The Role of Surgery

Despite medical progress in the diagnosis and antimicrobial therapy of IE, more than half of patients with IE suffer a serious complication, and the mortality rate is unacceptably high: ~20% during the initial hospitalization and ~40% at one year [485]. The major causes of death are structural complications and hemody-namic instability. As such, cardiac surgery, principally valve replacement, has become an important adjunct to medical therapy. Cardiac surgery is currently used in 25-50% of cases, and several studies suggest that combined medical and surgical therapy can reduce both early and late mortality in patients with a complicated course.

Several indications for surgery in patients with IE have been proposed by Olaisson and Peterson [1], as well as the AHA, with varying strengths of evidence. The former are provided in Table 9.8 Consensus indications for surgery during IE include the following: acute anatomical cardiac destruction; congestive heart failure (CHF); hemodynamically significant valvular dysfunction; perivalvular extension of infection (abscess or fistula); persistent (uncontrolled) infection; and lack of effective antimicrobial therapy available (or alternatively, difficult-to-treat pathogens). Surgery is also indicated for the majority of cases of prosthetic valve endocarditis (discussed in chapter 11) and for the management of mycotic aneurysms (see above). There is a lack of consensus on the indications of surgery in the management of embolic complications.

CHF, regardless of the pathogenesis, is the strongest predictor of mortality in patients with IE. As such, it is the strongest indication for surgery [1]. Among patients with NVE who develop moderate-to-severe (New York Heart Association III or IV) CHF and are treated with medical therapy alone, the mortality rate is 56-86%; among patients treated with combination medical and surgical therapy, the mortality rate is 11-35% [1]. Therefore, CHF is a bad prognostic factor. Furthermore, patients with IE who undergo cardiac surgery have higher perioperative mortality rates if they do so in CHF

Table 9.7. Classification of Myocardial Abscesses (Adapted from

Chakrabarti [452])

1. Endocarditis-related

I.Contiguous from

a. Valvular IE (perivalvular abscess)

b. Mural IE

2. Hematogenous seeding of myocardium

2. Septicemia-related

Hematogenous seeding of myocardium,

usually in association with abscesses


3. Miscellaneous

I.Trauma and penetrating injuries

2. Iatrogenic (e.g., catheterization,


3.Anatomic abnormalities (e.g., aneurysm

infection, infection of infarcted

myocardium, infection of myxoma)

(15-35%), when compared to patients without CHF (5-10%) [455]. As such, early cardiac surgery, ideally at the onset of CHF and before the onset of physiologic compromise, should be performed. Of note, the beneficial effect of surgery persists even in the presence of co-morbidities;

as such, the development of other complications (e.g., acute renal failure) is not a contraindication to proceed to surgery [486].

Anatomical destruction, such as acute valvular destruction with insufficiency, rupture of the chordae tendinae or papillary muscles, will usually manifest as CHF, necessitating cardiac surgical intervention. Other sequelae of acute destruction include rupture into the pericardium and septal perforation; these may manifest with acute hemodynamic compromise. In these situations, emergent surgery is indicated.

Physiologically significant valve dysfunction can manifest as insufficiency, producing a syndrome of CHF, or with valvular obstruction. The latter may occur, for example, as a result of large vegetations or thrombi superimposed on a stenosed native or on a prosthetic valve. Such obstruction can compromise cardiac output; hence the need for urgent surgery.

Perivalvular extension of infection can develop as paravalvular myocardial abscess or

Table 9.8. Indications for Surgical Intervention in the Management of IE (Olaison and Petersson [1]) Reprinted from Infectious Disease Clinics of North America V.16, Olaison L, Pettersson G, Current best practices and guidelines: Indications for surgical intervention in infective endocarditis,453-475,Copyright (2002),with permission from Elsevier.

Indications for surgery in patients with infective endocarditis


Evidence Based

Emergency indication for cardiac surgery (same day)

1. Acute AR with early closure of mitral value


2. Repture of a sinus Valsalva aneurysm into the right heart chamber


3. Rupture into the pericardium


Urgent indication for cardiac surgery (within 1-2 d)

4. Valvular obstruction


5. Unstable prosthesis


6. Acute AR or MR with heart failure, NYHA III-IV


7. Septal perforation


8. Evidence of annular or aortic abscess, sinus or aortic true or false aneurysm,fistula formation, or new onset

conduction disturbances


9. Major embolism + mobile vegetation > 10mm + appropriate antibiotic therapy < 7-10


10. Mobile vegetation > 15 mm + appropriate antibiotic therapy < 7-10 d


11. No effective antimicrobial therapy available


Elective indication for cardiac surgery (earlier is usually better)

12. Staphylococcal prosthetic valve endocarditis


13. Early prosthetic v alve endocarditis (<2 mo after surgery


14. Evidence of progressive paravalvular prosthetic leak


15. Evidence of valve dysfunction and persistent infection after 7-10 d of appropriate antibiotic therapy, as

indicated by presence of fever or bacteremia, provided there are no noncardiac causes for infection


16. Fungal endocarditis caused by a mold


17. Fungal endocarditis caused by a yeast


18. Infection with difficult-to-treat organisms


19. Vegetation growing larger during antibiotic therapy > 7 d


Abbreviations: A, Strong evidence or general agreement that cardiac surgery is useful and effective; AR, aortic regurgitation; B, Inconclusive or conflicting evidence or a divergence of opinion about the usefulness/efficacy of cardiac surgery, but weight of evidence/opinion of the majority is in favor;C, Inconclusive or conflicting evidence or a divergence of opinion; lack of clear consensus on the basis of evidence/opinion of the majority. MR, mitral regurgitation; NYHA, New York Heart Association classification.

as an intracardiac fistula. The former has been previously discussed. Intra-cardiac fistulous tracts usually develop from either aortic root abscesses or pseudoaneurysms that rupture into adjacent chambers. These fistulae may be single or multiple and generally extend from the aorta to the right atrium, right ventricle, or the left atrium [456]. As well, aortic insufficiency from IE may produce a septic regurgitant jet that strikes subaortic structures, creating secondary sites of infection. Abscesses form at such sites in the left ventricular outflow tract, especially in the mitral-aortic intervalvular fibrosa or junctional tissue between the anterior mitral leaflet and the aortic valve. This leads to pseudoa-neurysm formation and rupture into the left atrium, creating a left ventricular-left atrial shunt [487]. The diagnostic modality for detection of these fistulous tracts is TEE [487].

Persistent bacteremia has been defined as bacteremia with an organism identical to the initial isolate, despite > 7 days of antimicrobial therapy to which the isolate was susceptible [1,488,489]. However, positive blood cultures after 1-4 days of antibiotic therapy have been predictive of complicated bacteremias [490492]. In the absence of an extracardiac source (e.g., metastatic septic foci), persistent bacteremia indicates a failure of antimicrobial therapy and the most likely source would be intracardiac. As such, diagnostic imaging (e.g., TEE) should be pursued. Persistent fever is not synonymous with persistent bacteremia. In acute uncomplicated infective endocarditis, defervescence occurs within 1 week of effective antimicrobial therapy in 75% of patients and by two weeks in 90% of patients [493]. The presence of fever during therapy should be categorized as "persistent" if there has been no defervescence after one to seven days, or as "recurrent" if there was an initial period of decreased temperature [492,494]. Persistent fever after the first week of hospitalization suggests a septic embolic focus (e.g., visceral abscess) or an intracardiac complication, either of which may or may not be the result of inadequate antibiotic therapy [492]. Recurrence of fever suggests a focal septic complication, non-infectious embolic phenomenon (e.g., visceral infarct), a drug-hypersensitivity reaction (drug fever), or, least commonly, the emergence of a resistant strain [492]. In a single-center, prospective study of 193 patients with IE, 57% of patients had "persistent" or "recurrent" fever.

Of the patients with "persistent" fever, 56% were due to cardiac complications. "Recurrent" fever was most often caused by hypersensitivity reactions to P-lactams [494].

The presence of difficult-to-treat pathogens is an indication for surgical intervention (1). Frequently cited examples include Pseudomonas aeruginosa, fungi (e.g., Candida spp., Aspergillus spp.), Coxiella burnetti, and Brucella spp., organisms for which antimicrobial therapy exists, but when used alone, unlikely to lead to eradication. It is becoming clear, however, that even for pathogens with "adequate" antimicrobial agents available, surgical intervention combined with medical therapy may be the superior treatment of choice. Examples of such situations include NVE with S. aureus, certain coagulase-negative staphylococci, and P-hemolytic streptococci (see previous sections). This decision is particularly true in the prsence of any of the above complications.

The role of surgery in preventing CNS complications remains ill-defined. Neurological complications occur in 20-40% of patients with IE [52,495], and can manifest as brain infarction, mycotic aneurysms with/without intracerebral hemorrhage, bacterial meningitis, or toxic encephalopathy. The purpose of surgery would be to prevent septic embolic phenomena. Emboli, however, can occur before diagnosis, during therapy, or after treatment is completed. Identification of predictive factors to estimate an individual patient's risk of embolization has been difficult. Previous attempts to use echocardiography to identify high-risk vegetation characteristics, and thus to identify a subgroup of patients who may benefit from prophylactic surgery, have produced conflicting results. More recent studies have demonstrated that the large majority of embolic complications occur before the diagnosis and institution of antimicrobial therapy [213,214]. Even with antibiotic treatment, the risk of embolization remains elevated for the first two weeks [496]: in one study [162], 65% of embolisms occurred during this period. The risk decreases to 15% after one week of treatment, and then to 1% after four weeks of treatment [455]. Thus, the preventative effect of surgery would be maximal in the first few days of treatment. However, this potential benefit is tempered by the fact that early cardiac surgery would expose a number of patients, who would not have otherwise developed this complication, to the risks inherent with surgery. As well, these patients would be exposed to the risks associated with prosthetic valves (i.e., lifelong anticoagulation for metallic prosthetic valves, re-do surgery for bioprosthetic devices, risk of prosthetic valve endocarditis). As such, the traditional indication for valvular surgery for IE to avoid embolization has been the development of > 2 major embolic events, although this recommendation is arbitrary [52]. Objective risk factors that may aid in decision-making include the size of the vegetation at baseline, the progression of the vegetation size on therapy, and the infecting microorganism.

Vegetation size intuitively should correlate with risk of embolization. Larger, pedunculated vegetations are potentially more friable at the surface or the neck; when such pieces are disrupted, it results in emboli. Although early data correlating vegetation size to risk of emboliza-tion were inconsistent, several subsequent large studies [162,497-499] and a meta-analysis [500] have shown that vegetation size (specifically >10 mm) is a strong predictor of thrombo-embolic events. There is some concern, though, that this "threshold" size not be dogmatic in determining the need for surgery. Vegetation size alone does not precisely identifiy all high-risk patients: not all patients with large vegetations invariably develop embolic complications, and conversely, some patients with relatively small vegetations do. Therefore, other factors clearly must be contributing to the likelihood of embolization. In addition to vegetation size, valvular location has been reported to be important in some studies [162]. As well, the infecting microorganism may play a role, but the data is not adequately powered [162,496]. In addition to vegetation size, vegetation echogenicity theoretically may contribute to predicting a patient's risk for embolizaton. Low-density vegetations are fresh, and thus friable, and would have a greater capacity to embolize than a high-density vegetation, which is more typical of a chronic and healed vegetative mass [162]. Several studies [162,501], however, demonstrated that there was no relationship between vegetation echogenicity and the risk of embo-lization.

Change in vegetation size is a useful sign. One study suggests that a decrease in vegetation size on antimicrobial therapy is indictive of a rapid healing process [502]. In practical terms, however, most vegetations remain constant in size, despite appropriate antimicrobial therapy; this occurred in ~84% of vegetations in one study [162]. Failure of the vegetation to regress, however, was not associated with a worse prognosis. Growth of vegetation on antimicrobial therapy is ominous. Several studies [162,502] have demonstrated that this feature is associated with poor control of the infection and a higher incidence of embolization.

In conclusion, future studies are required to better delineate the risk factors that most accurately predict embolization and whether prophylactic cardiac surgery in such patients is beneficial.

For the patient with IE who has already developed neurological deficit(s), cardiac surgery may still be indicated if the risk of recurrent embolism is high or if there are concomitant complications. Management thus is determined by the nature of the neurologic lesion, as well as the nature of these other complications. Although the most common CNS complication is embolic disease without hemorrhage [503], a CT scan of the head should be the first step to determine the presence of intracranial hemorrhage [52,495,503].

In the absence of any hemorrhage, only small studies exist to guide management. Maruyama and colleagues report the development of severe neurologic deterioration in 29% (4/14 patients) who underwent valve replacement within five days of an acute, non-hemorrhagic, cardiogenic embolism [504]. Matsushita et al. also reported fatal neurologic deterioration in two patients who underwent emergency cardiac surgery within five days of their stroke [505]. They also noted better outcomes among patients with ischemic events if they were medically treated for 11 days prior to surgery and for 23 days prior to surgery if they had hemor-rhagic strokes. Other groups have demonstrated similar results [495,503]. Thus, it has been recommended that, when possible, cardiac operation be delayed two to four weeks for patients who have non-hemorrhagic, cardiogenic emboli [52,495,503,506].

If hemorrhage is identified on CT, the most likely cause is ruptured mycotic aneurysm. As such, angiogram (e.g., MRA, CTA, or conventional) should be performed. Neurosurgical consultation should be obtained to assist in management. Cardiogenic embolism with hemorrhage is associated with an increased risk for perioperative stroke in cardiac surgery [507]. Therefore, surgical management of the aneurysm (e.g., clipping) may be necessary. In patients who undergo aneurysm clipping, subsequent valve replacement should be delayed for two to three weeks, if the patient is stable [503]. Cardiac operations should be performed only when there is stabilization of the neurologic status clinically, and CT imaging demonstrates resolution of cerebral edema with no ongoing bleeding. If surgical intervention for the aneurysm is not deemed necessary, and the patient is stable, an interval of four weeks between the neurologic event and cardiac surgery is recommended [495,503].

For patients with intracerebral hemorrhage and progressive cardiac failure, the prognosis is extremely poor. In this situation, the benefit from cardiac surgery may outweigh the risk of cerebral deterioration associated with the surgery.

Splenic involvement in IE can be divided into two complications: splenic infarct and splenic abscess. These two conditions are not mutually exclusive, but represent a pathophysiological spectrum. Splenic infarct in IE occurs as a result of arterial compromise, due to embolization of portions of sterile fibrinous vegetations embo-lizing into the terminal arteries of the spleen. Splenic abscess is a suppurative collection which can develop in patients with IE either as a result of septic emboli or infection of prior infarct. The incidence of splenic complications of IE is unclear, largely because septic infarcts typically have no symptoms or localized findings, and thus may go unrecognized, whereas the incidence rates for splenic abscesses have been based on retrospective studies, and is thus influenced by recall bias. With these limitations, the incidence rate of splenic complications in IE has been estimated at 35-40% [52,508]. Clinically recognized splenic abscess occurred in 2-5% of IE cases [508,509]. Among cases of splenic abscess from all causes, 10-20% are due to endocarditis [510]. One study has demonstrated that the risk of splenic emboliza-tion in IE is equivalent for aortic and for mitral vegetations [511].

Splenic infarcts are most often asymptomatic [511], although in patients at high risk for venous thromboembolism, such as IE, the most common presenting symptom is left upper quadrant abdominal pain [512]. The diagnosis can be easily obtained by abdominal ultrasonography (U/S) or CT. CT demonstrates superior sensitivity when compared to U/S (~96% vs. 75-90%, respectively) [510]. On CT, splenic infarcts typically appear as multiple, peripheral-based, wedge-shaped hypodense lesions without significant contrast enhancement [511,513]. They may vary in size, but they rarely involve the entire organ. CT also has the capacity to identify lesions as small as several millimeters [510]. The clinical significance of splenic infarcts is that these lesions are at risk for intraabdominal hemorrhage during valvular surgery for the IE, as a result of anticoagulation during cardiopulmonary bypass [511]. Furthermore, splenic infarcts may predispose to splenic rupture. Other complications include pseudocyst formation, as well as superinfection with subsequent development of splenic abscess [514]. In the absence of any complications, an isolated splenic infarction can be managed safely with medical treatment [511,514].

Splenic abscesses, on the other hand, are usually symptomatic, with evidence of sepsis being most prominent [511]. The classic triad consists of fever, leukocytosis, and left-upper-quadrant abdominal pain [510,515]. Fever is by far the most common symptom, occurring in >90% of cases [510,515]. Thus, patients with endocarditis, abdominal complaints, signs of sepsis (e.g., recurrent or persistent fever), or recurrent or persistent bacteremia should be evaluated for any potential foci for relapse, particularly the spleen. CT is very useful for identification of a splenic abscess, which typically appears as a solitary, round-to-irregular shape, centrally located, hypodense lesion that is contrast enhancing [511]. Air within the cavity is pathognomonic of abscess [511]. There is, however, considerable overlap between the CT patterns of splenic infarcts and abscesses. In addition to the morbidity to the patient, the major clinical sign-ficance for a splenic abscess is that it may serve as a source of subsequent bacteremia and seeding of a prosthetic valve inserted for management of IE. The other major complications of splenic abscesses include rupture into the peritoneal cavity, which is the most common, as well as rupture into contiguous spaces, producing visceral abscesses, peritonitis, or empyema [515].

The management of a splenic abscess requires a combined medical and surgical approach. Splenic abscesses respond poorly to antibiotic therapy alone. Although antibiotics are effective in clearing the bacteremia of IE, they do not penetrate well into splenic abscesses; consequently, organisms in the abscess are not eradicated and can still be cultured. Previous studies have demonstrated 100% mortality rates for patients undergoing medical therapy alone. Robinson and colleagues [509] identified 27 patients who developed splenic abscesses among 564 patients with IE between 1970 to 1990. Of these, there were 13 deaths: 10/13 (77%) of the patients who did not undergo splenec-tomy died, compared to 3/17 (18%) of the patients who underwent splenectomy. A literature review by Johnson etal. [516] demonstrated that the survival rate for 17 patients with splenic abscess who did not undergo splenectomy was 0%, compared to 95% who did. In situations in which antimicrobial therapy alone appears successful initially, recurrence of abscess formation is common. Based on this evidence, the recommended definitive management of splenic abscesses in patients with IE has been splenec-tomy [52,509,515,517-519], of which the goal is to eradicate the extra-cardiac focus of infection as a prerequisite to successful management of IE. If possible, the AHA 2005 guidelines [52] recommend that splenectomy be perfomed prior to valve replacement surgery, to minimize the risk of contaminating the valve prosthesis as a result of bacteremia from manipulation of the abscess. This recommendation, although conceptually logical, is not based on evidence in the literature. However, in one series of ten patients with IE in whom splenectomy was performed for splenic abscesses, the splenectomies were staged and performed at a mean time interval of 11.2 days after valve replacement (range: 3-24 days) [511]. Although follow-up data is not completely provided, three of ten (30%) of the patients who underwent splenectomy died: one in the postoperative period from bleeding, and two at unspecified times from "cardiac causes." Another study suggests that splenectomy can be performed before or after valvular surgery, depending on the patient's clinical status [509]. Laparoscopic splenectomy for splenic abscess, although potentially more difficult technically, appears to be a safe and effective alternative to open surgery [517,518].

More recently, radiographically guided percutaneous aspiration or catheter drainage has become popular. The advantage is that it spares the spleen, and thus avoids the risks of the hyposplenic state (e.g., overwhelming post-splenectomy sepsis). Success rates with this procedure have ranged from 75% to 100%, although several catheterizations may be needed to achieve cure [510]. Furthermore, this procedure has been associated with high rates of failed attempts, which subsequently have required rescue splenectomies [515]. However, the need for a rescue splenectomy does not appear to be significantly associated with increased mortality rates [510,515]. It has been recommended that percutaneous aspiration or catheter drainage be contraindicated in a select subgroup of patients, namely those with multi-loculated abscesses, septations, tenaciously thick abscess contents, or abscess rupture/ bleeding [510,515].

In conclusion, the role for surgery in the management of IE or its complications is expanding. Although the risks for surgical intervention in patients with complicating features such as those discussed in this chapter are real, there is ample evidence that combined modality treatment is beneficial in specific instances.

All Natural Yeast Infection Treatment

All Natural Yeast Infection Treatment

Ever have a yeast infection? The raw, itchy and outright unbearable burning sensation that always comes with even the mildest infection can wreak such havoc on our daily lives.

Get My Free Ebook

Post a comment