Surgery for prosthetic valve endocarditis

Surgery for prosthetic valve endocarditis

Author
Adolf W Karchmer, MD
Section Editor
Catherine M Otto, MD
Scott E Kasner, MD
Gabriel S Aldea, MD
Deputy Editor
Elinor L Baron, MD, DTMH
Susan B Yeon, MD, JD, FACC



Last literature review version 16.2: May 2008 | This topic last updated: April 29, 2008 (More)


INTRODUCTION — Infection of a prosthetic heart valve can be difficult to diagnose and manage. Optimal treatment of prosthetic valve endocarditis (PVE) requires:

• Identification of the causative microorganism.
• Selection of a bactericidal antimicrobial regimen of proven efficacy.
• A clear understanding of the intracardiac pathology and attendant complications of PVE. (See "Presentation and diagnosis of prosthetic valve endocarditis").
• Surgical intervention in nearly all cases, especially that in which infection has extended beyond the valve to contiguous cardiac tissue.


The surgical management of prosthetic valve endocarditis will be reviewed here. The antimicrobial treatment of prosthetic valve endocarditis and the role of surgery in native valve endocarditis are discussed separately. (See "Antimicrobial therapy of prosthetic valve endocarditis" and see "Surgery for native valve endocarditis").

GENERAL PRINCIPLES — Cardiac surgery plays a major role in the effective therapy of many patients with PVE. Heart failure, persistent fever for 10 or more days despite appropriate antibiotic therapy, systemic embolization, and new onset electrocardiographic conduction abnormalities are associated with high mortality rates in patients with PVE and are clinical indications of invasive infection [1-4] . (See "Presentation and diagnosis of prosthetic valve endocarditis" and see "Surgery for native valve endocarditis").

Invasive infection is demonstrated far more predictably by transesophageal echocardiography than by a transthoracic evaluation. (See "Role of echocardiography in infective endocarditis").

Invasive infection is common in PVE, especially when infection arises within 12 months of surgery or involves an aortic prosthesis [2] . For patients with complicated PVE, the survival rates are higher among surgically than medically treated patients, and relapses, rehospitalization for valve surgery, and delayed mortality due to endocarditis are less common among those treated surgically [2,3,5-7] .

INDICATIONS — Indications for cardiac surgery in patients with PVE have been developed based upon the intracardiac pathology of PVE and the risk of recrudescent infection on the new prosthesis. The 2006 American College of Cardiology/American Heart Association (ACC/AHA) guidelines on the management of valvular heart disease included recommendations for surgery in patients with prosthetic and native valve endocarditis (show table 1A-1B) [8] .

Some of the indications for surgical treatment are absolute, while others are relative and require careful risk benefit analysis. Persistent bacteremia despite optimal antimicrobial treatment requires surgical intervention.

A multicenter prospective study of 104 patients with PVE examined the influence of medical versus surgical therapy on outcome in order to identify patients for whom surgery may be beneficial [9] . Multivariate analysis identified the following independent predictors of in-hospital mortality:


• Severe heart failure (odds ratio [OR] 5.5; 95% CI, 1.9-16.1)
• S. aureus infection (OR 6.1; 95% CI, 1.9-19.2)


Independent predictors of long-term mortality (32 month follow-up) included:

• Comorbidity (risk ratio [RR] 3; 95% CI, 1.4-6.6)
• Early PVE (<60 days post-surgery, RR 2.1; 95% CI, 1.1-4.3)
• Severe heart failure (RR 4.2; 95% CI, 2.2-8.0)
• Staphylococcus infection (RR 2.0; 95% CI, 1.0-4.0)
• New prosthetic dehiscence (RR 2.4; 95% CI, 1.3-4.7)


Mortality was not significantly different between surgical and nonsurgical patients overall (17 versus 25 percent), however, in-hospital mortality was reduced by a surgical approach in patients with staphylococcal (27 versus 73 percent) or complicated PVE (18 versus 48 percent).

An international multi-center prospective study of 355 patients with prosthetic valve endocarditis found the following variables independently associated with surgical therapy: intracardiac abscess, heart failure, younger age, coagulase negative staphylococci, and S. aureus. Unadjusted in-hospital mortality rates in this study were similar for surgical and medical treatment, 25 and 23 percent, respectively. When a propensity analysis was performed to compare surgically treated patients with a population of medically treated patients selected to have similar clinical features, there was a trend toward reduced in-hospital mortality with surgical treatment (OR 0.51; 95% CI 0.23-1.36). A similar trend was noted when all patients with high propensity score for cardiac surgery were examined [10] .

In contrast, antibiotic therapy alone is often successful in patients with PVE who have no evidence of heart failure, significant prosthetic valve dysfunction, or invasive infection, and who are infected by less virulent organisms. These patients are characterized by later onset of infection (more than 12 months after prosthesis implantation), and infection by viridans streptococci, HACEK, or enterococci (that can be treated with bactericidal therapy) [9,11,12] .

Valve dysfunction — The outcome of PVE in patients who experience moderate to severe heart failure due to prosthesis dysfunction is improved if patients are treated surgically. Few survive beyond six months if treated with antibiotics alone, whereas 44 to 64 percent survive with timely surgical intervention [1,3,11,13,14] .

An unstable hypermobile prosthesis due to dehiscence of anchoring sutures is a surgical emergency, requiring urgent intervention, because the valve is likely to become increasingly unstable with acute severe valve regurgitation.

As with NVE, surgical intervention to correct valve dysfunction and heart failure must be performed before it becomes severe and intractable. There is no evidence that delaying surgery in this setting improves outcome or reduces the frequency of recurrent endocarditis [15-17] . In fact, the operative mortality of these patients is proportional to the severity of hemodynamic disability at the time of surgery [15,18] .

Patients with PVE complicated by perivalvular invasion experience high mortality rates and are rarely curable with medical treatment alone. By contrast, complex reconstructive procedures have been associated with survival rates of 80 percent [6,19] . Echocardiographic findings are enhanced with transesophageal (versus transthoracic) but can still be limited by "shadowing" from prosthetic valve sewing rings. Echocardiographic findings of perivalvular invasion include [1,2] :

• Valve dehiscence
• Paravalvular abscess
• Aortic aneurysm or pseudoaneurysm
• Fistula formation


Relapse after optimal medical therapy — If PVE relapses after appropriate antimicrobial therapy, surgical intervention should usually be performed since relapses often reflect unrecognized perivalvular infection [1,2] .

Microorganisms usually requiring surgery — Retrospective studies suggest that S. aureus PVE is associated with significant mortality, up to 70 percent in patients treated with antibiotics alone [20-24] . While mortality outcomes in patients who had surgery in these studies have varied, intracardiac complications in patients with S. aureus PVE are consistently associated with an increased mortality and surgical intervention reduces mortality in this group. In one study, for example, mortality was reduced 20-fold by surgical intervention during antimicrobial therapy (odds ratio [OR] 0.5, 95% CI 0.005-0.42) [23] . Multiple studies have concluded that PVE caused by S. aureus is most effectively treated by antibiotics and prompt surgical intervention [3,9,20,21,25] .

Another study, while not finding an improvement in overall mortality with surgical treatment of S. aureus PVE, noted that patients who developed cardiac complications defined as heart failure and/or intracardiac abscess (a group that would be considered at high risk for endocarditis related death) and underwent early valve replacement had the lowest mortality rate (28.6 percent) [24] .

Early surgical intervention is considered by most experts to be a standard element of treatment for fungal PVE [26] . Among 15 patients with fungal PVE treated with antifungal agents and with surgery, 10 (67 percent) survived with an average follow-up of 4.5 years [27] . One review of 17 patients suggested that survival rates for Candida PVE were comparable with and without surgery (46 versus 50 percent) [28] . However, only patients with uncomplicated PVE survived in this review, and many remained on long term suppressive oral therapy. Patients who have fungal PVE have significant co-morbidities (such as end-stage renal disease on hemodialysis or immunosuppressed states from chemotherapy or HIV for example). Timing and duration of medical therapy and surgical intervention are frequently complex.

Some other pathogens, such as P. aeruginosa and probably multi-resistant enterococci for which there is no synergistic bactericidal regimen, are also less amenable to medical therapy. Surgery is generally advised for PVE caused by these microorganisms.

Embolization — The frequency of embolic complications is higher in patients with NVE who have vegetations exceeding 10 mm in diameter compared to those with smaller vegetations. Comparable data are not available for patients with PVE. However, the overall rate of embolic complications is similar for patients with PVE and NVE, and emboli decrease rapidly with effective therapy in both [29] . While prevention of emboli that cause highly morbid, irreversible end-organ damage (eg, central nervous system and myocardial infarction) is a laudable goal, it has not been established that surgical intervention achieves this aim in patients with PVE, large vegetations, and no other complications. The risk of emboli in this setting, rather than constituting an indication for surgery, should be weighed with other findings that might benefit from surgical intervention and then factored into the overall management plan. Recurrent emboli despite appropriate antibiotic therapy are an indication for surgical intervention.

OUTCOME — Complex reconstruction of the aortic or mitral valve apparatus and the supporting structures is often required to achieve an optimal outcome of PVE [19,30-32] . In the hands of experienced cardiac surgeons, operative mortality rates for patients with invasive PVE, treated with valve replacement and surgical reconstruction of paravalvular tissue, range from 10 to 30 percent; in contrast the projected mortality would approach 100 percent without surgery [6,19,30-33] . These data suggest that surgical intervention for PVE complicated by extensive invasion and tissue disruption should be performed in centers with extensive experience, when possible. Earlier surgical reintervention in patients with invasive organisms may also significantly limit morbidity by avoiding the need for complex reconstructions (such as root replacement for annular abscess or atrial-ventricular reconstructions for mitral valve prosthetic valve endocarditis).

The rate of recrudescent PVE after surgery is six to 15 percent; repeat surgery is required for recurrent PVE or for dysfunction of the newly implanted prosthesis in 18 to 26 percent [2,6,15,19,30,33] . While these figures are not insignificant, they are relatively small compared with the anticipated mortality with antibiotic therapy alone. Five-year survival rates ranging from 54 to 82 percent have been reported for patients undergoing surgery for PVE [15,19,30,32,34] .

ANTIBIOTIC TREATMENT FOLLOWING SURGERY — Following valve replacement for active bacterial endocarditis, the Task Force on Infective Endocarditis of the European Society of Cardiology (ESC) recommends another full course (six weeks) of antimicrobial treatment if the intraoperative valve culture is positive [35] . If the culture is negative, the ESC recommends that the full treatment course be completed (counting the duration of preoperative antibiotics).

ANTICOAGULANT THERAPY — Among patients with prosthetic valve IE, the potential benefit of preventing embolization with anticoagulation must be weighed against the increased risk of intracerebral hemorrhage. This issue is discussed in detail separately. (See "Anticoagulant and antiplatelet therapy in patients with infective endocarditis").

SUMMARY AND RECOMMENDATIONS


• Treatment of prosthetic valve endocarditis is more difficult than treatment of native valve endocarditis and often requires surgical replacement of the prostheses in addition to antibiotic therapy. (See "Introduction" above).
• Invasive infection is common in PVE, especially when infection arises within 12 months of surgery or involves an aortic prosthesis. (See "General principles" above).
• Some of the indications for surgical treatment are absolute, while others are relative and require careful risk benefit analysis (show table 1A). (See "General principles" above).
• The outcome of PVE in patients who experience moderate to severe heart failure due to prosthesis dysfunction or who have evidence of valve dehiscence, paravalvular abscess, or fistula formation is improved if patients are treated surgically. There is no evidence that delaying surgery in this setting improves outcome or reduces the frequency of recurrent endocarditis. (See "Valve dysfunction" above).
• Surgery is generally advised for PVE caused by S. aureus when accompanied by intracardiac complications, and also for fungi, gram-negative (nonHACEK) microorganisms (particularly P. aeruginosa), and multi-drug resistant enterococci. (See "Microorganisms usually requiring surgery" above).
• In the hands of experienced cardiac surgeons, operative mortality rates for patients with invasive PVE, treated with valve replacement and surgical reconstruction of paravalvular tissue, range from 10 to 30 percent; in contrast the projected mortality approaches 100 percent without surgery. (See "Outcome" above).
• The rate of recurrent PVE after surgery is six to 15 percent. (See "Outcome" above).