Methicillin-Resistant Staphylococcus aureus Prosthetic Valve Endocarditis: Pathophysiology, Epidemiology, Clinical Presentation, Diagnosis, and Management

Staphylococcus aureus prosthetic valve endocarditis (PVE) remains among the most morbid bacterial infections, with mortality estimates ranging from 40% to 80%. The proportion of PVE cases due to methicillin-resistant Staphylococcus aureus (MRSA) has grown in recent decades, to account for more than 15% of cases of S. aureus PVE and 6% of all cases of PVE.

SUMMARY

Staphylococcus aureus prosthetic valve endocarditis (PVE) remains among the most morbid bacterial infections, with mortality estimates ranging from 40% to 80%. The proportion of PVE cases due to methicillin-resistant Staphylococcus aureus (MRSA) has grown in recent decades, to account for more than 15% of cases of S. aureus PVE and 6% of all cases of PVE. Because no large studies or clinical trials for PVE have been published, most guidelines on the diagnosis and management of MRSA PVE rely upon expert opinion and data from animal models or related conditions (e.g., coagulase-negative Staphylococcus infection). We performed a review of the literature on MRSA PVE to summarize data on pathogenic mechanisms and updates in epidemiology and therapeutic management and to inform diagnostic strategies and priority areas where additional clinical and laboratory data will be particularly useful to guide therapy. Major updates discussed in this review include novel diagnostics, indications for surgical management, the utility of aminoglycosides in medical therapy, and a review of newer antistaphylococcal agents used for the management of MRSA PVE.

INTRODUCTION

Staphylococcus aureus prosthetic valve endocarditis (PVE) is a devastating infection. The mortality rate due to methicillin-resistant Staphylococcus aureus (MRSA) has climbed in recent decades, reaching more than 15% of cases of S. aureus PVE (12) and 6.6% of cases of PVE (4, 6).

Management for MRSA PVE is complex, and guidelines recommend both a multidisciplinary team and an individualized approach to care. Given the lack of clinical trials testing treatments for MRSA PVE, many aspects of management lack an empirical basis. For example, the timing and necessity of valve surgery remain unknown. In some studies, hospital mortality rates for S. aureus PVE were significantly higher in patients who had not undergone valve surgery (2, 3, 7, 8, 10, 13–15), prompting some investigators to conclude that early valve surgery (EVS) should be considered a standard treatment for S. aureus PVE, especially in patients with early-onset infection (2, 11). However, recent literature and the experience of the International Collaboration on Endocarditis (ICE) have called into question the value of EVS (1, 11, 16–18). Specifically, work by Hill et al. (16) suggested that uncomplicated S. aureus PVE cases might be successfully managed without early valve surgery, and a multicenter study reported in 2015 by the ICE found no association between EVS and a reduction of 1-year mortality rates in patients with S. aureus PVE (11).

Furthermore, although multiple treatment guidelines advocate for incorporating an aminoglycoside into MRSA PVE therapy (19, 20) to promote sustained susceptibility to rifampin, these recommendations are largely based on experimental models (21–23) and the use of aminoglycosides for treatment of coagulase-negative Staphylococcus (CoNS) PVE (24). Clinical data on the benefit of aminoglycoside combination therapy in humans for MRSA PVE are lacking. Furthermore, the use of an aminoglycoside for other causes of endocarditis, including methicillin-sensitive Staphylococcus aureus (MSSA) native valve endocarditis (NVE) and CoNS PVE, demonstrates either harm or a lack of a survival benefit (24–27). Although current guidelines discuss this lack of evidence, most continue to recommend the addition of an aminoglycoside based on expert opinion (28–31). Trimethoprim-sulfamethoxazole, clindamycin, ceftaroline, daptomycin, linezolid, telavancin, oritavancin, tigecycline, and combinations that might result in synergy could have a role in treatment but have not yet been thoroughly studied.

TIMING, PATHOPHYSIOLOGY, PATHOGENESIS, AND HISTOPATHOLOGY

Timing of Infection

MRSA PVE is often dichotomized based on duration of disease (32) into early, defined as the first year postsurgery, or late, defined as after 1 year postsurgery. These thresholds have been developed based on the risk of developing PVE and differences in the microbiology of the disease between periods (33, 34). The risk of PVE is greatest during the first 3 months after surgery. The risk peaks approximately 15 days after surgery, during which the PVE risk is estimated to be 45 cases/100,000 patient days (35). After this time period, it decreases steadily to approximately 1 case/100,000 patient days from 150 days to 20 years postoperatively (35–39). The cumulative proportions of patients developing PVE range from 1 to 3% in the first 365 days after surgery according to several studies with close follow-up and from 3 to 6% in 5 years (35–39). S. aureus is a frequently encountered pathogen in both early and late PVE cases (6, 16, 40, 41), accounting for approximately 12 to 36% of early cases (53 to 69% at the first 2 months from surgery) and 18 to 30% of late cases (6, 41). A large, multicenter, international study (6) showed that MRSA was the causative microorganism in 18.9% of early cases of PVE, versus 3.3% of late cases.

Pathophysiology

Early PVE infection is believed to be caused by accidental seeding during surgery or due to bloodstream dissemination in the first hours to months postoperatively. Early after surgery, the prosthetic sewing ring and cardiac connection tissue and sutures have not yet endothelized. Fibronectin and fibrinogen coat these areas and are believed to be a possible nidus for infection. In late PVE infection, these cardiac structures become fully endothelized, and the pathogenesis of disease more resembles that of NVE (42).

The location and distribution of PVE also differ by the type of prosthetic valve and the route of infection (Fig. 1). In cases where the infection is introduced via contamination around the surgical site, it typically affects the annulus and sewing ring union, causing pseudoaneurysms, dehiscence, fistulas, and abscesses around valves (43–45). In one postmortem study, perivalvular invasion, which is commonly coupled with prosthesis dehiscence and paravalvular regurgitation, appeared in approximately 40% of autopsies performed for patients with PVE, and frank extension into tissue leading to myocardial abscess was seen in 15% (46). Annular and/or myocardial abscesses were present in 68% of the 47 patients with PVE in another study (47). Dehiscence of the prosthesis was described in more than 80% of cases of aortic and mitral prostheses examined by Ben Ismail et al. (48), and vegetations were seen in 75% (48). More than 1 year postoperatively, infection related to biological PVE is more commonly found at the prosthesis leaflets. Complications of late PVE include leaflet rupture and perforation of an aortic valve prosthesis, which can extend through the intervalvar fibrosa and annulus to cause pericarditis or, more frequently, can spread into the membranous portion of the interventricular septum and cause arrhythmia (49–51). Large vegetations may also keep the prosthesis open, causing malfunction or encroachment on the valve orifice, resulting in functional stenosis or regurgitation via malcoaptation or perforation.

FIG 1.

Prosthetic valves explanted from patients with MRSA prosthetic valve endocarditis, and received at the microbiology laboratory to perform valve culture and 16S PCR (courtesy of Mercedes Marín, Hospital General Universitario Gregorio Marañón, Madrid, Spain). Shown are a mechanical valve, a mitral ring, an aortic bioprosthesis, and a mitral bioprosthesis.

Pathogenesis

Several independent factors are involved in the development of MRSA PVE (Fig. 2). MRSA has microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) that recognize and bind to adhesion molecules of the fibrin-platelet matrices of “nonbacterial thrombotic endocarditis,” such as fibronectin, laminin, and collagen, and these can also adhere to the normal endothelium or minimally injured tissue (40, 52, 53). Once the thrombus is colonized by MRSA, this microorganism can proliferate, creating the characteristic vegetation of infective endocarditis (IE). The presence of cardiac prostheses introduces an additional variable that favors PVE, since MRSA can adhere by forming biofilms (Fig. 2 and 3) (40, 53). At initial placement, the ring-prosthesis interface is not endothelized and favors fibrin-platelet thrombus formation. The suture points where the prostheses are placed constitute a mechanism whereby MRSA can invade the heart tissue and form abscesses. In addition, the continuous stress caused by the repetitive movement in the bioprostheses may disrupt the surface of the leaflets and predispose to infection of the fibrin-platelet thrombus. MRSA can reach the prosthesis by contamination of the prosthetic valve during surgery (Fig. 4) or by a hematogenous route via a catheter-related infection, intravenous drug use, a surgical wound, or pulmonary or urinary tract infection (40, 53, 54). During the surgical procedure (Fig. 4), the surgical site may be exposed to MRSA from the patient’s or health care practitioner’s skin. The ability of S. aureus strains to produce biofilms (Fig. 3) in vitro has been linked to clinically persistent MRSA bacteremia (≥7 days) and the evolution of prosthetic valve vegetation propagation (52, 55, 56).

FIG 2.

Pathogenesis of MRSA (methicillin-resistant Staphylococcus aureus) PVE. MSCRAMMs, microbial surface components recognizing adhesive matrix molecules.

FIG 3.

An original electron scanning microscope image of a methicillin-resistant Staphylococcus aureus biofilm on a patient’s mechanical heart prosthesis. The image was prepared at both the Clinical Microbiology and Infectious Diseases Department and the Pathology Department of the Hospital General Universitario Gregorio Marañón and was taken at the National Center of Electron Microscopy (JSM 6400, CNME, Madrid, Spain).

FIG 4.

Surgery image (courtesy of Gregorio Cuerpo, Cardiac Surgery, Hospital General Universitario Gregorio Marañón, Madrid). Shown is mitral prosthetic valve endocarditis caused by S. aureus.

Nonvalvular invasive infection may also cause bioprosthetic valve endocarditis. For example, annular and myocardial invasion was observed in 38 of 85 patients (45%) in one study and was more common among cases of bioprosthetic PVE occurring during the first year after valve placement than in cases presenting later (59 versus 25%) (57). Invasive disease was more frequent in patients with early than in those with late bioprosthetic PVE (79% versus 31%) in another series (58).

A large proportion of cases of PVE is nosocomial and correlates with a high proportion of MRSA infections (6). An international study including 556 patients with PVE demonstrated that 36.5% of infections were nosocomially acquired or related to frequent health care visits (6). Similarly, the use of transcatheter aortic valve implantation (TAVI) also increased the risk for MRSA PVE (59, 60), as did the use of orotracheal intubation and percutaneous self-expandable valves (59). Further studies are needed to better establish a relationship between MRSA virulence factors observed experimentally in vitro and in animal models (61) and clinical disease in humans.

Histopathology

There is no typical pattern of PVE histological characterization in bioprosthetic valves. As bioprosthetic valves degenerate, they often create noninfective, calcific, vegetative-like lesions with inflammatory infiltrates, which can result in a noninfectious process that can mimic and be misdiagnosed as PVE. A retrospective pathological study of inflamed bioprosthetic valve tissues from 88 cases of resected bioprosthetic valves (21 for probable endocarditis and 67 for noninfective dysfunction) was performed to better define the histological criteria for PVE (62). PVE was histologically characterized by neutrophil-rich inflammatory infiltrates and the presence of microorganisms. Inflammatory infiltrates in valve tissue samples from the noninfective control group consisted mainly of lymphocytes and macrophages. In that study, having a neutrophil percentage exceeding 1.5% of the valve surface area was associated with a high specificity (94%) for infectious PVE (62).

EPIDEMIOLOGY

PVE occurs in 1% to 6% of patients after prosthetic valve placement (63), accompanied by an incidence of 0.3% to 1.2% per patient year (6, 64–66) and accounting for 16% to 31% of IE cases in several studies (1, 6, 67–70). The etiology of 146 early-PVE clinical cases and 140 late-PVE cases was summarized from 17 published reports (32). In that study, early S. aureus PVE accounted for 19.2% of the cases, and late complications were less likely with increasing time after surgery, occurring in only 11.4% of the cases. The incidence of both early and late PVE was correlated with increasing underlying comorbidity at the time of valve placement, surgeon experience, extracorporeal circulation duration, sterility of the heart-lung machine and the operating theater, extracardiac postoperative infection, and the length of time that the patient was monitored after surgery (32). MRSA accounted for approximately 6.5% of PVE cases (n = 556) assessed by the ICE Prospective Cohort Study (6). Figure 5 shows a comparison of cases of S. aureus (P = 0.003) and MRSA (P = 0.001) involved in PVE across geographic regions. The data provided by Wang et al. (6) are consistent with the global epidemiology of S. aureus (P = 0.007) and MRSA (P = 0.001) PVE found and also kindly shared by Murdoch et al. (4).

FIG 5.

Global epidemiology of MRSA involved in prosthetic valve endocarditis (PVE). The causative agents of PVE differ geographically (4, 6). Data from Wang et al. (6) were collected between June 2000 and August 2005 from 556 patients with infective prosthetic valve endocarditis in 53 sites worldwide (P = 0.003 for Staphylococcus aureus; P = 0.001 for MRSA [methicillin-resistant Staphylococcus aureus]). MSSA, methicillin-sensitive Staphylococcus aureus.

In one observational study, the prophylactic use of penicillinase-resistant penicillins (methicillin and oxacillin) during the perioperative periods reduced early postoperative S. aureus PVE (71). Unsurprisingly, the incidence of PVE was higher when the valve replacement occurred in the setting of active or recently treated endocarditis (36, 39, 72, 73). In several observational studies, bioprosthetic valves appear to increase the risk for infection over mechanical valves after 18 months (37, 72, 73). However, three randomized trials including 1,418 patients monitored for 8 to 20 years could not demonstrate a statistically significant difference in PVE occurrence between biological and mechanical valves (P = 0.45 [74], P = 0.71 [75], and P = 0.70 [76]). Another observational study that included 38,000 patients ≥65 years of age demonstrated a higher risk of endocarditis after a median of 12 years of follow-up among those with bioprosthetic valves (2.2% versus 1.4%; unadjusted hazard ratio, 1.69 [95% confidence interval {CI}, 1.43 to 2.00]) (77). Finally, Calderwood et al. (37) showed a higher risk of occurrence of PVE among patients (n = 116 out of 2,608 evaluated) who received mechanical valves than among those who received bioprosthetic valves at 3 months postsurgery (P = 0.02), but in contrast, Grover et al. (73) could not demonstrate any difference between the valves in 66 patients who developed PVE out of 1,032 observed during a mean length of follow-up of 7.7 years.

CLINICAL PRESENTATION, ASSESSMENT, AND DIAGNOSIS

PVE signs and symptoms are similar to those of native valve disease. Yet the clinical presentation of MRSA PVE is often nonspecific, especially soon after surgery, when inflammation and fever might occur for other reasons. Due to the intracardiac complications described above, clinical manifestations of PVE frequently include hemolysis, heart failure, valvular dysfunction, and/or new arrhythmia (13, 78, 79). Calderwood et al. (78) studied the outcomes of 116 patients with PVE and found that 64% of individuals with PVE suffered some combination of worsened or new cardiac failure, a changed or new cardiac murmur, continuous fever, or irregularities upon electrocardiography (ECG). These complications were more common in the first year after valve replacement and in aortic valve prosthesis infections (Table 1). New or changing murmurs, heart failure, and new electrocardiographic conduction disturbances are noted more often in PVE than in NVE cases due to the likelihood of invasive infection. ECG (Fig. 6), chest radiograph, and blood cultures are considered part of the standard work-up if there is a clinical suspicion of PVE.

TABLE 1.

Key articles on epidemiology, diagnosis, management, conclusions, and prevention of S. aureus PVE and IE^a

Author(s) (yr) (reference)	Epidemiology	Diagnosis	Management	Conclusion(s) and/or prevention strategy
Béraud et al. (2011) (215)	137 physicians participated in the study	Infective endocarditis	Endocarditis was treated with gentamicin doses of 3 mg/kg/day by 61% of physicians, 4 mg/kg/day by 22.1%, and 5 mg/kg/day by 16.9%	Guidelines were not followed by most of the physicians for gentamicin dosing in these patients; instead, they used validated regimens from published studies
Bille (1995) (216)	Antimicrobial therapy review	Endocarditis due to Staphylococcus	A combination of 3 antibiotics (vancomycin or oxacillin + gentamicin and rifampin) is suggested for PVE during at least 6 wk	Further studies are needed to incorporate novel treatment options, especially in patients affected by MRSA
Calderwood et al. (1985) (37)	2,642 patients who underwent valve replacement for the first time were included in the study	116 patients with PVE (4.4%)	At 12 mo, the risk of PVE was 3.1%, and at 60 mo, the risk was 5.7%; porcine valves had a significantly lower risk of PVE during the first 90 days from surgery than mechanical valves but a significantly higher risk after 12 mo postsurgery	There were significant differences in the risk of PVE depending on the type of valve, but no significant differences between porcine and mechanical valves were observed in the risk of having PVE after 5 yr
Cervera et al. (2014) (217)	Analysis of a study cohort	93 cases of S. aureus infective endocarditis (left sided)	57% had a vancomycin MIC of <1.5 μg/ml, and 43% had an MIC of ≥1.5 μg/ml	Percentages of in-hospital death varied significantly between both groups, at 30 and 53%, respectively
Chirouze et al. (2004) (1)	Evaluation of mortality risk	61 cases of S. aureus PVE	Patients who had their valve replaced early, despite having heart complications, showed lower mortality rates (P = 0.09)	S. aureus PVE is a disease with high morbidity and mortality rates (28.6–85.7%)
Chirouze et al. (2015) (11)	Impact of early valve surgery on clinical outcome of S. aureus PVE within the International Collaboration of Endocarditis	747 cases of definite left-sided PVE	Non-S. aureus PVE caused significantly lower rates of death after 1 yr than S. aureus PVE; at this time, patients with S. aureus PVE and EVS also had lower mortality rates (P < 0.01); EVS did not diminish mortality at 1 yr	Different factors should be taken into account before deciding on EVS
Cosgrove et al. (2009) (218)	236 patients from 44 hospitals and 4 countries were prospectively evaluated	S. aureus bacteremia and native valve infective endocarditis	Vancomycin or an antistaphylococcal penicillin + low-dose gentamicin or daptomycin alone was administered to patients; renal adverse events were evaluated	Low-dose gentamicin should not be used routinely for S. aureus bacteremia and native valve infective endocarditis due to the nephrotoxicity shown
de Feiter et al. (2005) (219)	Fusidic acid, rifampicin, vancomycin, oxacillin, and gentamicin treatment failures	Patient with Staphylococcus epidermidis PVE	Despite the nonapproval for this indication, linezolid was administered to this patient	Patient had a favorable outcome with linezolid
Del Río et al. (2014) (220)	Rescue therapy with imipenem + fosfomycin	Complicated bacteremia and MRSA endocarditis	Treatment was successful in 69% of cases; the mortality rate due to MRSA was 1/5 (20%)	Combination therapy was safe and effective as rescue therapy
Fernández Guerrero et al. (2009) (15)	Incidence of infective endocarditis, epidemiology, clinical features, prognosis	Definite S. aureus endocarditis (right sided and left sided)	NVE was a less common hospital-acquired infection than PVE; for both types of endocarditis, renal and cardiac failure and central nervous system complications were detected	Valve replacement significantly improved outcomes for patients with PVE
Fowler et al. (2006) (157)	Daptomycin vs standard therapy	S. aureus bacteremia and endocarditis	Microbiological failure was more common in the group treated with daptomycin than in the one with standard therapy	A noninferiority rate was observed in the group treated with daptomycin compared to the one treated with standard therapy for bacteremia and right-sided endocarditis caused by S. aureus
Hasbun et al. (2003) (221)	Prognostic factors	Left-sided endocarditis (native valve) with complications	Factors related to mortality after 6 mo, including abnormal mental status, bacterial cause, comorbidities, medical treatment, moderate/severe congestive cardiac failure	4 groups of patients were identified depending on the mortality risk 6 mo after baseline
Holland et al. (2014) (222)	Review on hospital management	Bacteremia caused by S. aureus	Diagnostic methods and antibiotic treatment strategies	There are groups of patients who do not need TEE
John et al. (1998) (3)	Clinical strategies and prognostic factors	Definite PVE caused by S. aureus	Complications affecting the central nervous system (33%) and heart (67%) were found; the 3-mo mortality rate was 42%	More patients died due to heart problems than due to problems affecting the central nervous system, but this mortality was diminished when there was surgery for valve replacement during antibiotic treatment
Kang et al. (2012) (223)	6-wk occurrence of embolic events and mortality	Patients with large vegetations, severe disease in valves, left-sided endocarditis due to infection	Patients were randomized into 2 groups, conventional treatment or early surgery	Rates of embolic events and mortality significantly decreased in the group with early surgery compared to the group of patients treated conventionally
Karchmer et al. (1983) (24)	Retrospective study of 75 PVE cases	Staphylococcus epidermidis PVE	The gentamicin susceptibility rate was 78%, and those for rifampin and vancomycin were 100% for all isolates tested; dysfunction of valves and tissue progression were the most common problems, needing surgery in 30 cases	Antibiotic therapy including vancomycin + rifampin or an aminoglycoside increased favorable outcome rates; surgical treatment was also important
Karchmer (1991) (224)	Infection control	PVE	During a year postsurgery, the nosocomial risk of PVE was 1.4–3.0%; the most common reason for hospital-acquired PVE was methicillin-resistant coagulase-negative Staphylococcus	Further studies are needed to detect postsurgical causes of hospital-acquired PVE and diminish them
Le and Bayer (2003) (225)	Review on antibiotic treatment for endocarditis caused by frequently detected microorganisms		Few strategies for in vitro, experimental, and clinical evaluation of enterococcal endocarditis have been shown	Human clinical data are scarce on combination antibiotic treatment for infective endocarditis due to S. aureus
Mayer and Schoenbaum (1982) (226)	Review and approach	PVE	Higher rates of morbidity and mortality were detected in early than in late PVE cases; the etiology often included fungi, staphylococci, and Gram-negative rods in early PVE and streptococci in late PVE	Factors related to poor outcome were early PVE, paravalvular leakage, emboli, persistent fever, nonstreptococcal microorganisms, nonheterograft aortic valve, congestive cardiac failure
Muñoz et al. (2015) (12)	Epidemiology, clinical features, prognostic factors	Infective endocarditis (1,804 cases)	Previous cardiac surgery, atrial fibrillation, cardiac complications and failure, septic shock, age, cerebrovascular complications, or Candida or Staphylococcus cause was related to in-hospital deaths (28.9%); after 1 yr, association was found for cancer, cardiac failure, age, and renal insufficiency (11.2%)	The rates of in-hospital and 1-yr deaths were elevated, and surgery was the only protective factor
Murdoch et al. (2009) (4)	Global infective causes and clinical features	Infective endocarditis	Infections of mitral and aortic valves due to S. aureus were the most frequent presentation, also complicated with heart failure, stroke, and other emboli and abscess in the heart	Risk factors for in-hospital mortality were lung edema, age, prosthetic infection, S. aureus or coagulase-negative staphylococcal cause, mitral vegetation, and valve problems
Rajashekaraiah et al. (1980) (26)	Evaluation of tolerance (MBC/MIC > 16)	S. aureus bacteremia and endocarditis	Tolerant microorganisms were accompanied by more deaths, complications, hospitalization in ICU, prolongation of fever	Poor outcomes of endocarditis were more common in cases caused by tolerant microorganisms than in cases caused by sensitive ones
Ribera et al. (1996) (27)	Cloxacillin vs cloxacillin + gentamicin during 2 wk	S. aureus endocarditis (right sided)	Mortality occurred 1 and 2 cases, respectively	Combination treatment was not more effective than the single one
Sohail et al. (2006) (17)	Mortality rates in patients who received medical vs surgical treatment	S. aureus PVE	Mortality rates of 48% and 28%, respectively	The no. of deaths was lower in the surgical group; bioprosthetic valves and ASA class IV were prognostic factors
Wang et al. (2007) (6)	Global infective causes and clinical features	PVE	The most frequent microorganism was S. aureus; 36.5% of cases were related to health care; the of in-hospital death rate was 22.8% and was related to health care, age, persistent bloodstream infection, S. aureus cause, and cardiac and CNS problems	S. aureus is globally the main cause of PVE, and the presence of complications is an important prognostic factor
Wareham et al. (2005) (227)	Cases treated with linezolid	MRSE and VRE endocarditis	In vitro study of Staphylococcus epidermidis and Enterococcus faecalis infections treated with linezolid + gentamicin or vancomycin	Outcome was favorable in both cases
Watanakunakorn (1979) (32)	Treatment with penicillin vs penicillin + gentamicin	S. aureus endocarditis	The death rate was 40% in both groups of patients	There is no clinically demonstrated advantage of the use of combination therapy with gentamicin
Wilson et al. (1995) (228)	Treatment efficacy experience	Endocarditis caused by enterococci, staphylococci, streptococci, and members of the HACEK group	Recommendations of treatment based on previously reported studies	Literature is scarce
Yaw et al. (2014) (229)	Clinical outcome evaluation	MRSA and MSSA bacteremia	Rehospitalization rates associated with infection were similar in both groups of patients	MSSA bacteremia outcomes were more favorable than MRSA bacteremia ones; patients colonized with MRSA should be treated carefully

^aASA, American Society of Anesthesiologists; EVS, early valve surgery; MBC, minimal bactericidal concentration; MRSA, methicillin-resistant Staphylococcus aureus; MRSE, methicillin-resistant Staphylococcus epidermidis; NVE, native valve endocarditis; PVE, prosthetic valve endocarditis; TEE, transesophageal echocardiography; VRE, vancomycin-resistant Enterococcus; ICU, intensive care unit; HACEK group, a group of Gram-negative bacilli consisting of Haemophilus spp., Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens, and Kingella spp.

FIG 6.

Electrocardiogram at 25 mm/s of a patient with aortic MRSA PVE with nonspecific findings, including tachycardia (heart rate of approximately 125 beats per minute), a first-degree block with a PR interval exceeding a duration of 0.2 s (indicated by the arrow), and ST segment changes (indicated by the circle).

Pulmonary, neurological, kidney, and musculoskeletal complications or complications associated with systemic infections might also occur, at the same time, in some PVE cases (80–87). Of note, S. aureus is the pathogen most commonly related to IE complications compared to others (88). Lung complications are more frequently seen in patients with right-sided endocarditis, in several cases presenting manifestations such as pneumonia, abscesses, pleural effusion, and/or atelectasis.

Central nervous system (CNS) complications, primarily embolic infarcts or hemorrhage, are also more common in PVE than in NVE cases, ranging from 20 to 40%, with overt arterial emboli being noted for 40% of patients with PVE (48, 89–91). Similar rates of embolic stroke (18%) and ischemic stroke (20%) among patients with PVE were reported by Davenport and Hart (92) and Keyser et al. (89), respectively. Fernández Guerrero et al. (15) observed that, compared to NVE, S. aureus PVE clinical manifestations were characterized by less-frequent cardiac murmurs and a shorter symptomatic prodrome.

In PVE patients with no prior antibiotic therapy, blood cultures are positive in at least 90% of cases (63). Molecular techniques, such as DNA examination by pulsed-field gel electrophoresis or 16S rRNA sequencing, can be used when standard blood or tissue cultures have not revealed a pathogen (93, 94).

Transesophageal echocardiography (TEE) has poorer diagnostic validity for PVE (including MRSA PVE) than for NVE (95). However, TEE with a high-resolution biplane or multiplane transducer that allows continuous-wave and pulsed-wave Doppler and color flow imaging can increase the accuracy of diagnosis of PVE (96). TEE is considered the method of choice for diagnosis of PVE (97), because although transthoracic echocardiography (TTE) may lead to a diagnosis in some cases, TEE has a higher sensitivity for PVE (98, 99) (Fig. 7A). This increase in sensitivity is not accompanied by a loss of specificity and is independent of the valve type or position (98–101). For suspected PVE, TEE sensitivity and specificity have been estimated at 77 to 90% and 90%, respectively, compared to TTE, with a specificity and sensitivity estimated at 40 to 70% and 90%, respectively (102). Nonetheless, TTE has value for the assessment of ventricular size and function and severe hemodynamic lesions in valves and can often detect anterior aortic prosthetic valve abscesses (103) (Fig. 7B). Typically, the ventricular surfaces of prostheses in the mitral, tricuspid, and aortic positions are better viewed by TTE, whereas TEE has additional value for viewing the aortic valve surfaces, tricuspid and mitral valves, and the exit of the prosthesis in the aorta; fistula and abscess detection; detection of paraprosthetic leaks; and visualization of a mitral valve prosthesis (104, 105). However, TEE will miss some cases of prosthetic valve dehiscence. In a study of 26 patients with PVE, there were 14 (56%) cases of aortic valve dehiscence intraoperatively, 4 (29%) of which were not detected upon TEE (106). One prospective study that compared the paired use of TTE with TEE in 114 episodes of clinically suspected IE (34 PVE and 80 NVE) found that the results of the two tests were concordant in only 55% of cases (107). TEE prompted reclassification for 34% of patients with prosthetic valves, compared with 11% of patients with native valves.

FIG 7.

Imaging modalities for MRSA prosthetic valve endocarditis diagnosis. (A) Transesophageal echocardiography demonstrating a 1.4- by 1.2-cm highly mobile echodensity at the ventricular side of the bioprosthetic aortic valve (indicated by the arrow), without significant valvular dysfunction. (B) Transthoracic echocardiogram image showing a parasternal view with the prosthetic aortic valve, right ventricular outflow tract, and aorta on the top; the left atrium and mitral valve at the bottom; and the left ventricle on the left. The prosthetic aortic valve is not well visualized, but there is a vegetation on the ventricular side (indicated by the circle) and anterior aortic root thickening (indicated by the arrow), suggestive of an aortic root abscess. There is also mild prosthetic aortic valve regurgitation, but it cannot be appreciated in the still image. (C) Electrocardiogram-gated multidetector CT angiography demonstrating a 4- by 8-mm vegetation on the bioprosthetic aortic valve. (D) PET/CT image at the posterior prosthetic aortic valve, after 16.29 mCi [¹⁸F]fluorodeoxyglucose uptake. The arrow notes an area of hyperintensity, suggesting a focus of inflammation or infection consistent with prosthetic valve endocarditis.

The Duke criteria have been described as being less useful for the diagnosis of PVE because of low sensitivity (108, 109) compared to the sensitivity of 70 to 80% (110, 111) in the diagnosis of NVE. Because of diagnostic challenges in PVE with standard testing and clinical scores, other techniques have been assessed. Magnetic resonance imaging (MRI) and multislice computed tomography (MS-CT) might improve the detection of an intra- or pericardiac anatomical complication (112). The value of ECG-gated multidetector CT angiography (MDCTA) was addressed by three studies (113–115), which demonstrated a >90% sensitivity for the diagnosis of PVE (113), which was improved to a sensitivity and a specificity of 100% and 83%, respectively, after completion of routine testing for endocarditis and resulted in modification of therapy in 25% of cases (115) (Fig. 7C).

Recently, nuclear imaging has shown promise for improving diagnostics. [¹⁸F]fluorodeoxyglucose positron emission tomography electrocardiogram-gated computer tomography ([¹⁸F]FDG PET/CT) scans have shown utility as an additional diagnostic criterion for PVE in cases where a diagnosis cannot be made with standard echocardiography (116–119) (Fig. 7D). FDG PET/CT detects inflammation early in the infection process (120). For suspected PVE, [¹⁸F]FDG PET/CT demonstrated a 67 to 100% positive predictive value, a 50 to 100% negative predictive value, 73 to 100% sensitivity, and 71 to 100% specificity (116–119, 121–123). As such, an algorithm reported by Saby et al. (116) shows that PET/CT is useful to assess probable PVE cases. This algorithm incorporates the PET/CT 2013 modified Duke criteria for the diagnosis of possible PVE cases that do not meet criteria for endocarditis by the modified Duke criteria but remain under high clinical suspicion. The sensitivity of the modified Duke criteria significantly increased with the addition of PET imaging to the scoring system, from 70% (95% CI, 52% to 83%) to 97% (95% CI, 83% to 99%) (P = 0.008). This result was the consequence of a significant reduction (P < 0.0001) in the number of possible PVE cases from 56% to 32% (116).

Three retrospective studies have addressed the value of technetium-99m-hexamethylpropylene amine oxime (^99mTc-HMPAO)-labeled leukocyte scintigraphy with single-photon emission tomography/computed tomography (SPECT/CT) for the detection of PVE (119, 124, 125). Probable (125) or definite (119, 124) IE cases were included. Globally leukocyte scintigraphy showed an 85 to 100% positive predictive value, a 47 to 81% negative predictive value, 64 to 90% sensitivity, and 36 to 100% specificity (126). Further evaluations with increased sample sizes will be helpful for better defining a role and optimal scenario of nuclear imaging in the diagnosis of PVE.

TREATMENT

The S. aureus PVE mortality rate remains high (25 to 42%) (3, 54) despite advances in antibiotic treatment (1). Both the American Heart Association (AHA) and its European counterpart, the European Society for Cardiology (ESC), suggest a triple-drug regimen (class I; level of evidence C) with vancomycin and rifampin for ≥6 weeks and gentamicin during the first 2 weeks for the management of MRSA PVE (Table 2) (29, 30).

TABLE 2.

International guidelines for therapy of MRSA PVE^a

ESC guidelines for adults	AHA guidelines
	Pediatric	Adult
Vancomycin at 30–60 mg/kg Q24h i.v., BID or TID during ≥6 wk	Vancomycin at 40 mg/kg Q24h i.v. (maximum dose, 2 g Q24h), BID or TID during ≥6 wk	Vancomycin at 30 mg/kg Q24h i.v., BID during ≥6 wk
Rifampin at 900–1,200 mg Q24h i.v./orally, BID or TID during ≥6 wk	Rifampin at 20 mg/kg Q24h i.v. (maximum dose, 900 mg Q24h), TID during ≥6 wk	Rifampin at 900 mg Q24h i.v./orally, TID during ≥6 wk
Gentamicin at 3 mg/kg Q24h i.v./i.m., once a day or BID during the first 2 wk	Gentamicin at 3 to 6 mg/kg Q24h i.v./i.m., TID during the first 2 wk	Gentamicin at 3 mg/kg Q24h i.v./i.m., BID or TID during the first 2 wk

^aThe doses of these drugs must be adjusted in the setting of renal insufficiency. The intravenous (i.v.) route is preferred, particularly in infants and children. Q24h, every 24 h; BID, twice a day; TID, three times a day; i.m., intramuscular.

These recommendations are based on CoNS PVE therapeutic regimen activity, IE experiments, and retrospective clinical series (57, 127–130). A single retrospective study of valve cultures from 61 patients with staphylococcal PVE (29 due to S. aureus and 32 due to CoNS) treated surgically showed that those receiving combination therapy were 5.9 times more likely to have culture-negative valves than patients receiving monotherapy (adjusted by the length of therapy prior to surgery) (128). In this study, the six patients treated with a triple-drug regimen including rifampin prior to surgery had negative valve cultures. Regarding surgical indication, all of the groups of therapy evaluated for S. aureus NVE and any staphylococcal NVE or PVE were similar (P = 0.53, P = 0.51, and P = 0.66, respectively). No data on clinical outcomes restricted to those with S. aureus PVE were reported.

Antibiotic recommendations for treatment of MRSA PVE are also partially based on ex vivo and animal studies of Staphylococcus epidermidis PVE. A retrospective evaluation of 23 cases of methicillin-resistant S. epidermidis (MRSE) PVE (129) showed an increase in serum bactericidal activity when rifampin was added to vancomycin regimens (129). Studies regarding MRSE endocarditis in rabbit models also demonstrated that gentamicin, rifampin, and vancomycin in combination increased the efficacy of eradication of S. epidermidis from vegetations compared to beta-lactam antibiotics alone (131, 132). Another study (130) found that when rifampin was added to the combination of vancomycin and gentamicin in broth, there was an enhanced bactericidal effect in rabbits despite antagonism of the bactericidal rate. Notably, similar data for these relationships on MRSA are lacking.

Vancomycin

Vancomycin remains the mainstay of therapy for MRSA PVE (29, 30). Vancomycin dosing should be based on actual body weight and adjusted to achieve troughs of 15 to 20 μg/ml (29). Vancomycin binds the terminal d-alanyl-d-alanine moieties of N-acetylmuramic acid (NAM)/N-acetylglucosamine (NAG) peptides in the bacterial cell wall and thus prevents the addition of the NAM/NAG peptide subunits into the peptidoglycan matrix of MRSA. It also acts by altering bacterial cell membrane permeability and RNA synthesis (133). Common vancomycin toxicities include hypersensitivity (“red man syndrome,” related to the infusion rate) after intravenous administration and nephrotoxicity when used concomitantly with aminoglycosides (134). Significant weight gain can also occur during prolonged treatment, particularly in older men (135).

Rifampin

The rifampin mechanism of action is based on the suppression of RNA synthesis through the inhibition of the bacterial DNA-dependent RNA polymerase. Rifampin is believed to bind to a pocket of the RNA polymerase β-subunit within the DNA/RNA channel. This noncompetitive inhibitor prevents RNA synthesis by directly blocking RNA elongation and thus preventing the synthesis of host bacterial proteins (136). Hepatic and immunoallergic toxicities are the most common rifampin adverse effects. Hepatotoxicity usually affects patients with previous liver damage and is associated with the dose administered. Adverse immunoallergic events might be minor or major and are predominantly observed after prolonged or intermittent treatment (137).

Animal model experiments, in vitro data, and clinical observations have each demonstrated an outsized beneficial effect of rifampin on foreign-material infections (24, 57, 127, 130, 138–140). Nevertheless, S. aureus has a high intrinsic mutation rate at the rpoB gene, encoding the rifampin-binding site. When large numbers of S. aureus bacteria are exposed to rifampin alone or in combination with ineffective antimicrobials, single mutations in this region allow the rapid selection of a rifampin-resistant subpopulation (57, 138). Consequently, recommended regimens support the selection of two additional antimicrobials with rifampin, to protect against the development of resistance to rifampin. Similarly, many have recommended waiting until vancomycin and gentamicin have been administered for 3 to 5 days, bacteremia has resolved, and undrained abscesses or collections have been debrided prior to initiation of rifampin (30). This approach is supported by the in vitro antagonism demonstrated when rifampin is exposed to replicating bacteria in combination with other antimicrobials (141) and the synergism detected when bacteria are instead in a latent state (142). This effect can be seen in biofilm-mediated foreign-body infections, such as those associated with orthopedic hardware, prostheses, or vascular grafts (143). Lowy et al. (131) prevented the development of rifampin-resistant S. epidermidis by adding either vancomycin or gentamicin in a rabbit endocarditis model, but these observations have unclear clinical utility in treating PVE in humans.

Rifampin is recommended for a minimum of 6 weeks, at a dosing regimen of 300 mg every 8 h, in combination with gentamicin for the first 2 weeks and vancomycin for the full treatment course of 6 weeks (29, 31). Notably, some authors alternatively recommend rifampin at 600 mg once daily or 300 to 450 mg every 12 h with another antistaphylococcal antibiotic for S. aureus infections (31).

The data on synergy between rifampin and other antimicrobials are conflicting (144, 145). Although several studies (146–148) have demonstrated that incorporating rifampin into failing therapies can increase bactericidal rates and the chances of eradication of serious S. aureus and S. epidermidis infections, others have reported both in vitro synergy and antagonism for rifampin in combination with beta-lactam agents, vancomycin, or gentamicin against S. aureus (149, 150). Another study concluded that there was no synergy or antagonism against MRSE regarding the concomitant use of other antibiotics (cephalothin, nafcillin, vancomycin, or gentamicin) with rifampin (127).

MRSE isolates harboring rifampin mutations appeared to be as virulent as their rifampin-sensitive antecedents in rabbit endocarditis models, although a similar effect was not seen in rifampin-resistant S. aureus isolates in a mouse model (151). The mechanism of decreased virulence may be due to decreased production of toxins by rifampin-resistant S. aureus (127). The emergence of rifampin resistance was not prevented in vivo by the combination of rifampin with a beta-lactam antibiotic, although it was prevented in vitro. There was no decrease in virulence of rifampin-resistant methicillin-resistant S. epidermidis in comparison to rifampin-sensitive antecedent strains (127). The MRSE study showed good bactericidal efficacy of rifampin in vitro when there was a prevention of emergent rifampin-resistant mutants by a beta-lactam antibiotic.

Gentamicin

European and U.S. guidelines suggest intramuscular or intravenous gentamicin at a dose of 3 mg/kg of body weight every 24 h, once daily or in 2 or 3 divided doses, for the treatment of MRSA PVE. Gentamicin serum levels and renal function should be assessed at least once weekly, and more-frequent measurements are suggested in cases of renal insufficiency. Gentamicin doses as high as 4 mg/kg/day have been demonstrated to be successful in MRSA PVE treatment without additional toxicity in both human and animal studies (152, 153). In nonobese adults, the gentamicin dose is based on ideal body weight. Gentamicin is not distributed into adipose tissue, as it is highly hydrophilic. Therefore, corrected body weight should be used for dosing calculations for obese patients, rather than ideal body weight. The gentamicin mechanism of action is based on irreversible binding to specific proteins on the 30S subunit of the MRSA ribosome and the decoding site of the 16S rRNA. This leads to a misreading of the mRNA, the addition of incorrect amino acids in the growing peptide chain, and the interruption of MRSA protein synthesis (154).

Ototoxicity and nephrotoxicity are important adverse events limiting gentamicin clinical use. Both side effects are associated with the dose administered and might not appear until the end of treatment. Kidney damage, in contrast to inner ear damage, is usually reversible but can be fatal (155, 156). Gentamicin trough and peak values lower than 1 μg/ml and 3 to 4 μg/ml, respectively, are recommended when gentamicin is administered every 8 h. When gentamicin is dosed daily, serum troughs should be lower than 1 μg/ml. ESC guidelines recommend that peaks be assessed once after infusion, with a goal range of 10 to 12 μg/ml (per AHA/IDSA guidelines, there is no role for measuring peak gentamicin concentrations following single daily dosing).

The risk-to-benefit ratio of gentamicin therapy for MRSA PVE remains controversial. In a study of 35 ex vivo strains of S. aureus isolated from blood cultures of septic patients (23), vancomycin or nafcillin combined with tobramycin, gentamicin, or kanamycin had improved activity against most of the strains. In a rabbit model of S. aureus endocarditis, nafcillin and gentamicin altogether achieved a faster eradication of S. aureus at the heart vegetation than nafcillin in monotherapy (22).

Nonetheless, clinical data in support of gentamicin use in this setting are lacking. In a recent study comparing daptomycin therapy with an antistaphylococcal penicillin or vancomycin in combination with gentamicin for MSSA or MRSA bacteremia and/or right-sided endocarditis (157), significantly more patients who received standard therapy with gentamicin suffered nephrotoxicity (18.1% versus 6.7% with renal tubular necrosis and 46.8% versus 19.8% with worsening creatinine clearance), without an improvement in clinical outcomes for those who received gentamicin (157). Whether these results pertain to MRSA PVE is unknown, but this topic warrants urgent study given the ongoing guideline recommendations for the use of gentamicin in the treatment of MRSA PVE.

Alternative Therapies

If an isolate is resistant to gentamicin and all available aminoglycosides, a fluoroquinolone to which the strain is highly susceptible has been recommended (138–140). If the patient is treated with a fluoroquinolone instead of an aminoglycoside, a three-drug regimen for the entire course of treatment is preferred. In cases of resistance to aminoglycosides and fluoroquinolones, ceftaroline, trimethoprim-sulfamethoxazole, or linezolid (158) has been proposed as the third agent during the first 2 weeks of treatment, if the isolate is susceptible in vitro.

For cases of MRSA PVE with reduced vancomycin susceptibility (MIC > 1.0 μg/ml), substantial toxicity, or failure of vancomycin, the optimal treatment is not established. Options include high-dose daptomycin (8 to 10 mg/kg once per day, if the isolate is daptomycin susceptible), linezolid, telavancin, ceftaroline, and daptomycin combined with ceftaroline, nafcillin, or fosfomycin, combinations that might result in synergy (159–165). Yet reported clinical experience with these treatments and combination therapies in MRSA PVE is limited. Daptomycin combination with rifampin and gentamicin has been recommended in these cases as a second-line therapy for MRSA PVE (30). Synergy between β-lactams and daptomycin is associated with several characteristics, including increased daptomycin binding and β-lactam-mediated potentiation of innate immunity, but the precise molecular mechanism is unknown (139). Dhand et al. (166) reported a series of seven cases with rapid clearance of persistent MRSA bacteremia when high-dose nafcillin was added to high-dose daptomycin. Other experts (167) suggested the use of high-dose daptomycin combined with fosfomycin for MRSA PVE.

There are several case reports of ceftaroline being used to successfully treat refractory or drug-resistant MRSA PVE, either alone (168) or in combination therapy (169). In one case report, the efficacy and tolerability of ceftaroline were demonstrated in one patient with osteomyelitis and endocarditis caused by an MRSA strain that was not susceptible to daptomycin (170). A patient with MRSA aortic PVE was also cured with prolonged high-dose daptomycin plus ceftaroline after other therapeutic failures (169).

Ceftaroline fosamil was the most active bactericidal drug in a rabbit model of MRSA endocarditis (171). For each MRSA strain, rabbits were randomized to no therapy (controls), a ceftaroline fosamil dose equivalent to 10 mg/kg/12 h in humans (600 mg twice daily), daptomycin at a dose comparable to 6 mg/kg/24 h in humans, or a tigecycline dose equivalent to 100 mg/24 h in humans plus 50 mg/12 h. Both ceftaroline and daptomycin exhibited high bactericidal efficacy based on MRSA vegetation reduction rates (>5 log₁₀ CFU/g), whereas tigecycline did not show bactericidal efficacy, and MRSA vegetation reduction rates were <2 log₁₀ CFU/g in comparison to controls. However, the MRSA vegetation sterilization rate by ceftaroline was 100%, in contrast to the rate of 57% reached by daptomycin, with resistant mutants being seen only in the daptomycin therapy group. Recent clinical data also demonstrate ceftaroline as an alternative for MRSA bacteremia salvage treatment (170, 172–177). Two observational studies also suggest that ceftaroline therapy alone (178) or in combination with trimethoprim-sulfamethoxazole (179) can be used to treat invasive MRSA infection, although more experience and, if possible, adequately designed clinical studies are needed before there is widespread recommendation for its use.

Linezolid is an alternate treatment in cases of MRSA PVE complicated by drug allergy or intolerance, although its use is limited by a relative scarcity of data and side effects of prolonged use. Among 33 cases of endocarditis treated with linezolid, 21 (63.6%) had a favorable outcome. PVE accounted for 25% of the reviewed cases, and MRSA accounted for 24.2% (1 case of MRSA PVE among 8 PVE cases [12.5%]). The one patient with MRSA PVE treated with linezolid had a favorable outcome. Additional efficacy and tolerability data are required to better support the use of linezolid for PVE (180).

In an experimental rabbit model, Miró et al. (181) suggested that telavancin could be as effective as vancomycin in the treatment of endocarditis caused by glycopeptide-intermediate S. aureus (GISA). Other experimental IE models also showed telavancin bactericidal activity against different MRSA strains, including daptomycin-resistant S. aureus, vancomycin-intermediate S. aureus (VISA), and GISA (182–184). Telavancin (185) and also quinupristin-dalfopristin (186) have been reported as favorable rescue therapies in MRSA IE patients after vancomycin clinical failure.

The efficacy of oritavancin was also recently assessed in animal models of left-sided MRSA endocarditis. The drug has gained attention for its single intravenous dosing schedule at 1,200 mg over 3 h, which allows treatment of complicated MRSA infections without the need for indwelling central venous catheters, a particular concern for patients with recent or active injection drug use, and has shown promise in other refractory and drug-resistant cases of PVE (187, 188). A left-sided MRSA endocarditis rabbit model suggested that oritavancin was superior to vancomycin in resolving bacteremia and reducing bacterial counts in vegetations and tissues (189). Those investigators concluded that oritavancin was microbiologically effective and might be an alternative to vancomycin in treating similar infections in humans (189). Of note, Stewart et al. (190) reported a case series of 10 patients treated with oritavancin, including 1 patient with NVE due to group B Streptococcus, who unfortunately failed treatment. Although oritavancin is currently indicated only for acute bacterial skin and skin structure infections, it has the potential to play an increasing role as an agent against MRSA PVE, particularly in light of the ongoing international opioid epidemic (191, 192).

Tigecycline has activity against MRSA; however, it has been demonstrated to have a lower efficacy than vancomycin against MRSA strains (193), and its peak serum concentrations do not exceed 1 µg/ml (194, 195). As such, tigecycline is not typically recommended as an agent for MRSA bacteremia or endocarditis. Similarly, although clindamycin and trimethoprim-sulfamethoxazole are theoretically alternative options for MRSA PVE, experience with these agents for this condition is limited, and they are typically considered only in the case of severe drug intolerance or as considerations in synergistic regimens (196).

The “Endocarditis Team”

A multidisciplinary team including specialists from different clinical fields (neurology, microbiology, infectious diseases, cardiology, imaging, surgery, and congenital heart disease) (30) has been recommended for the management of PVE (197, 198). Early involvement of all team members is crucial to this strategy. Some studies have reported reductions in 1-year mortality rates of PVE with the implementation of such teams (199, 200).

Choice and Timing of Surgery

Surgery is recommended for high-risk patients, particularly those with valvular failures complicated by heart failure, abscess formation, or fistulas and those not responding to maximally effective antimicrobial therapy (30). The timing of surgery for PVE is a topic of ongoing debate (2, 7, 16, 17, 201–203) and is a decision based on surgeon- and patient-specific factors, including complications of disease, such as CNS emboli and hemorrhage, and overall surgical risk. A recent study of 4,166 cases of infective PVE and NVE in patients with heart failure found decreases in both hospital and 1-year deaths with early surgery (during the index hospitalization) (65). More recently, two prospective cohort studies, reported by Chirouze et al. (11) and Lalani et al. (204), found a crude reduction in 1-year mortality rates with early versus delayed surgery but similar rates after consideration of confounding and survivor bias. Thus, the decision on the timing of therapy continues to largely depend on surgical risk and input from multidisciplinary teams.

PROGNOSIS

The PVE mortality rate remains high in the current era, ranging from 30 to 80% for early PVE and 20 to 40% for late PVE (54, 63). The identification of high-risk subgroups is critical to establishing more-effective treatment strategies (9). Health care-associated disease, staphylococcal or fungal etiologies, older age, diabetes mellitus, early PVE, heart failure, CNS embolic disease, and intracardiac abscess are predictive of worse clinical outcomes (1, 3, 8, 13, 65, 205). Among these, staphylococcal infection (5) and complicated PVE (78) are the strongest indicators of poor outcomes. Patients with complications require aggressive management, including antibiotic therapy and, often, early surgery (30).

PREVENTION

The utility of antimicrobial prophylaxis for prevention of bacteremia and IE in humans remains unclear (30, 206). Since the relaxation of antibiotic prophylaxis after the 2007 guideline revisions, no incremental increase in the IE incidence was observed (207, 208), with the exception of a recent ecological study showing increased rates of Streptococcus but not staphylococcal IE cases since the guideline change (209). According to current guidelines, antimicrobial prophylaxis is recommended for high-risk patients (30, 33, 71, 210–213) undergoing high-risk procedures (30), including those with prosthetic heart valves. Probably as important as antibiotic prophylaxis is careful attention to skin and dental hygiene. Moreover, procedures affecting gastrointestinal, respiratory, musculoskeletal, genitourinary, and dermatological systems do not typically require antimicrobial prophylaxis unless they are invasive (30).

CHALLENGES AND FUTURE PERSPECTIVES

MRSA is an increasingly common cause of PVE and continues to be among the most morbid infections in the modern era. Recent data have changed our approach to this disease. Multidisciplinary teams can improve outcomes (200, 214), and specialized centers are an important aspect of optimizing care. Moreover, both nuclear imaging modalities and molecular techniques that show promise in improving PVE diagnostic sensitivity are emerging. Finally, a number of newer agents that are active against MRSA have recently been approved and have shown early efficacy in the management of this condition. However, because prospective and randomized clinical data on the management of MRSA PVE remain scarce, recently updated international practice guidelines for IE from the ESC and the American College of Cardiology (30) continue to rely largely on nonhuman data and expert opinion. As such, important priorities for the field include (i) validating newer diagnostic modalities, such as advanced cardiac imaging, that may increase the sensitivity and specificity of PVE diagnosis; (ii) clarifying the role and optimal timing of valve replacement; (iii) performing comparative effectiveness studies to assess newer alternatives to vancomycin with gentamicin and rifampin as the mainstays of therapy; and (iv) clarifying the role of aminoglycosides in MRSA PVE therapy.

Biographies

Alicia Galar, Pharm.D., Ph.D., studied Pharmacy at the University of Navarra and did both her residency in Clinical Microbiology and Parasitology and Ph.D. at the Clínica Universidad de Navarra. She was a Postdoctoral Clinical Research Fellow/Specialist at Brigham and Women’s Hospital and Harvard Medical School. Dr. Galar worked at the World Health Organization (WHO) Collaborating Centre for Surveillance of Antimicrobial Resistance and joined the Transplant and Oncology Team at the Infectious Diseases Division of Brigham and Women’s Hospital/Dana-Farber Cancer Institute/Massachusetts General Hospital. She completed her education with several courses at the Harvard School of Public Health, Massachusetts General Hospital, and Harvard University. She is currently working at the Department of Clinical Microbiology and Infectious Diseases of the Hospital General Universitario Gregorio Marañón in Madrid. Dr. Galar’s main research interests include antimicrobial resistance, therapeutic drug monitoring, pharmacokinetic/pharmacodynamic (PK/PD) strategies, antimicrobial stewardship, vaccines, prosthetic-device-related infections, infective endocarditis, and infections in patients with heart diseases.

Ana A. Weil holds an M.P.H. from the Johns Hopkins School of Public Health and an M.D. at Tufts University. She trained in internal medicine and infectious diseases at Massachusetts General Hospital (MGH), where she also served as a chief resident. Dr. Weil is currently a physician-scientist in the Infectious Diseases Division at MGH. The Weil laboratory is focused on understanding the influence of the gut microbiome in susceptibility to enteric infections and pathogen-gut microbe interactions at the mucosal surface, including identifying gut species that may protect against infection and relationships between gut microbial species and mucosal immune responses. Dr. Weil also treats general medicine and infectious disease patients at MGH and has held several leadership positions in medical education, including as a faculty advisor for an elective course at Harvard Medical School.

David M. Dudzinski studied Medicine at Harvard Medical School and trained at Massachusetts General Hospital, where he was a chief resident. He is board certified in internal medicine, cardiovascular diseases, nuclear cardiology, adult comprehensive echocardiography, and critical care medicine. Dr. Dudzinski is a cardiologist, cardiac intensivist, and echocardiographer at Harvard Medical School and Massachusetts General Hospital and serves as a staff cardiac intensivist in the cardiac and cardiac surgical intensive care units, with a clinical expertise centered on the nascent field of critical care cardiology. Dr. Dudzinski’s academic interests are in pulmonary embolism, right ventricle function, procedural echocardiography, aortic disease, quality improvement, medical education, resuscitation science and emergency cardiovascular care, and critical care cardiology. Dr. Dudzinski is also an attorney with expertise in the scientific and legal aspects of the regulation of protein-based therapeutics, and he serves on the American College of Cardiology’s national medical professional liability working group.

Patricia Muñoz, M.D., Ph.D., studied Medicine at the Complutense University of Madrid (UCM) and trained at the Hospital Clínico San Carlos. She was deputy editor of Clinical Microbiology and Infection and was awarded the Young Investigator Award of the European Society of Clinical Microbiology and Infectious Diseases. She is currently Professor in Medicine in Clinical Microbiology at UCM and Head of the Section of Clinical Microbiology and Infectious Diseases at the Hospital General Universitario Gregorio Marañón (HGUGM). Dr. Muñoz’s research interests include fungal infections; infective endocarditis; infections in solid-organ transplant recipients, immunocompromised hosts, and heart surgery patients; and nosocomially acquired infectious diseases. She is an active member of the European Study Groups for Nosocomial Infections and Infection in Compromised Host and the Spanish Network of Infection in Transplantation. She is the Secretary of the Group for the Management of Infective Endocarditis at HGUGM and has been President of the Spanish Society for Cardiovascular Infections.

Mark J. Siedner is an Associate Professor of Medicine at Harvard Medical School and holds an M.D. from the Johns Hopkins University School of Medicine and an M.P.H. from the Johns Hopkins Bloomberg School of Public Health. He trained in internal medicine at Columbia University Medical Center and in infectious diseases at Massachusetts General Hospital (MGH). His clinical work focuses on clinical infectious disease both at MGH and as an HIV care provider in southwestern Uganda and Kwazulu-Natal, South Africa. He leads a research program in Uganda and South Africa in partnership with MGH and the Africa Health Research Institute, aimed at mitigating the causes of morbidity and mortality among people living with HIV in low-income countries.