile vegetations, and those undergoing hemodialysis as the cohort with the highest risk. ⁵⁶

Clinical Manifestations

The prevalences of the clinical features observed in patients with IE are summarized in Table 25-6. Many of these features were espoused by Osler in the Gulstonian lectures^4,87 but are rarely seen in an era when diagnostic testing is better and there is antimicrobial therapy. Most patients continue to have an initial indolent course from 2 weeks to many months with vague symptoms. Symptoms include fever, chills, anorexia, weight loss, night sweats, and malaise.

Fever is the most common symptom, occurring in from 64% to 93% of patients with NVE, 85% with PVE, and 75% to 88% of patients with IDU and IE. It is less common in elderly patients and in patients with HF, renal failure, severe debility, or previous antibiotic therapy. ⁸⁸ Persistent fever more than 1 week after therapy requires further investigation into its cause (e.g., an abscess somewhere), a nosocomial infection, drug fever, or inadequate IE therapy.

A murmur is apparent in 80% to 85% of patients, ²⁸ although auscultation is a dying art in cardiology and a murmur may not always be recognized by health care providers even when present. The murmur of acute and fulminant aortic regurgitation may be particularly difficult to hear because there is little diastolic gradient. Whereas tricuspid regurgitation should be evident from examination of the jugular venous pulse, the murmur is often quite soft when right ventricular systolic pressure is normal.

A variety of peripheral cutaneous manifestations highlight the classic endocarditis examination ( Figure 25-2). Unfortunately many of these peripheral stigmata are rare today. Emboli can be observed in many areas, such as mouth and conjunctival petechiae, nail bed splinters, skin Janeway lesions, Osler nodes, and Roth spots. Splinter hemorrhages tend to occur in the proximal half of the nail bed, as opposed to splinters due to trauma, which occur in the outer half. Janeway lesions are painless, erythematous skin lesions that often appear in crops on the hands or feet. They represent embolic events similar to those in splinters. Biopsies reveal that they are microabscesses without arteritis, and organisms can often be cultured from them.

Osler nodes are painful lesions that manifest as nodules on the pads of the toes or the fingertips and may persist for days. The cause for Osler nodes is unclear, but the fact that they may be seen in other settings, such as systemic lupus erythematosus (and there is histologic evidence for perivasculitis on biopsy), has led many to regard them as immunologic phenomena. Rarely, organisms have been cultured from Osler nodes, suggesting that emboli may at least be the inciting mechanism. ⁸⁹

Roth spots are retinal hemorrhages with a white center. They most likely represent septic emboli, but like Osler nodes, they have been described in other clinical settings and especially as a manifestation of systemic lupus erythematosus, anemia, diabetes, multiple myeloma, and HIV infection. ⁹⁰ Frank retinal artery occlusion may also occur.

Musculoskeletal aches and pains are common in IE and often occur early in the course. ⁹¹ Any joint can be involved, but back and shoulder pains are most frequently cited. ⁹² Septic emboli may result in osteomyelitis or bone abscess formation (especially in the spine). Musculoskeletal pain must be taken seriously if it persists during the course of therapy.

Neurologic symptoms are common, being seen in as many as 30% to 50% of patients with IE. Symptoms appear to be more common in patients with IDU and those with staphylococcal IE. ⁹³ Embolic stroke is the most common and serious manifestation. Intracranial hemorrhage may occur from a ruptured arterial vessel, a ruptured mycotic aneurysm, or bleeding into a thrombotic stroke distribution. ⁹⁴ In addition, neurologic symptoms may be related to cerebritis or meningitis or to toxic or immune-mediated injury. Brain abscess is rare, but microabscesses from virulent organisms, such as S. aureus, occur with some frequency. ⁹⁵ Meningitis may be a major feature in IE due to Streptococcus pneumoniae.

Splenic emboli are probably underreported. In the preantibiotic era, splenomegaly was common. Now about 25% to 50% of patients may have evidence of enlarged spleen. Autopsy series suggest that splenic infarcts are often present without clinical symptoms. ⁹⁶

The Evolution of the Duke Criteria

IE has traditionally been defined as an infection of the valves and chordae within the cardiac chambers. This definition has now been expanded to include infection on any structure within the heart, including normal endothelial surfaces (e.g., myocardium and valvular structures), prosthetic heart valves (e.g., mechanical or bioprosthetic valves, homografts, and autografts), and implanted devices (e.g., pacemakers, implantable defibrillators, and ventricular assist devices). Much of the increase in the incidence of IE described over the last couple of decades has come from infections on intracardiac devices, and the number of devices being implanted has grown dramatically.

The diagnosis of IE hinges on clinical suspicion and the demonstration of continuous bacteremia. It was not until the late 1970s that Pelletier and Petersdorf ⁹⁷ developed a case definition based on a 30-year experience of caring for patients with IE in Seattle. Although this definition was highly specific, it lacked sufficient sensitivity. In 1981, von Reyn et al ⁹⁸ published an analysis that provided four diagnostic categories for cases of suspected IE (rejected, possible, probable, and definite), and the effort improved both the sensitivity and the specificity of the previous case definition. The definition did not incorporate imaging information, however.

In 1994, Durack et al ⁹⁹ from Duke University Medical Center incorporated echocardiography into the criteria for the first time, giving rise to what have come to be known as the Duke criteria. These criteria have been validated subsequently by many other studies,^100–102 including the latest modifications ( Table 25-7).^103,104 There are now three diagnostic categories. Definite endocarditis is considered to be present if there is pathologic evidence (surgical pathologic histology or culture or vegetation histology or culture) or if there is clinical evidence as demonstrated by the presence of two major criteria or one major criterion and three minor criteria or five minor criteria. Possible endocarditis is defined as having one major criterion and one minor criterion or three minor criteria. Rejected diagnosis is defined as having a firm alternative diagnosis or sustained resolution of the evidence for endocarditis after 4 or fewer days of antibiotic therapy or no pathologic evidence of endocarditis at surgery or autopsy after 4 or fewer days of therapy.

Stay updated, free articles. Join our Telegram channel

Join