INTRODUCTION — Prosthetic valve endocarditis (PVE) refers to infection of one or more prosthetic heart valves [1-4]. The timing of the infection after surgical valve replacement reflects different pathogenic mechanisms that, in turn, influence the clinical presentation.
The pathogenesis, epidemiology, microbiology, pathology, clinical manifestations, and diagnosis of PVE (both surgical aortic valve replacement and transcatheter aortic valve implantation [TAVI]) will be reviewed here.
Issues related to management of PVE (including antimicrobial therapy and surgery) are discussed separately, as are issues related to prevention. (See "Antimicrobial therapy of prosthetic valve endocarditis" and "Surgery for prosthetic valve endocarditis" and "Prevention of endocarditis: Antibiotic prophylaxis and other measures".)
General issues related to TAVI are discussed separately. (See "Indications for valve replacement for high gradient aortic stenosis in adults", section on 'Choice of surgical or transcatheter intervention' and "Transcatheter aortic valve implantation: Periprocedural and postprocedural management" and "Transcatheter aortic valve implantation: Complications".)
DEFINITIONS AND PATHOGENESIS — Postsurgical PVE can be early (≤12 months postoperatively) or late (>12 months postoperatively) [5-9]. The timing of the infection reflects different pathogenic mechanisms that, in turn, influence the epidemiology, microbiology, pathology, and clinical manifestations of the infection.
Infection engrafted on a transcatheter aortic valve replacement occurs with greatest frequency during the initial year after placement, but studies have not correlated time of onset with pathology or microbiology.
Early infection — In early PVE, particularly within six months of surgery, microorganisms reach the prosthetic valve via direct intraoperative contamination or via hematogenous spread; this time window includes the period of surgical admission as well as the period of postoperative rehabilitation following hospital discharge. (See 'Health care-associated infection' below.)
Early after valve implantation, the valve sewing ring, cardiac annulus, and anchoring sutures have not yet become covered with endothelium; therefore, organisms have direct access to the prosthesis-annulus interface and to paravalvular tissue along suture pathways. These structures are coated with host proteins, such as fibronectin and fibrinogen, to which organisms can adhere. Paravalvular abscesses are particularly common with prosthetic valves because the annulus is commonly the primary site of infection involving both mechanical and bioprosthetic valves, especially in early PVE [10].
Late infection — The pathogenesis of late PVE has been postulated to resemble native valve endocarditis (NVE). (See "Pathogenesis of vegetation formation in infective endocarditis".)
Beginning six months after surgery, the pathogenesis of infection shifts to that postulated for late infection. As the sewing ring, sutures, and adjacent tissues become endothelialized following valve replacement, alterations in the surface and flow characteristics of valve leaflets may facilitate deposition of microthrombi (comprised of platelets and fibrin) on the bioprosthetic leaflets and the anchoring stent of bioprosthetic and mechanical valves. These microthrombi serve as hospitable surfaces for organisms to adhere. The pathogens associated with late PVE tend to be bacteremic isolates able to survive serum bactericidal activity and adhere to these microthrombi, and are similar to those inducing NVE. (See "Native valve endocarditis: Epidemiology, risk factors, and microbiology", section on 'Microbiology'.)
With time after surgery, the paravalvular tissues become endothelialized and are somewhat protected from infection. Therefore, unless the infecting organism is Staphylococcus aureus or another highly virulent or invasive pathogen, the paravalvular tissues are less likely to be affected in late PVE. Accordingly, late-onset infections are less often complicated by paravalvular abscess and valve dehiscence, and are more commonly restricted to the sewing ring or the bioprosthetic leaflet.
EPIDEMIOLOGY
General principles — PVE represents 20 percent of all cases of endocarditis; it occurs in 1 to 6 percent of patients with valve prostheses [11], with an incidence of 0.3 to 1.2 percent per patient-year [4,11,12]. In general, PVE in patients who have undergone surgical valve replacement occurs with equal frequency at aortic and mitral sites [6,7,13-16].
The epidemiology of aortic valve PVE depends on whether the valve replacement was a surgical aortic valve replacement (SAVR) or a transcatheter aortic valve implantation (TAVI). (See "Indications for valve replacement for high gradient aortic stenosis in adults", section on 'Choice of surgical or transcatheter intervention'.)
Risk factors for endocarditis are discussed further separately. (See "Native valve endocarditis: Epidemiology, risk factors, and microbiology".)
Surgical replacement — Patients who undergo SAVR are at risk of developing PVE. In data from Danish national registries between 1996 and 2015, the cumulative incidence of endocarditis among patients following SAVR was 6.0 per 1000-patient years (PY). This is less than the incidence of infective endocarditis in individuals with a prior history of endocarditis (16.1 out of 1000 PY), but greater than the incidence of native valve endocarditis in non-valve replacement matched controls (hazard ratio [HR] 19.1, 95% CI 15.0-24.4) [9].
In early studies, the frequency of PVE during the initial postoperative year among patients with a mechanical valve is similar to or greater than the frequency of PVE among patients with a bioprosthetic valve; with increasing time after valve implantation, PVE may be slightly more common among patients with a bioprosthetic valve [5,6,13]. In several early studies, the cumulative percentage of patients with a bioprosthetic valve who developed PVE ranged from 1 to 3 percent during the initial postoperative year and 3 to 6 percent by five years after valve replacement [5,7,8,13,14,17].
Subsequent studies have demonstrated that the risk of infection is higher for bioprosthetic valves than mechanical valves. In three randomized trials including more than 1400 patients with 8 to 20 years of follow-up, rates of PVE were nominally (but not significantly) higher in patients receiving bioprosthetic versus mechanical valves [18-20].
In data from the Swedish National Patient Registry for patients age 50 to 69 years old who underwent aortic valve replacement between 1997 and 2012, the rate of PVE was higher among recipients of bioprosthetic valves (8.6 percent; followed for a mean of 5.0 years) than among recipients of mechanical valves (7.3 percent; followed for a mean of 8.8 years) [21]. In addition, the Swedish data are notable for a greater incidence of PVE among recipients of bioprosthetic than mechanical valves at multiple time points (1 year, 2 to 5 years, and 6 to 10 years) (table 1) [22,23].
Similarly, in an American observational study including more than 38,000 patients ≥65 years of age with prosthetic valves implanted from 1991 to 1999, the cumulative risk of endocarditis at 12-year follow-up was higher among those with bioprosthetic valves than those with mechanical valves (2.2 versus 1.4 percent) at 12-year follow-up (adjusted HR 1.60, 95% CI 1.31-1.95) [24].
Transcatheter aortic valve implantation — Factors that may confer risk for TAVI-PVE include extensive implanted foreign material, residual paravalvular regurgitation following implantation, and mucosal injury associated with device placement.
The median time from implantation to onset of PVE symptoms in one registry including 250 TAVI-PVE episodes was 5.3 months (interquartile range 1.5 to 13.4 months) [25]. In another report including 53 TAVI-PVE episodes, 76 percent of episodes occurred within one year after valve placement [26].
In data from prospective clinical trials and national registries (in which propensity scoring or multivariable analysis was used to overcome differences in baseline patient characteristics), the incidence rates of TAVI-PVE are similar to the incidence rates of SAVR-PVE with a bioprosthetic valve [27-30]. Rates of PVE for TAVI and SAVR are highest during the initial year after placement and decrease over time thereafter (table 2) [29,30].
In a meta-analysis including 10 randomized trials comparing TAVI and SAVR, the overall incidence of PVE was similar in patients who underwent TAVI or SAVR at one year (representing early PVE) as well as at a mean follow-up of 2 and 3.4 years (representing late PVE) [31]; among patients with intermediate surgical risk, there was a trend toward increased risk of PVE among those who underwent TAVI (2.3 versus 1.2 percent; odds ratio 1.92, 95% CI 0.99-3.77). There was no difference in incidence of PVE among patients who underwent TAVI with a self-expanding valve (SEV), TAVI with a balloon expendable valve (BEV), or SAVR. Similarly, in a registry including more than 6200 patients who underwent TAVI, the cumulative incidence of PVE at one year after implantation for SEV and BEV was 0.95 and 1.25 percent, respectively [32].
The incidence of TAVI-PVE and SAVR-PVE is similar among older patients with severe aortic stenosis and intermediate operative risk. In a propensity-matched comparison of more than 1000 such patients who underwent TAVI with a BEV or SAVR, the rates of PVE at one year were 0.8 and 0.7 percent, respectively [33]. Similarly, in a trial including more than 280 patients at lower surgical risk managed with a self-expanding bioprosthesis by TAVI or a bioprosthesis by SAVR, the cumulative five-year incidences of PVE was 6.2 and 4.4 percent, respectively [34].
In contrast to the aforementioned findings, a large retrospective cohort study using linked databases (without propensity score matching) found an overall infective endocarditis incidence of 4.81 (95%CI 4.61-5.03) per 1000 PY among patients undergoing SAVR compared with 3.57 (95% CI 3.00-4.21) per 1000 PY among patients undergoing TAVI. In a multivariable analysis, SAVR was an independent predictor of infective endocarditis. The patients who underwent SAVR had a longer follow-up period (53.9 versus 24.5 months). More than one-third of patients who underwent SAVR underwent concomitant coronary artery bypass grafting and 7.7 percent had more than one valve intervention [35]. In addition, in a review of data from three randomized trials, a higher cumulative incidence of PVE at five years was observed among patients who underwent SAVR versus TAVI [36]. These studies illustrate the challenges of comparing outcomes among patient groups, including differences in baseline characteristics, surgical procedure, and duration of follow-up.
A number of studies have evaluated risk factors for TAVI-PVE using multivariable analysis [23,25,28,30,35,37,38]. Patient factors associated with TAVI-PVE include male sex, chronic renal disease (creatinine clearance <30 mL/min/1.75m2), pulmonary disease, cirrhosis, and endocarditis within the prior year. Procedure factors include postprocedure aortic regurgitation (moderate to severe), need for cardiac electrical device placement, hospital complications (such as cardiac arrest, major bleeding, and sepsis), low valve placement, and transapical access. In addition, concerns have been raised about nonstandardized preprocedure antibiotic prophylaxis and variable sterility in implant facilities.
Health care-associated infection — In addition to the typical nosocomial nature of early onset PVE, some cases of PVE (including TAVI-PVE) are health care-associated (in that they result from infection acquired in outpatient health care settings or in the context of ongoing invasive care).
In one cohort study including more than 550 patients with PVE, health care-associated infection (non-nosocomial) was defined as PVE diagnosed within 48 hours of admission in a patient with extensive nonhospital health care contact (as defined as follows) [4]:
●Intravenous therapy, wound care, or specialized nursing care at home or intravenous chemotherapy within the prior 30 days
●Residence in a nursing home or other long-term care facility
●Hospitalization in an acute care hospital for two or more days within the prior 90 days
●Attendance at a hospital or hemodialysis clinic within the prior 30 days
Health care-associated infection was observed in 37 percent of cases; of these, 70 percent were nosocomial and 30 percent were acquired in an outpatient context. Approximately 70 percent were diagnosed within the first year after valve implantation; the majority occurred within the first 60 days. S. aureus was the most common pathogen, identified in 34 percent of cases.
In patients with prosthetic valves, nosocomial bacteremia is associated with significant risk for seeding the prosthesis. In one study including 115 prosthetic valve recipients with nosocomial bacteremia judged not to be the sentinel event of endocarditis, PVE due to the bacteremic organism developed in 16 percent of cases, between 7 and 170 days thereafter (median interval 28 days) [39]. Similarly, in another study including 37 patients with prosthetic valves and postoperative candidemia without evidence of endocarditis, fungal endocarditis developed in 11 percent of cases, between 26 and 690 days later [40]. The patients who developed candida PVE had persistent fungemia (mean 8.1 days) without evidence of endocarditis during the month after cardiac surgery.
MICROBIOLOGY — The microbiology of PVE involving surgically implanted valves is relatively predictable, depending on the time since implantation (table 3) [4,25,26,41-45]:
●During the initial two months of implantation, the most frequently encountered pathogens were S. aureus and coagulase-negative staphylococci (CoNS); next in frequency were gram-negative bacilli and Candida species. This spectrum of organisms reflects the typical nosocomial origin of these infections.
●Between 2 and 12 months after implantation, the most frequently encountered pathogens were coagulase-negative staphylococci, S. aureus, and streptococci, followed by enterococci. In general, cases occurring 2 to 12 months after surgery are a blend of delayed-onset nosocomial and community-acquired infections.
●Beyond 12 months after implantation, the most frequently encountered pathogens were streptococci and S. aureus, followed by coagulase-negative staphylococci and enterococci. In general, the range of pathogens is similar to that of native valve endocarditis (NVE) in patients who are not injection drug users. This is because late PVE, like NVE, usually results from transient bacteremia occurring among ambulatory patients. (See "Native valve endocarditis: Epidemiology, risk factors, and microbiology", section on 'Microbiology'.)
●Culture-negative PVE occurs in all time intervals after surgery. (See 'Culture-negative endocarditis' below.)
●Sporadic cases of PVE due to a variety of other bacteria, fungi, and nontuberculous mycobacteria (some related to intraoperative exposure to aerosolized organisms due to heater-cooler contamination by Mycobacterium chimaera; others related to contamination in the manufacture of bioprostheses) have also been reported. One case of infective endocarditis due to enterovirus has been described [46].
The above descriptions of PVE microbiology are illustrated by observations of PVE due to CoNS. Among patients with PVE due to CoNS during the initial year after surgery, the cause is almost exclusively Staphylococcus epidermidis, of which most are methicillin resistant (reflecting the likely nosocomial origin); in contrast, among patients with PVE due to CoNS more than one year after surgery, almost half are non-epidermidis species, and most are methicillin susceptible (reflecting the likely community origin) [13,47]. (See "Infection due to coagulase-negative staphylococci: Epidemiology, microbiology, and pathogenesis" and "Infection due to coagulase-negative staphylococci: Treatment".)
Based on available data thus far, the microbiology of transcatheter aortic valve implantation (TAVI)-PVE is generally similar to that of surgical PVE; however, one difference is the increased frequency of enterococci causing TAVI-PVE (table 3). It is uncertain whether this reflects use of femoral vascular access, increased likelihood of a genitourinary portal of entry in older adults, or other factors.
PATHOLOGY — The intracardiac pathology of infection involving surgically placed valves, particularly when PVE presents during the early months after surgery or when it is caused by invasive organisms, shapes the requirement for therapy. Paravalvular invasion, commonly with associated dehiscence of the prosthesis and paravalvular regurgitant flow, occurs in approximately 40 percent of patients and frank extension into tissue causing myocardial abscess is seen in 15 percent [48,49]. In one series, invasion of paravalvular tissue was noted in more than 80 percent of cases [50].
In some patients, infection of an aortic valve prosthesis extends through the annulus to cause pericarditis or, more commonly, into the membranous portion of the interventricular septum where it disrupts the conduction system, resulting in various degrees of heart block [51-53]. Large vegetations may prevent closure of the prosthesis producing incompetence or encroach upon the valve orifice causing functional stenosis.
Bioprosthetic valve endocarditis also is associated with invasive infection. Annular and myocardial invasion was noted in 38 of 85 patients (45 percent) in one study and was more frequent among bioprosthetic PVE occurring in the first year after valve replacement than in cases presenting later (59 versus 25 percent) [54]. Similarly, in another series, invasive disease was more common in patients with early compared with later bioprosthetic PVE (79 versus 31 percent) [55].
The histologic features that characterize PVE in bioprosthetic valves are not well defined. As bioprosthetic valves degenerate, they may form noninfective, calcific, vegetative-like lesions with inflammatory infiltrates, thus potentially causing a noninfectious process that can be misdiagnosed as PVE. In one study of the histopathology of bioprosthetic valve tissue removed at surgery from 88 patients (21 for suspected endocarditis and 67 for noninfective dysfunction), PVE was characterized histologically by surface vegetations with microorganisms and neutrophil-rich inflammatory infiltrates; in the non-infected group, valve pathology was characterized by calcific inflammatory infiltrates consisting mainly of macrophages and lymphocytes [56].
Thus far, the pathology of transcatheter aortic valve implantation (TAVI)-PVE has been largely inferred from transesophageal echocardiography and other cardiac imaging, given relatively limited autopsy and surgical reports. In one review including 33 patients with TAVI-PVE, echocardiographic and surgical findings indicated invasive intracardiac pathology including abscesses, fistulas, and pseudoaneurysms in 44 percent of cases; among these patients, the median interval from TAVI to PVE was 4 months [57].
In another study including more than 240 patients with TAVI-PVE, echocardiography demonstrated vegetations and paravalvular complications in 68 and 18 percent of cases, respectively [25]. Invasive complications appear similar with infection of self-expanding valve (SEV) and balloon expandable valve (BEV) [26,32]; vegetations were attached to the stent in 18 percent and/or on leaflets in 48 percent of cases. Vegetations were more common on stents of SEV compared with BEV (26 and 11 percent, respectively); vegetations were found more commonly on leaflets of BEV compared with SEV (59 versus 36 percent, respectively) [25]. In addition, mitral valve infection (including leaflet perforation and regurgitation) has been observed in 20 percent of patients with TAVI-PVE patients [25,26].
CLINICAL MANIFESTATIONS — Patients with PVE present with symptoms and signs similar to those encountered in native valve endocarditis (NVE); however, many patients with PVE present with nonspecific symptoms. Clinical manifestations of PVE include fever, chills, anorexia, and weight loss. Other symptoms include malaise, headache, myalgias, arthralgias, night sweats, abdominal pain, dyspnea, cough, and pleuritic pain.
In patients with PVE, the frequency of new or changing murmurs, heart failure, and new electrocardiographic conduction disturbances is higher than in patients with NVE [4,54,58]. In addition, patients with early PVE may present with postoperative manifestations that are more prominent than features of endocarditis. Supportive signs of infective endocarditis include cutaneous manifestations such as petechiae or splinter hemorrhages. (See "Clinical manifestations and evaluation of adults with suspected left-sided native valve endocarditis", section on 'Symptoms and signs'.)
Systemic complications of PVE include cardiac and neurologic complications, septic emboli, metastatic infection, and systemic immune reactions. Clinical manifestations reflecting these complications may be present at the time of initial presentation and/or may develop subsequently. The incidence of clinically overt arterial emboli is 40 percent; central nervous system complications, primarily embolic infarcts or hemorrhages, occur in 20 to 40 percent of cases [50,59-61]. Clinical manifestations of a complication warrant independent diagnostic evaluation, concurrent with evaluation for PVE. (See "Clinical manifestations and evaluation of adults with suspected left-sided native valve endocarditis", section on 'Complications as presenting symptoms' and "Complications and outcome of infective endocarditis".)
Manifestations of invasive infection include valvular dysfunction, persistent fever ≥9 days despite appropriate antibiotic therapy, new electrocardiographic conduction disturbances, and echocardiographic evidence of abscess formation [58,62]. In one study including more than 100 patients with PVE after surgical valve replacement, these findings were observed in 64 percent of patients; they occurred more frequently in the setting of aortic valve involvement and in the first year after valve replacement [58].
In transcatheter aortic valve implantation (TAVI)-PVE, rates of heart failure range from 22 to 60 percent [25,26]. Older patients with TAVI-PVE may present with nonspecific symptoms (anorexia, weight loss, and fatigue) and fever may be blunted. Thus far, clinical manifestations of invasive infection noted in surgical PVE (described above) have not been similarly correlated in patients with TAVI-PVE; however, echocardiography and other imaging have demonstrated that invasive infection does occur with significant frequency in these patients. (See 'Pathology' above.)
DIAGNOSIS
Overview of diagnostic approach — The diagnosis of PVE should be suspected in patients with history of valve replacement who present with any of the following:
●Bacteremia with an organism commonly causing PVE (see 'Microbiology' above)
●Persistent unexplained bacteremia with an organism uncommonly associated with PVE
●Persistent nonspecific symptoms (fever, chills, anorexia, weight loss) in the absence of bacteremia
●New prosthetic valve dysfunction, particularly regurgitation, even in the absence of other signs of infection
The diagnosis of PVE is established based on clinical manifestations, blood cultures (or other microbiologic data), and echocardiography. The accepted criteria for diagnosis of infective endocarditis are the modified Duke criteria, which are summarized in the tables (table 4 and table 5) [63]. However, the sensitivity of these criteria for diagnosis of PVE is lower than their sensitivity for native valve endocarditis (NVE). Therefore, in the setting of persistent clinical suspicion for PVE but 'possible' or 'rejected' endocarditis based on the modified Duke criteria, additional cardiac imaging should be pursued if feasible [2]. (See 'Diagnostic (modified Duke) criteria' below and 'Cardiac imaging' below.)
At least three sets of blood cultures should be obtained from separate venipuncture sites prior to initiation of antibiotic therapy. For patients who are clinically stable, antimicrobial therapy may be deferred while awaiting the results of blood cultures and other diagnostic tests. For patients with signs of clinical instability or sepsis, initiation of empiric antimicrobial therapy (after blood cultures have been obtained) is appropriate. Follow-up blood cultures should be obtained 48 to 72 hours after antimicrobial therapy is begun and repeated every 48 to 72 hours until clearance of bacteremia is documented. (See 'Blood cultures' below and "Detection of bacteremia: Blood cultures and other diagnostic tests" and "Antimicrobial therapy of prosthetic valve endocarditis".)
Echocardiography should be performed in all patients with suspected PVE (algorithm 1) [1,64-66]. Transthoracic echocardiography (TTE) is often the initial study; however, transesophageal echocardiography (TEE) has higher sensitivity than TTE for both the diagnosis and the detection of paravalvular extension of infection. Accordingly in the absence of contraindications, TEE should be pursued when TTE findings are nondiagnostic as well as to assess paravalvular infection. If initial TEE is negative or indeterminate and clinical suspicion for PVE persists, repeat TEE should be pursued (5 to 7 days later). (See 'Echocardiography' below.)
In patients for whom the diagnosis of PVE remains uncertain following TEE, and/or in circumstances where TEE is not feasible, electrocardiogram-gated cardiac computed angiography (CTA) and 18F-fluorodeoxyglucose positron emission computed tomography (18F-FDG PET/CT) may be useful imaging tools. (See 'Additional imaging tools' below.)
Additional diagnostic evaluation for patients with suspected or known PVE includes electrocardiography (ECG), chest radiography, and additional radiographic imaging tailored to clinical manifestations. (See "Clinical manifestations and evaluation of adults with suspected left-sided native valve endocarditis", section on 'Overview of diagnostic approach'.)
●ECG – Baseline ECG should be performed as part of the initial evaluation for all patients with suspected infective endocarditis, with subsequent telemetry monitoring or serial electrocardiograms. The presence of heart block or conduction delay (which may manifest initially as a prolonged PR interval) may provide an important clue to paravalvular extension of infection to the valve annulus and adjacent septum (which should prompt further evaluation with echocardiography as discussed below). In addition, the presence of findings consistent with ischemia or infarction may suggest the presence of emboli to the coronary circulation. (See "ECG tutorial: Basic principles of ECG analysis".)
●Chest radiography – Chest radiography is warranted to evaluate for an infiltrate, congestive heart failure, and potential alternative causes of fever and systemic symptoms.
●Computed tomography (CT) – CT of the torso (chest, abdomen, and pelvis) may be used to evaluate for sites of metastatic infection (such as splenic infarct, renal infarcts, psoas abscess, or other sites of infection) that may warrant localized drainage [1,67]. However, given risk for nephrotoxicity, the decision to pursue this imaging should be guided by a careful history and clinical assessment.
●Additional radiographic imaging to evaluate for complications of PVE should be tailored to findings on history and physical examination [1]. As examples, patients with back pain should be evaluated for vertebral osteomyelitis with magnetic resonance imaging (MRI), and patients with headache, neurologic deficits, or meningeal signs should be evaluated with head MRI/magnetic resonance angiogram for neurologic complications (including intracranial mycotic aneurysm or central nervous system bleeding). (See "Overview of infected (mycotic) arterial aneurysm".)
Routine brain imaging with CT or MRI is not standard of care in the absence of focal neurologic signs or symptoms. However, as asymptomatic cerebral infarcts are quite common in patients with IE; the presence of such lesions can be diagnostically important if they are detected in patients with other clinical findings suggestive of infective endocarditis [1,68-70]. For example, in one prospective study of MRI in 130 patients with suspected endocarditis, of whom only 16 had acute neurologic symptoms, MRI findings led to upgraded classification of infective endocarditis to definite or possible in 32 percent of patients with non-definite endocarditis and to a modification of planned therapy in 18 percent of cases [68].
Lastly, patients with PVE should undergo a thorough dental evaluation; the examination should focus on periodontal inflammation, pocketing around teeth, and caries that may result in pulpal infection and subsequent abscess [1]. All active sources of oral infection should be eradicated, and patients should be counseled regarding the importance of daily dental hygiene with serial dental evaluation. (See "Antimicrobial therapy of left-sided native valve endocarditis", section on 'Follow up'.)
Diagnostic (modified Duke) criteria — The accepted criteria for diagnosis of infective endocarditis are the modified Duke criteria, which are summarized in the tables (table 4 and table 5) [63]. (See "Clinical manifestations and evaluation of adults with suspected left-sided native valve endocarditis", section on 'Diagnostic (modified Duke) criteria'.)
However, the sensitivity of the modified Duke criteria for diagnosis of PVE is lower than the sensitivity for diagnosis of NVE, given limitations of echocardiography, including TEE. In two studies of pathologically confirmed cases of PVE involving surgically implanted valves, assessed by the Duke criteria, 76 percent were categorized as definite PVE and 24 percent as possible PVE [71,72].
For this reason, we are in agreement with the 2015 guidelines issued by the European Society of Cardiology, which propose use of additional imaging techniques, particularly imaging focused on intracardiac pathology, in the setting of persistent clinical suspicion but 'possible' or 'rejected' endocarditis based on the Duke criteria [2]. (See 'Additional imaging tools' below.)
Diagnostic tools
Blood cultures
General principles — At least three sets of blood cultures should be obtained from separate venipuncture sites prior to initiation of antibiotic therapy [73,74]. In the absence of prior antibiotic therapy, blood cultures will be positive in ≥90 percent patients with PVE. Because bacteremia is continuous, cultures will be positive regardless of whether or not blood cultures are obtained in proximity to the fever. When all or most blood cultures drawn over a period of hours to days in a patient with a prosthetic valve are positive, PVE is highly probable.
The duration of documented bacteremia is particularly important when the isolate is an organism that is commonly considered a contaminant, such as coagulase-negative staphylococci or diphtheroids. A high rate of blood culture positivity, or molecular evidence that a sporadically isolated organism represents a single clone, helps to distinguish infecting pathogens from contaminants [75].
Blood cultures for diagnosis of endocarditis are discussed further separately. (See "Clinical manifestations and evaluation of adults with suspected left-sided native valve endocarditis", section on 'Blood cultures'.)
Culture-negative endocarditis — Culture-negative endocarditis is defined as endocarditis with no definitive microbiologic etiology following inoculation of at least three independently obtained blood samples in a standard blood-culture system, with negative cultures after five days of incubation and subculturing.
If antibiotics have not been administered prior to obtaining blood cultures, it is unusual to have persistently negative blood cultures in patients with PVE caused by the commonly encountered bacteria causing endocarditis. Nevertheless, culture-negative PVE can occur when infection is caused by fastidious organisms such as Legionella species, Bartonella species, Coxiella burnetii, Mycoplasma hominis, and fungi. (See "Endocarditis caused by Bartonella".)
Rare cases of blood culture-negative PVE (and other cardiac surgery related focal infections) have been caused by M. chimaera, a non-tuberculosis mycobacterium. Contamination of heater-cooler units used during cardiac surgery has resulted in exposure of patients during cardiac surgery to aerosolized organisms [76-78]. Onset of these infections after cardiac surgery has often been delayed and indolent [78]. (See "Overview of nontuberculous mycobacterial infections", section on 'M. chimaera associated with cardiac surgery'.)
Detection of these and other unusual fastidious pathogens causing PVE relies upon the same evaluation used to assess culture-negative NVE. Next-generation sequencing of pathogen deoxyribonucleic acid (DNA) from plasma is an emerging technology that may facilitate rapid detection of fastidious pathogens causing infective endocarditis [79].
Cardiac imaging
Echocardiography — Echocardiography is an important tool for diagnosis of PVE. TTE is often the initial study and provides useful information regarding cardiac function. However, TEE is the study of choice for detection of vegetations as well as other complications including abscess, fistula, leaflet perforation, pseudoaneurysm, and paraprosthetic leak [80]; therefore, TEE should always be performed in patients with suspected PVE (in the absence of contraindications):
●Following surgical valve replacement; the sensitivity of TEE is greater than that of TTE (86 to 92 percent versus 17 to 52 percent, respectively), particularly for assessing mitral valve prosthesis or paravalvular complications [81-89]. In one study including more than 100 patients with surgical valve replacement and suspected endocarditis (34 PVE, 80 NVE) who underwent TTE and TEE, the results were concordant in 55 percent of cases; TEE results led to a reclassification of the case more frequently among patients with prosthetic valves versus native valves (34 versus 11 percent), including 10 patients reclassified as definite PVE [87].
●Following transcatheter aortic valve implantation (TAVI), the sensitivity of echocardiography (including TEE) is limited due to artifacts and shadowing caused by the metal stents anchoring the valve; alternative imaging modalities may be needed for assessment of PVE. (See 'Additional imaging tools' below.)
The negative predictive value of the modified Duke criteria, including a complete echocardiographic evaluation, in a patient with PVE may be as low as 60 to 65 percent [88,89]. If PVE is strongly suspected in the setting of negative echocardiography, repeat echocardiography (5 to 7 days later) may establish a diagnosis of PVE in some cases [83,90]. In one study among patients with suspected PVE, repeat TEE resulted in a shift from possible to definite endocarditis in 42 percent of cases; there was no diagnostic benefit beyond a third TEE [91]. If the diagnosis of PVE remains uncertain, other imaging technologies, if available, should be considered. (See 'Additional imaging tools' below.)
Issues related to echocardiography for diagnosis of endocarditis are discussed further separately. (See "Clinical manifestations and evaluation of adults with suspected left-sided native valve endocarditis", section on 'Echocardiography'.)
Additional imaging tools — The most useful alternative cardiac imaging tools for diagnosis of PVE include CTA and 18F-FDG PET/CT [2]; thus far, other imaging modalities (including cardiac MRI, radiolabeled autologous white blood cell single-photon emission CT [SPECT/CT]) appear to be less useful.
CTA or 18F-FDG PET/CT may be useful in patients for whom the diagnosis of PVE remains uncertain following TEE, and/or in circumstances where TEE is not feasible. The use of such imaging tools depends in large part on their local availability and the expertise of local staff and is often guided by the diagnostic or management questions that remain after traditional echocardiographic imaging has been performed.
Cardiac CTA may be helpful for cases in which definitive evidence of PVE and its complications cannot be fully assessed with TEE, or in planning a surgical strategy for patients with extravalvular complications. In general, TEE is superior to CTA for detection of vegetations (especially small ones) or valve perforations, whereas CTA is comparable or superior to TEE for detection of paravalvular infection, abscess, or pseudoaneurysm [92-95]. CTA may also provide satisfactory coronary imaging for patients at intermediate risk of coronary artery disease who require surgical intervention [94,95].
18F-FDG PET/CT is useful for diagnosis of select cases of PVE in which echocardiography is not diagnostic. With this modality, uptake of positron-labeled glucose by inflammatory leukocytes allows anatomic localization of infection. The sensitivity of this modality likely diminishes as inflammation resolves with administration of antimicrobial therapy; in addition, images must be reviewed carefully to distinguish between PVE and non-infectious post-surgical inflammation [96]:
●In one study including more than 90 patients (with indwelling prosthetic valve[s] and/or other intracardiac devices) with suspected endocarditis who were evaluated with TEE, 18F-FDG PET/CT, and CTA, patients were initially classified according to the modified Duke criteria [88]. Addition of 18F-FDG PET/CT to the modified Duke criteria as an additional major criteria was associated with increased diagnostic sensitivity for PVE (from 52 to 91 percent) with little loss in specificity (from 95 to 90 percent), and with an increase in the negative predictive value (from 60 to 88 percent). In a subgroup of patients who underwent both 18F-FDG PET/CT and CTA, the classification was improved further by CTA detection of abscesses, pseudoaneurysms, and fistulas.
●In another study including more than 180 patients with suspected PVE, 18F-FDG PET/CT findings were used as an additional major criterion in the modified Duke schema [89]. Inclusion of the 18F-FDG PET/CT results improved the sensitivity (42 to 91 percent), positive predictive value (74 to 86 percent), and negative predictive value (65 to 93 percent) of the modified Duke criteria for diagnosis of PVE. Of the 62 episodes initially classified as possible PVE, 18F-FDG PET/CT findings allowed reclassification to definite PVE in 76 percent of cases, and increased conclusive diagnoses (definite or rejected) from 67 to 92 percent.
●A prospective study examined the impact of FDG PET/CT on the diagnosis and management of patients with suspected IE, including 70 patients with suspected PVE [97]. FDG PET-CT led to modification of the classification by Duke criteria in 24 percent of patients with PVE, mostly due to perivalvular uptake. Patient management was modified in 21 percent of cases, resulting in a change in antibiotic therapy in 15 of 70 patients and change in cardiac surgery management in 4 of 70 patients. Those who benefitted the most from FDG PET-CT were those with noncontributory baseline echocardiography or initial classification of possible IE.
For patients with suspected TAVI-PVE, use of 18F-FDG PET/CT and/or CTA may be especially beneficial given diminished sensitivity of echocardiography in this context. In one study including 22 patients with possible TAVI-PVE (based on echocardiography), additional imaging with 18F-FDG PET/CT and CTA confirmed or excluded TAVI-PVE in 10 cases, but 12 cases remained classified as possible PVE, requiring a clinical decision regarding treatment [98]. In another report including 16 patients with suspected TAVI-PVE, multimodality imaging increased the sensitivity of the modified Duke criteria from 50 to 100 percent [99].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Treatment and prevention of infective endocarditis".)
INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or email these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)
●Basics topic (see "Patient education: Endocarditis (The Basics)")
SUMMARY
●Prosthetic valve endocarditis (PVE) refers to infection of one or more prosthetic heart valves. PVE can be early (≤12 months postoperatively), or late (>12 months postoperatively). In early PVE, especially that occurring during the initial two months after surgery, microorganisms reach the prosthetic valve via direct intraoperative contamination or via hematogenous spread during the initial days and weeks after surgery; paravalvular abscess is common. In late PVE, the pathogens tend to be bacteremic isolates similar to those inducing native valve endocarditis (NVE). The microbiology of PVE depends on the time since implantation (table 3) (see 'Definitions and pathogenesis' above and 'Microbiology' above).
●PVE represents 20 percent of all cases of endocarditis; it occurs in 1 to 6 percent of patients with valve prostheses, with an incidence of 0.3 to 1.2 percent per patient-year; the risk is greatest during the initial year after implantation. PVE occurs with slightly greater frequency on bioprosthetic than mechanical valves (table 2). (See 'Epidemiology' above.)
●Transcatheter aortic valve implantation PVE (TAVI-PVE) is a new and increasingly frequent form of PVE. Cases primarily present in the initial year after valve replacement. The microbiology of TAVI-PVE is similar to that seen with PVE involving surgically implanted valves, with the exception of an increased frequency of enterococcal infections. The incidence of TAVI-PVE is similar to that of PVE involving surgically implanted aortic bioprosthetic valves. (See 'Transcatheter aortic valve implantation' above.)
●Patients with PVE present with symptoms and signs similar to those encountered in NVE; however, many patients with PVE present with nonspecific symptoms. Clinical manifestations of PVE include fever, chills, anorexia, and weight loss. The frequency of new or changing murmurs, heart failure, and new electrocardiographic conduction disturbances in patients with PVE is higher than in patients with NVE. (See 'Clinical manifestations' above.)
●The diagnosis of PVE should be suspected in patients with history of valve replacement who present with any of the following (see 'Overview of diagnostic approach' above):
•Bacteremia with an organism commonly causing PVE
•Persistent unexplained bacteremia with an organism uncommonly associated with PVE
•Persistent nonspecific symptoms (fever, chills, anorexia, weight loss) in the absence of bacteremia
•New prosthetic valve dysfunction, particularly regurgitation, even in the absence of other signs of infection
●The diagnosis of PVE is established based on clinical manifestations, blood cultures (or other microbiologic data), and echocardiography. The accepted criteria for diagnosis of infective endocarditis are the modified Duke criteria (table 4 and table 5). However, the sensitivity of these criteria for diagnosis of PVE is lower than the sensitivity for NVE; therefore, in the setting of persistent clinical suspicion for PVE but 'possible' or 'rejected' endocarditis based on the modified Duke criteria, additional cardiac imaging should be pursued if feasible. (See 'Overview of diagnostic approach' above and 'Diagnostic (modified Duke) criteria' above and 'Additional imaging tools' above.)
●At least three sets of blood cultures should be obtained from separate venipuncture sites prior to initiation of antibiotic therapy. For patients who are clinically stable, antimicrobial therapy may be deferred while awaiting the results of blood cultures and other diagnostic tests. For patients with signs of clinical instability, initiation of empiric antimicrobial therapy (after blood cultures have been obtained) is appropriate. (See 'Overview of diagnostic approach' above.)
●Echocardiography should be performed in all patients with suspected PVE (algorithm 1). In general, transthoracic echocardiography (TTE) is the first imaging test for patients with suspected PVE; however, transesophageal echocardiography (TEE) has higher sensitivity than TTE and should be pursued in the absence of contraindications. In patients with TAVI, the sensitivity of echocardiography is limited due to artifacts and shadowing caused by the metal stents anchoring the valve. (See 'Echocardiography' above.)
●Additional cardiac imaging tools include 18F-fluorodeoxyglucose positron emission computed tomography (18F-FDG PET/CT) and electrocardiogram-gated cardiac computed angiography (CTA); these tools may be useful in patients for whom the diagnosis of PVE (particularly TAVI-PVE) remains uncertain following echocardiography. With 18F-FDG PET/CT, uptake of positron-labeled glucose by inflammatory leukocytes allows anatomic localization of infection. CTA is useful for detection of paravalvular infection, abscess, or pseudoaneurysm. (See 'Additional imaging tools' above.)
●Additional evaluation for patients with suspected PVE includes electrocardiography, chest radiography, and other radiographic imaging tailored to clinical manifestations. (See 'Overview of diagnostic approach' above.)
