Introduction

Infective endocarditis (IE) is a serious cardiac condition associated with significant morbidity and mortality, including a high rate of surgical treatment. Despite advancements in the diagnosis and treatment of IE, epidemiologic changes such as an increasing rate of health care-associated infection, causation by virulent Staphylococcus aureus, and the co-morbid medical conditions in an older patient population have influenced a persistently high in-hospital mortality rate (approximately 15%–20%) [1•].

The American College of Cardiology (ACC) and American Heart Association (AHA) published guidelines for the management of valvular heart disease, including IE, in 2006. In 2008, a revision of these guidelines focused largely on recommendations for antibiotic prophylaxis for the prevention of IE (Table 1) [2]. These revisions significantly reduced criteria for prophylaxis because of limited outcome data regarding its efficacy and have also generated controversy because of their divergence from previous clinical practice and recommendations.

Table 1 Indications for infective endocarditis prophylaxis
Full size table

In 2009, the European Society of Cardiology (ESC) published its recommendations that further reduced the types of pre-existing cardiac conditions as well as types of procedures for which IE prophylaxis is recommended [3]. In addition, these recent guidelines provided similar recommendations for early surgery during the active phase of infection for patients with heart failure, those with complications (e.g., intracardiac abscess) unlikely to be cured with antibiotic therapy alone, and large vegetations that pose a high risk of embolic complications and mortality (Table 2) [3].

Table 2 Indications for surgical intervention in infective endocarditis and potential reasons for delay or lack of surgery
Full size table

With these guidelines as background, the purpose of this review is to summarize studies of IE published in 2010 and 2011, focusing on those studies related to treatment and outcome of this deadly condition.

Epidemiology and prevention of IE

Changes in the epidemiology of IE have been recently reported. The International Collaboration on Endocarditis (ICE) study group previously reported their multinational, prospective registry data that demonstrated 1) a shift toward acute rather than sub-acute presentation of IE at the time of diagnosis, 2) similar shift toward more virulent causation by S. aureus rather than streptococcal species, influenced by the prevalence of health care-associated infection, and 3) a high rate of surgical intervention during the initial hospitalization, with a high in-hospital mortality rate of 18% [1•]. Lomas et al. [4•] further investigated this effect of health care-associated infection on the epidemiology and outcome of 795 IE cases from a seven-hospital cohort study in Spain. Health care-associated infection was identified in 16% of cases, predominantly in older patients with more co-morbid conditions and staphylococcal infections (58% vs 25% in community-acquired infection) [4•]. Vascular access was the main cause of bacteremia responsible for health care-associated infection, especially peripheral venous catheter infection in one third of these instances. Health care-associated infection was associated with a much higher rate of in-hospital mortality than community-acquired infection (45% vs 24%, P < 0.001) [4•]. In patients with health care-associated infection, the presence of septic shock and the absence of surgical intervention was associated with higher mortality [4•]. Thus, the rate of IE, especially IE due to health care interventions, may be reduced by more careful use of vascular access.

A common, predisposing heart condition for developing native valve IE is bicuspid aortic valve (BAV). A multicenter study from two tertiary centers in France compared 50 patients with BAV to 260 patients with tricuspid aortic valve IE [5•]. Patients with BAV IE were younger, had fewer co-morbidities, and a higher frequency of aortic perivalvular abscess (50%) [5•]. However, the presence of BAV IE was not independently predictive of in-hospital mortality or 5-year survival. Surgery was performed in a high percentage of patients with BAV early in the course of IE, with a similar operative mortality rate (8.3%) as patients with non-BAV IE [5•]. Given the high rate of perivalvular abscess in BAV IE, patients with BAV should be carefully evaluated for this complication of IE, with consideration for routine transesophageal echocardiography.

Presence of a prosthetic heart valve is another important predisposition to IE, one for which antibiotic prophylaxis is still recommended before dental procedures by both ACC/AHA and ESC guidelines. Recent work from the International Collaboration on Endocarditis has found that S. aureus was the most common cause of prosthetic valve infection, with a major influence from health care-associated infection [6]. An extensive pathologic study of prosthetic endocarditis at the Mayo clinic from 1985 to 2004 consisted of 116 cases of surgically explanted valves, predominantly during the active phase of IE [7]. Interestingly, even across this long time-frame of inclusion, S. aureus was the most prevalent cause of both early-onset (38%) and late-onset (30%) prosthetic valve IE [7]. Coagulase-negative staphylococci were more common in early- and intermediate-onset IE, whereas Viridans streptococci and enterococci were more prevalent in late-stage IE [7]. Whereas health care-associated infection may have contributed to the high prevalence of S. aureus, selection bias for surgery in prosthetic valve infection likely contributed as well.

An uncommon cause of “culture-negative” IE is infection with Coxiella burnetii, also known as Q fever. Over a 23-year period, 104 cases of Q fever endocarditis with long-term outcome were reported by the French National Referral Center [8]. The median duration of follow-up was extensive (100 months), including serologic follow-up of immunoglobulin titers. In this series, 18 months of treatment with doxycycline and hydroxychloroquine was sufficient to treat Q fever IE in the vast majority of cases. Relapse was associated with shorter (<18 months) duration of therapy and prosthetic valve infection [8]. Mortality was associated with age, prosthetic valve infection, stroke at presentation, failure of fourfold decrease in phase 1 IgG or IgA levels during treatment, or persistence of IgM at 1 year [8].

Despite general consensus between ACC/AHA and ESC guidelines for IE prophylaxis, a recent survey of cardiologists suggests persistent uncertainty or variance in the understanding and adoption of these recommendations. In a survey of approximately 200 congenital heart disease cardiologists in Canada, the United States, Australia, and New Zealand, there was disagreement regarding prophylaxis for rheumatic mitral valve stenosis (not recommended by guidelines) and perimembranous ventricular septal defect status post-surgical patch closure without residual shunt 3 months post-operatively.

Diagnosis and prognosis of IE

The modified Duke Criteria [9] continue to serve as the case and clinical definition of IE, incorporating the major elements of IE such as pathologic findings, continuous bacteremia with microorganisms associated with IE, and evidence of endocardial involvement by echocardiography. However, approximately 10% of patients with definite IE by these criteria have negative blood culture results or absence of bacteremia, termed “culture-negative endocarditis.” In a large series of 759 cases of culture-negative IE from France, a causative microorganism was identified in 63% of cases, including 48% by serologic diagnosis of Q fever and Bartonella infections and 14% by polymerase chain reaction (PCR) of valvular biopsies diagnosing streptococci, Tropheryma whipplei, Bartonella species, and fungi [10•]. In addition, 2.5% of patients had non-infective etiologies such as malignancy or autoimmune disease [10•]. Based on these findings, serologic analysis for Q fever and Bartonella species and PCR of valvular tissue should be considered in cases of culture-negative IE.

Delays in the diagnosis of IE may occur due to such factors as non-virulent organism, non-specific manifestations (e.g., fever), and negative blood cultures due to recent antibiotic therapy. Diagnostic delays may allow progression of endocardial involvement without appropriate antibiotic therapy, leading to greater adverse hemodynamic consequences due to progressive regurgitation and higher embolic risk. A rapid, serologic biomarker for IE may reduce delay to diagnosis and treatment and thus, improve the outcome of this condition. Procalcitonin is a peptide released by the lungs or intestinal tract in response to bacterial inflammatory stimulus, and has been evaluated as a potential biomarker for bacteremia and differentiation of sepsis from systemic inflammatory response syndrome (SIRS). In a study of 759 consecutive patients referred for echocardiography for the clinical suspicion of IE, 147 (19%) were diagnosed with IE [11•]. Procalcitonin level was significantly higher in patients with IE than in cases of rejected IE [11•]. A number of clinical factors were independently associated with high procalcitonin level, including positive blood culture for typical IE microorganism, fever, symptoms ≤5 days, immunocompromised status, and male gender [11•]. Although a level above 0.04 ng/mL was found to have 95% sensitivity for the diagnosis of definite IE, 12% of the IE patients in this cohort had concentration lower than this threshold [11•]. Greater evaluation of this biomarker is needed to determine if this peptide is an earlier diagnostic finding in IE that may optimize the use of antibiotic therapy in patients with suspected or possible IE.

B-type natriuretic peptide (BNP) is another biomarker released by ventricular myocardium in conditions of increased wall stress, and an elevated serum level has been associated with the diagnosis of heart failure as the cause of dyspnea as well as worse prognosis in a number of cardiac conditions, including left ventricular systolic dysfunction, myocardial infarction, valvular heart disease, and hypertrophic cardiomyopathy. In a single center study of 45 patients with IE who had BNP measured for clinical indications, an elevated BNP level (≥400 pg/mL) was present in 32 patients and was associated with severe, left-sided valvular regurgitation [12]. An elevated BNP level was also associated with a higher incidence of the composite endpoint (death, abscess, or central nervous system event), with strongest association between BNP and abscess [12]. This relationship between BNP and clinical events remained present when only patients with normal left ventricular systolic function were evaluated [12].

In a separate study from the same institution, the prognostic association between a novel, high-sensitivity cardiac troponin T measurement and adverse outcome (death, central nervous system event, or cardiac abscess) was evaluated in 42 patients with definite IE [13]. Cardiac troponin T was detectable in 93% of IE patients using this novel highly sensitive assay, and a higher level was associated with the composite endpoint, which occurred in 15 of 42 patients [13]. Analysis of the receiver operating curve demonstrated that a level of ≥0.08 ng/mL had optimal specificity for the composite endpoint [13]. Further studies are needed to evaluate the utility of biomarkers, either individually or in combination, in the prognosis of IE and whether treatment such as surgical therapy based on prognostic biomarker elevation is associated with improved outcome (lower risk of IE complications).

Treatment

  • Recent studies have evaluated the effect of surgical intervention on outcome of IE. Recommended indications for surgery in IE (predominantly for native valve IE) by the ACC/AHA and ESC are largely similar (Table 2). In the absence of randomized clinical trials of surgery with medical therapy as compared to medical therapy alone, observational trials provide information regarding the appropriate indications for surgery in IE, typically being complications such as heart failure, intracardiac abscess, embolic events, severe valvular regurgitation, or failure of antibiotic therapy. In these studies, propensity adjustment for clinical characteristics associated with surgery has been used to reduce selection bias for the use of this intervention. In a prospective cohort study of 1552 patients with native valve IE from the International Collaboration of Endocarditis, surgical therapy was performed in 46% of cases during the index hospitalization and was associated with a significant reduction in in-hospital mortality (12.1% vs 20.7% in patients treated with medical therapy alone, P < 0.001) [14]. After propensity-based matching and adjustment for survivor bias, the survival benefit of surgery remained (absolute risk reduction, −5.9%), particularly for patients with complications such as paravalvular complications, systemic embolization (including stroke), and S. aureus infection [14].

  • More recently, two studies have added insights regarding the outcome of surgery for IE. In a very large study of 19,543 operations for IE from the Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database from 2002 through 2008, Gaca et al. [15•] examined the factors related to operative mortality (8.2%). Notably, nearly half of patients did not have active IE, defined as treatment with antibiotic therapy at the time of surgery for IE, and data on causative microorganism were not collected. Multiple variables were found to be independently associated with mortality, including urgency of surgery, hemodynamic status, renal failure, active infection, multiple valve involvement, diabetes mellitus, arrhythmia, and previous cardiac surgery [15•]. Given the low overall operative mortality and the negative influence of active infection in this study, the timing of surgery for IE needs further evaluation but a delay in surgery to reduce operative mortality must be tempered by the potential adverse consequences, including hemodynamic decompensation requiring urgent rather than elective surgery. Although recommendations from the ESC for surgery in IE suggests timing of surgery (immediate, urgent, or elective) [3], these recommendations have been based on consensus clinical judgment rather than outcome data. Potential reasons for delay or lack of surgical intervention in patients with IE and indications are shown in Table 2.

  • A single-center, retrospective study compared the results of early surgery or “conventional treatment” for left-sided, native valve IE with vegetation [16•]. From 1998 to 2006, early surgery was performed on 64 patients within 7 days of diagnosis and conventional management was pursued for 68 patients. Notably, the mean age of the patients was less than 50 years, they had few co-morbid medical conditions, and Viridans Streptococcus and culture-negative IE were the most common causes. The early surgery group had larger vegetations and 88% had severe valvular regurgitation. In the conventional management strategy, 47% of patients had surgery during the initial treatment period due to heart failure or worsening valve regurgitation. Five-year overall survival rates were very high (95%) and similar for the two groups, but patients treated with early surgery had a significantly lower rate of embolic event without a higher rate of recurrent IE (Fig. 1). Although the clinical characteristics of this cohort suggest a lower-risk patient group not typical of larger, multicenter study patients and influencing the highly favorable outcome in this study, the results do support current recommendations for surgery in IE with large vegetations in the absence of other surgical indications for the prevention of embolic events.

    Figure 1
    figure 1

    Results of early surgery (OP) or conventional treatment (CONV) for left-sided, native valve infective endocarditis with vegetation [16•]. (From Kim DH, Kang DH, Lee MZ, et al. Impact of early surgery on embolic events in patients with infective endocarditis. Circulation. Sep 14 2010;122(11 Suppl):S17–22; with permission).

    Full size image