Presentation and diagnosis of prosthetic valve endocarditis

http://www.uptodateonline.com/online/content/topic.do?topicKey=endocard/4909&selectedTitle=37~150&source=search_result

Presentation and diagnosis of prosthetic valve endocarditis

Author
Adolf W Karchmer, MD
Section Editor
Stephen B Calderwood, MD
Deputy Editor
Elinor L Baron, MD, DTMH


Last literature review version 16.2: May 2008 | This topic last updated: May 16, 2007 (More)


INTRODUCTION — Prosthetic valve endocarditis is a serious problem with potentially fatal consequences [1] . (See "Complications and outcome of infective endocarditis", section on Mortality).

The pathogenesis, microbiology, pathology, clinical features and diagnosis of prosthetic valve endocarditis (PVE) will be reviewed here. The antimicrobial and possible surgical treatment and prevention of PVE are discussed separately. (See "Antimicrobial therapy of prosthetic valve endocarditis" and see "Surgery for prosthetic valve endocarditis").

PATHOGENESIS — PVE can arise early or late after surgery. The timing of the infection reflects different pathogenic mechanisms that, in turn, influence the epidemiology, microbiology, pathology and clinical manifestations of the infection.

Early infection — Microorganisms can reach the valve prosthesis by direct contamination intraoperatively or via hematogenous spread during the initial days and weeks after surgery. These pathogens have direct access to the prosthesis-annulus interface and to perivalvular tissue along suture pathways because the valve sewing ring, cardiac annulus, and anchoring sutures are not endothelialized early after valve implantation. These structures are coated with host proteins, such as fibronectin and fibrinogen, to which some organisms can adhere and initiate infection.

Late infection — As the sewing ring, sutures, and adjacent tissues become endothelialized over the months after valve replacement, sites for adherence of microorganisms and access to host tissues adjacent to the prosthesis are altered. The pathogenesis of late PVE has been postulated to resemble native valve endocarditis (NVE). (See "Pathogenesis of vegetation formation in infective endocarditis").

Platelet-fibrin thrombi probably form on the prosthesis and serve as sites for the organisms to adhere and cause infection. Thus, the pathogens producing PVE tend to be bacteremic isolates with the ability to adhere to these thrombi and are similar to organisms inducing NVE. Unless the infecting organism is Staphylococcus aureus or another highly virulent or invasive pathogen, the perivalvular tissues are less likely to be affected in late PVE since endothelialization limits access to these tissues. Later onset infections tend to be less invasive, are less often complicated by perivalvular abscess formation and valve dehiscence, and are more commonly restricted to the sewing ring or the bioprosthetic leaflet.

However, the leaflets of porcine bioprosthetic valves experience age-related alterations in their surface characteristics. These aging leaflets become sites for platelet-fibrin thrombus deposition and subsequent infection. As a result, the infection rates on bioprostheses increase during this later period when compared to mechanical valves [2-4] .

EPIDEMIOLOGY — The risk of developing PVE, typically expressed as a hazard function or the likelihood of infection at a given point in time, is not uniform after valve replacement. The risk is greatest during the initial three months after surgery, remains high through the sixth month, and then falls gradually with an annual rate of approximately 0.4 percent from 12 months postoperatively onward for the life of the valve recipient [3,5-8] . The percentage of patients developing PVE during the initial year after valve replacement ranges from one to three percent in studies with active follow-up; by five years, the cumulative percentage ranges from three to six percent (show table 1).

Although conclusions vary among studies, infection generally occurs with equal frequency at aortic and mitral sites and on mechanical and bioprosthetic devices during the first postoperative year [3-6,9,10] . As noted above, bioprosthetic valves carry a greater risk for infection than mechanical after the first 18 months [2-4] . When a valve is replaced in a patient with endocarditis, whether the infection was active or healed at the time of surgery, there is an increased risk for both early (some of which is recrudescence of active endocarditis) and late PVE [2,4,6,8] .

Healthcare associated infection — Some cases of prosthetic valve endocarditis result from infection acquired in healthcare settings, both inpatient (nosocomial) and outpatient (non-nosocomial). The relative frequency of health care-associated PVE was evaluated in a prospective, observational cohort multinational study of 556 patients with PVE [1] . Non-nosocomial health care-associated infection was defined as PVE diagnosed within 48 hours of admission in an outpatient with extensive healthcare contact, as defined by:

• 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


Healthcare-associated infection was present in 37 percent of cases: 70 percent were labeled as nosocomial and 30 percent as outpatient acquired. Approximately 70 percent of the health care-associated infections were diagnosed within the first year after valve implantation, with the majority occurring within the first sixty days. S. aureus was the most common organism, being identified as the pathogen in 34 percent of cases.

Nosocomial bacteremia occurring in patients with prosthetic valves carries a significant risk for seeding the prosthesis. Among 115 prosthetic valve recipients experiencing nosocomial bacteremia that was judged not to be the sentinel event of endocarditis, 18 (16 percent) developed PVE with the bacteremic organism between seven and 170 days thereafter (median interval 28 days) [11] . Similarly, among 37 prosthetic valve patients with postoperative candidemia without evidence of endocarditis, fungal endocarditis developed in four (11 percent) between 26 and 690 days later [12] . The patients who developed candida PVE had persistent fungemia (8.1 mean days fungemia) without evidence of endocarditis during the month after cardiac surgery. (See "Candida endocarditis").

MICROBIOLOGY — While many different organisms have caused sporadic cases of PVE, the microbiology is relatively predictable when PVE is categorized by time after implantation. The largest reported contemporary experience comes from the multinational study of 556 patients cited above [1] :

• The most frequently encountered pathogens in early PVE (within two months of implantation) were S. aureus (36 percent) and coagulase-negative staphylococci (17 percent); next in frequency were culture-negative (17 percent) and fungal infection (9 percent) (show table 2).

• When PVE presented after two months, the culture results in decreasing order of frequency were coagulase-negative staphylococci and S. aureus (18 to 20 percent each); next in frequency were no organism identified, enterococci, and viridans streptococci (10 to 13 percent each) (show table 2).


Cases occurring more than 12 months after valve surgery are usually caused by the same pathogens that produce NVE in patients who are not injection drug users, since late PVE, like NVE, usually results from transient bacteremia occurring among ambulatory patients. (See "Infective endocarditis: Epidemiology and risk factors", section on Microbiology).

The coagulase-negative staphylococci causing PVE during the initial year after surgery are almost exclusively S. epidermidis. Between 84 and 87 percent of these organisms are methicillin-resistant and thus resistant to all of the beta-lactam antibiotics [3,13] . By contrast, almost half of the coagulase-negative staphylococci causing PVE one year or more after surgery are non-epidermidis species, and only 22 to 30 percent are methicillin-resistant. (See "Microbiology, pathogenesis, and epidemiology of coagulase-negative staphylococci").

The mechanisms leading to methicillin-resistance are the same among coagulase-negative staphylococci as for methicillin-resistant S. aureus (MRSA). (See "Coagulase-negative staphylococci: antimicrobial therapy and resistance" and see "Microbiology of methicillin-resistant Staphylococcus aureus").

However, methicillin-resistant coagulase-negative staphylococci, while uniformly possessing the resistant genotype, do not always express the phenotype, a phenomenon called heteroresistance. This makes detection of methicillin-resistance among coagulase-negative staphylococci more difficult. As a result, before embarking upon therapy with a beta-lactam antibiotic, the microbiology laboratory must be asked to thoroughly exclude possible methicillin-resistance. If this cannot be done, methicillin-resistance should be assumed. (See "Overview of antibacterial susceptibility testing", section on Heteroresistance).

PATHOLOGY — The intracardiac pathology of infection involving mechanical valves, particularly when PVE presents during the early months after surgery or when it is caused by invasive organisms, shapes the requirement for therapy. Perivalvular 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 [14,15] . In one series, invasion of perivalvular tissue was noted in more than 80 percent of cases [16] .

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 [17-19] . Large vegetations may prevent closure of the prosthesis producing incompetence or encroach upon the valve orifice causing functional stenosis.

Bioprosthetic valve endocarditis also causes invasive infection. As an example, 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 cases presenting later (59 versus 25 percent) [20] . In another series, invasive disease was more common in patients with early compared with later bioprosthetic PVE (79 versus 31 percent) [21] .

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 and inflammatory infiltrates thus potentially causing a noninfectious process to be misdiagnosed as PVE. A retrospective pathologic analysis of inflamed bioprosthetic valve tissue from 88 patients who underwent surgical removal of a bioprosthetic valve (21 for suspected endocarditis and 67 for noninfective dysfunction) was conducted to better define the histologic criteria for PVE [22] . Histologically, PVE was characterized by microorganisms, vegetations, and neutrophil-rich, inflammatory infiltrates. In contrast, inflammatory infiltrates in valve tissue samples from the noninfective control group consisted mainly of macrophages and lymphocytes. A neutrophil surface area with a cutoff value of >1.5 percent of total valve tissue surface was highly specific for PVE (94 percent).

CLINICAL MANIFESTATIONS — Patients with PVE present with symptoms and signs similar to those encountered in NVE. (See "Infective endocarditis: Historical and Duke criteria").

However, when PVE develops prior to discharge from the hospital, findings related to surgery or to perioperative complications may predominate over the subtle features of endocarditis.

Signs and symptoms of invasive infection — The high frequency of invasive infection results in higher rates of new or changing murmurs, heart failure, and new electrocardiographic conduction disturbances in patients with PVE than in those with NVE. 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 [16,23-25] . (See "Complications and outcome of infective endocarditis").

Valvular dysfunction, fever which persists for nine days or more despite appropriate antibiotic therapy, new electrocardiographic conduction disturbances, and echocardiographic evidence of abscess formation are clinical manifestations of invasive infection [26,27] . These findings were noted in 64 percent of 116 patients with PVE in one study and occurred significantly more frequently when infection involved an aortic valve prosthesis and when it arose within the first year after valve replacement [27] .

DIAGNOSIS — Laboratory findings in patients with PVE are similar to those noted in patients with NVE of comparable duration. (See "Diagnostic approach to infective endocarditis").

Blood cultures — In the absence of prior antibiotic therapy, blood cultures will be positive in 90 percent or more of 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 blood cultures are drawn over a period of hours to days in a patient with a prosthetic valve and all or most are positive, PVE is highly probable. The duration of documented bacteremia is particularly important when the isolate is a common 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 [28] .

If antibiotics have not been administered prior to obtaining blood cultures, it is unusual to have persistently negative blood cultures in patients with clinically overt PVE. Nevertheless, culture-negative cases can occur when infection is caused by fastidious organisms such as Legionella species, Bartonella species, Coxiella burnetii, Mycoplasma hominis, fungi other than Candida species, and the HACEK organisms. (See "Endocarditis caused by Bartonella" and see "Q fever endocarditis").

Detection of these and other unusual fastidious pathogens causing PVE relies upon the same evaluation used to assess culture-negative NVE [29] . (See "Diagnostic approach to infective endocarditis").

Echocardiography — The availability of transesophageal echocardiography (TEE) has greatly improved the diagnosis and management of PVE and is considered the procedure of choice if only one type of echocardiogram is to be obtained. Although transthoracic (TTE) and transesophageal echocardiograms are complementary in the diagnosis of PVE, TEE has a sensitivity of 82 to 96 percent compared with 17 to 36 percent for TTE [30-32] . This increased sensitivity is achieved without regard to valve position and without loss of specificity [30-34] (show table 3). (See "Diagnostic approach to infective endocarditis").

One prospective study compared TTE and TEE in 114 episodes of infective endocarditis suspected on clinical grounds (34 PVE, 80 NVE) [35] . The results of the two tests were concordant in 55 percent of cases, but TEE led to a reclassification of the case in 34 percent of patients with prosthetic valves compared to 11 percent of patients with native valves. Twenty-two patients were reclassified as definite infective endocarditis based upon the TEE results; 10 of these patients had prosthetic valves. In two patients with PVE the diagnosis would have been rejected if the evaluation had been limited to TTE data.

Echocardiography, including both TTE and TEE, is essential for both the diagnosis and management of PVE [36] . A high resolution biplane or multiplane transducer that allows continuous-wave and pulse-wave Doppler and color-flow imaging should be employed. TTE provides superior images of the ventricular surfaces of prostheses in the mitral, tricuspid, and aortic positions, whereas the atrial surfaces of the mitral and tricuspid valves and the aortic surface and outflow track of prostheses in the aortic position are better viewed by TEE. TEE is unquestionably superior to TTE for visualizing a mitral valve prosthesis.

TEE is also superior to TTE in the detection of abscesses, fistulae, and paraprosthetic leaks [37] . In one study that correlated the results of both types of echocardiography with anatomic findings at surgery in 88 valves from patients with endocarditis, the sensitivity of TEE for the detection of valve perforation was significantly higher than that of TTE [38] . TEE detected perforation in 21 of 22 cases compared to only 10 of 22 by TTE.

In patients with failing prosthetic valves, echocardiography may define the intracardiac pathology, which can provide evidence of a need for surgical therapy. However, TEE may miss some cases of prosthetic valve dehiscence. In a study that included 26 patients with prosthetic valve endocarditis, 14 (54 percent) had surgically identified valve dehiscence [43] . Dehiscence was missed on TEE in 4 cases which all involved an aortic prosthetic valve.

The negative predictive value of a full echocardiographic evaluation of a patient with suspected PVE is 86 to 94 percent. Nevertheless, if PVE is strongly suspected in the face of a negative evaluation, repeat echocardiography one week later can provide evidence of PVE in some cases [32,39] . Echocardiography has replaced other modalities of cardiac imaging in the diagnosis of PVE, including cardiac catheterization (which is now reserved for defining coronary artery anatomy).

Summary — Identifying patients with PVE requires a high index of suspicion and an appreciation of the subtle symptoms and signs of endocarditis. Obtaining three or four blood cultures over time, without confounding their utility by antibiotic administration, and an echocardiogram are paramount in the evaluation.

When sporadic blood cultures yield coagulase-negative staphylococci or when only several cultures have been obtained and are positive for these organisms, molecular techniques, such as DNA examination by pulsed-field gel electrophoresis, should be used to compare the isolates. Establishing clonality of isolates from separate cultures argues against contamination and favors that the isolates are true pathogens [28] . However, there is a potential for polyclonal infections to arise, especially from direct contamination in the operating room.

The Duke criteria for the diagnosis of endocarditis provide a systematic approach for diagnosing both NVE and PVE [40] . Among pathologically confirmed cases of PVE, 76 percent have been categorized as definite PVE and 24 percent as possible PVE when judged by the Duke criteria [41,42] . The expanded use of TEE in the assessment of patients with possible PVE has enhanced the sensitivity of these criteria for the diagnosis of PVE [35] . The standard for assessing possible PVE should include appropriately obtained blood cultures and echocardiographic imaging, with the results interpreted according to the Duke criteria. Using this approach, few cases of PVE should be missed. (See "Infective endocarditis: Historical and Duke criteria" and see "Diagnostic approach to infective endocarditis").