Complications and outcome of infective endocarditis

http://www.uptodateonline.com/online/content/topic.do?topicKey=endocard/5997&selectedTitle=36~150&source=search_result

Complications and outcome of infective endocarditis

Author
Denis Spelman, MBBS, FRACP, FRCPA, MPH
Daniel J Sexton, MD
Section Editor
Stephen B Calderwood, MD
Gabriel S Aldea, MD
Scott E Kasner, 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: May 19, 2008 (More)


INTRODUCTION — Infective endocarditis (IE) is associated with a myriad of complications, one or more of which occur in the majority of patients. This was illustrated in a review of 223 episodes of IE in which 57 percent of patients had one complication, 26 percent two, 8 percent three or more, and 6 percent six or more complications [1] . Complications such as heart failure and stroke are relatively common and feared outcomes of IE, while other complications such as blindness and septic arthritis are, fortunately, rare in modern practice.

The frequency and type of complications due to IE has changed with advances in diagnosis and therapy. Renal failure and uncontrolled intracardiac or metastatic infection, for example, which were previously common complications of IE, are infrequent in the antibiotic era. The frequency of specific complications depends upon variables such as the infecting pathogen, duration of illness prior to therapy, and the type of treatment facility (eg, referral versus community hospital). It is often difficult to assess the true incidence of complications despite extensive literature on the subject because different reviews and case series were generally based upon retrospective chart reviews and used different diagnostic criteria to define cases of IE. (See "Infective endocarditis: Historical and Duke criteria").

The major complications of IE will be reviewed here but will only attempt to describe the frequency of these complications as relatively common, rare, or very rare because accurate numbers are not available. Complications can occur before, during and, rarely, even after the end of therapy (eg, ruptured mycotic aneurysm).

DEFINITIONS — Complications of IE can be broadly categorized as:

• Cardiac
• Septic
• Embolic
• Neurologic
• Musculoskeletal
• Renal
• Associated with medical treatment


While these categories are broad and organize the complications in an understandable fashion, they do not take into account significant overlapping features. As an example, patients with neurologic involvement can simultaneously have embolic and septic processes.

One can also consider complications in terms of their pathogenesis, which leads to different groupings:

• Embolic (eg, cerebral infarct)
• Local spread of infection (eg, heart valve destruction)
• Metastatic infection (eg, vertebral osteomyelitis)
• Immune-mediated damage (eg, glomerulonephritis)


CARDIAC COMPLICATIONS — Cardiac complications are the most common complications seen in IE, occurring in one-third to one-half of patients in most recent case series [2] .

Heart failure — Heart failure (HF) remains the most common cause of death due to IE in the modern era and is the most frequent reason for cardiac surgery in patients with IE. (See "Surgery for native valve endocarditis").

The usual cause of HF in patients with IE is valvular insufficiency resulting from infection-induced valvular damage. Rarely, embolism of fragments of valvular vegetations or vegetation-induced stenosis of the coronary ostia can cause acute myocardial infarction and subsequent HF [2] .

Perivalvular abscesses — Among patients with IE, the reported incidence of perivalvular abscess at surgery or autopsy has ranged from about 30 to 40 percent [3-5] . The aortic valve and its adjacent annulus are more susceptible to abscess formation and the complications of perivalvular extension of infection than are the mitral valve and ring [3-5] . This was illustrated in an autopsy study of patients with native valve endocarditis: annular extension of infection was far more common in patients with aortic valve compared to mitral valve endocarditis (41 versus 6 percent) [3] .

Injection drug use may be another risk factor for perivalvular abscess [4] . In contrast, although large vegetation size had been implicated as a risk factor in some series, subsequent studies have shown no correlation between the presence or size of the vegetation and the development of periannular extension [4,6] .

Paravalvular abscesses can extend into adjacent cardiac conduction tissues, possibly leading to various forms of heart block. Involvement of the conducting system is most common with infection of the aortic valve, especially when there is involvement of the valve ring between the right and non-coronary cusp (this anatomic site overlies the intraventricular septum that contains the proximal ventricular conduction system).

Paravalvular abscess should be suspected when fever persists despite appropriate antimicrobial therapy and/or when conduction abnormalities appear on the ECG [7] . Transesophageal echocardiography (TEE) has a much greater likelihood of detecting a myocardial abscess than transthoracic echocardiography (TTE). One study, for example, evaluated 118 patients with IE, 37 percent of whom had an abscess documented at surgery or autopsy [5] . The sensitivity, specificity, and positive and negative predictive values of TEE imaging were 87, 95, 91, and 92 percent, respectively; the sensitivity of TTE was much lower (28 versus 87 percent) although the specificity was 99 percent. However, some perivalvular abscesses may be missed by TEE. (See "Role of echocardiography in infective endocarditis", section on Perivalvular abscess or fistula).

Patients with perivalvular abscesses appear to have higher rates of systemic embolization and fatal outcomes. In one study comparing outcomes of patients with and without perivalvular abscesses, the rate of embolization was approximately twice as high (64 versus 30 percent) [4] . These patients also have a higher mortality rate (23 versus 14 percent in those without abscesses in a report of 118 patients, 44 of whom had a perivalvular abscess) [5] . Mortality may be particularly high in patients with at least moderate valvular regurgitation [8] .

Other extravalvular complications — Other rare extravalvular cardiac complications of IE include:


• Pericarditis, which may be suppurative or nonsuppurative, can rarely cause pain or even cardiac tamponade [9,10] . (See "Purulent pericarditis" and see "Evaluation and management of acute pericarditis").
• Fistulous intracardiac connections (eg, aorta-atrial or aorta-ventricular) due to extension of infection from the valve to adjacent myocardium may rarely result in large aneurysms, a pseudoaneurysm if the aortic wall is involved [11] , or even myocardial perforation.


The incidence of fistulous intracardiac complications was 1.6 percent in a retrospective, multicenter study of 4681 episodes of IE [12] . Surgery was performed in 66 of the 76 patients with a mortality of 41 percent. Multivariate analysis identified heart failure (odds ratio [OR] 4.3), prosthetic IE (OR 4.6), and urgent or emergent surgical treatment (OR 4.3) as being significantly associated with an increased risk of death.

• Aortic valve dissection [13] .
• Descending thoracic aorta intraluminal infectious masses [14] .


EMBOLIZATION — Embolization remains a distressingly common complication of IE and can occur even after appropriate therapy is well underway. This section will provide a general discussion of embolization in patients with IE. Issues related to embolization in patients with IE who undergo surgery are discussed separately (See "Surgery for native valve endocarditis" section on Embolization, and see "Surgery for prosthetic valve endocarditis" section on Embolization).

Systemic emboli most commonly complicate left-sided IE, but rarely can occur in tricuspid valve endocarditis via a patent foramen ovale [15] . However, emboli to the lung with subsequent abscess formation occur frequently in patients with tricuspid endocarditis. Small, clinically inapparent embolization probably occurs in most, if not all, patients with IE, but clinically recognized embolism has been reported in 13 to 44 percent of patients in published reports [16,17] .

Emboli consisting of vegetation fragments can occlude or damage virtually any blood vessel, large or small, in the systemic or pulmonary arterial circulation. As a result, emboli can produce:


• Stroke
• Blindness
• Painful ischemic or frankly gangrenous extremities
• Unusual pain syndromes (eg, due to splenic or renal infarction)
• Hypoxia (due to pulmonary emboli in right-sided endocarditis)
• Paralysis (due to embolic infarction of either the brain or spinal cord)


Emboli can occasionally cause symptoms or signs that mimic other common conditions such as kidney stones, Bell's palsy, dizziness, or pleurisy.

Endocarditis should be considered as a possible etiology in virtually all patients who present with signs or symptoms of systemic arterial embolization. In one study, cerebral infarction was the presenting sign of IE in 4 to 14 percent of all cases of IE [15] . The vast majority of patients with an acute stroke do not have endocarditis although the occurrence of a stroke in a younger patient or evidence of simultaneous or sequential cerebral and systemic arterial embolization, increase the probability of IE. As an example, in one report of 60 patients with IE and cerebral complications, 28 (47 percent) also had clinically identifiable systemic emboli [18] compared to only 2 percent of all stroke patients in a different large series [19] .

Symptomatic embolization appears to be more common with IE due to fungal pathogens. Whenever emboli to large systemic arterial vessels occur in a patient with IE, the possibility of a fungal etiology should be entertained. In a literature review of 270 patients with fungal endocarditis, peripheral arterial embolization occurred in 45 percent. The most common sites were the cerebral circulation (17 percent) and femoral artery (16 percent) [20] .

Histopathologic or microbiologic examination of occluding embolic material in such large vessels may lead to a diagnosis of an underlying fungal IE. Haemophilus species and other slow growing fastidious gram-negative organisms also seem to predispose to embolization with some frequency. (See "Candida endocarditis").

Effect of antibiotic therapy on embolic risk — The risk of embolization tends to decline after the institution of effective antimicrobial therapy, and serious embolic events rarely occur several weeks after such therapy is instituted [16,21,22] . The relationship between the initiation of antibiotic therapy and the risk of embolism is illustrated by the following observations:

In a series of 629 patients with left sided endocarditis, 131 patients (21 percent) had one or more embolic events [21] . The event occurred before the initiation of antimicrobial therapy in 42 percent, on the day therapy commenced in 14 percent, and within 15 days of initiating therapy in 82 percent.
In a study that used TTE to detect vegetations, the rate of embolization fell from 13 per 1000 patient days during the first week of therapy to less than 1.2 per 1000 patient days after two weeks of therapy [16] .
These findings suggest that surgery may not be necessary for prevention of embolic stroke in the early weeks following initiation of appropriate antibiotic therapy, if there are no other indications for surgery (such as a large mobile vegetation or congestive heart failure due to valvular leak) [22] .

Predictors of embolization — The size of a vegetation as determined by echocardiography has been assessed as a risk factor for embolization in IE. Although some data are conflicting, vegetation size is generally a risk factor for embolization. (See "Role of echocardiography in infective endocarditis", section on Echocardiographic estimation of outcome).

However, the presence or absence of vegetations, the location of the vegetations, the characteristics of vegetations by TEE, the specific etiologic microorganism, or antiphospholipid antibodies may have predictive value [23-29] :

• A multicenter prospective European study of 384 patients with definite IE by Duke criteria found that emboli were more frequently observed in cases due to Streptococcus bovis and S. aureus by multivariate analysis [29] . Multivariate analysis identified vegetation size >10 mm and severe vegetation mobility as additional risk factors for embolic events that occurred in 28 patients after the initiation of antibiotic therapy.
• Embolization occurs more frequently with left-sided than right-sided vegetations [24] . In a review of 281 patients with clinically suspected IE, the incidence of embolic events was greater with mitral than aortic valve vegetations (25 versus 10 percent) [25] . The risk was highest with vegetations on the anterior mitral leaflet (37 percent), suggesting that the mechanical effects of broad and abrupt leaflet excursion may contribute to the risk of embolization [24] .
• A prospective study of patients with IE due to S. aureus suggested that the risk of embolization was significantly greater in patients who had visible vegetations by both TTE and TEE compared to patients who had vegetations visualized only by TEE [26] .
• The absence of valvular abnormalities on TTE may be associated with a decreased incidence of complications [27] .
• The presence of antiphospholipid antibodies were correlated with an increased risk of embolization (62 versus 23 percent) in a series of 91 patients with IE, perhaps due to increased endothelial cell activation, generation of thrombin, and defective fibrinolysis [28] .

Effect of prior antiplatelet therapy — The possible protective effect of prior antiplatelet therapy (one or more of aspirin, dipyridamole, clopidogrel, or ticlopidine) on embolism in IE was evaluated in a retrospective cohort of 600 patients with IE, 147 of whom (25 percent) had a symptomatic embolic event [30] . The patients who had received continuous daily antiplatelet therapy for at least six months prior to hospitalization for IE had a significantly lower rate of a symptomatic embolic event (12 versus 28 percent without such therapy, adjusted odds ratio 0.36, 95% CI 0.19-0.68). The presumed mechanism is that platelet aggregation plays a role in vegetation formation.

In contrast to prior therapy, the initiation of aspirin after the diagnosis of IE is of no benefit and may be harmful. This was illustrated in randomized trial in which 115 patients with IE were assigned to aspirin (325 mg/day) or placebo for four weeks [31] . Aspirin did not reduce the incidence of embolic events, was associated with a trend toward an increased incidence of bleeding (odds ratio 1.92, 95% CI 0.76-4.86), and had no effect on vegetation resolution or valve function.

NEUROLOGIC COMPLICATIONS — Neurologic complications rank second to cardiac in importance, occurring in approximately 25 to 35 percent of patients [15,32-36] , although a lower rate of 10 percent was noted in one series [37] . In a review of 260 nondrug addicts with IE due to Staphylococcus aureus, 91 patients (35 percent) developed neurologic manifestations including 61 (23 percent) who presented with these symptoms [34] . (See "Complications of Staphylococcus aureus bacteremia").

Manifestations — The mechanism for and types of neurologic complications are diverse and include:

• Embolic stroke
• Acute encephalopathy
• Meningoencephalitis
• Purulent or aseptic meningitis
• Cerebral hemorrhage (due to stroke or a ruptured mycotic aneurysm)
• Brain abscess or cerebritis
• Seizures (secondary to abscess or embolic infarction)


Neurologic complications may be the presenting symptom in patients with IE. In a series of 68 patients with stroke and endocarditis, for example, two-thirds presented to the hospital with stroke, before the diagnosis of endocarditis was made [37] . Thus, the possibility of IE should be considered in all patients who present with strokes, meningitis, or a brain abscess. Unexplained fever accompanying a stroke in a patient with valvular disease is an important clue in some patients. (See "Diagnostic approach to infective endocarditis").

Outcomes — Reported patient outcomes after a neurologic complication are variable. The following findings have been noted in different series:

• Among survivors of cardiac surgery for IE, 70 percent of patients with a preoperative stroke experienced a full neurologic recovery [36] . Outcomes were worse in patients with stroke complicated by meningitis, abscess, or intracerebral hemorrhage.
• Patient mortality in more contemporary series has varied from approximately 20 to 50 percent at one year [36,37] to as high as 74 percent (time of follow-up not given) in patients with Staphylococcus aureus endocarditis [34] .

One of the dilemmas that often arises is whether neurologic complications of IE are a contraindication to valve replacement. This important issue is discussed elsewhere. (See "Surgery for native valve endocarditis", section on Effect of recent cerebral embolization).

MYCOTIC ANEURYSMS — Mycotic aneurysms can occur in the cerebral or systemic circulation of patients with IE, usually at points of vessel bifurcation. (See "Mycotic aneurysms").

RENAL DISEASE — Renal infarction (due to emboli), drug-induced acute interstitial nephritis, glomerulonephritis (due to deposition of immunoglobulins and complement in the glomerular membrane) and, rarely, renal abscess can occur in patients with IE. (See "Renal disease in infective endocarditis").

Acute renal failure, defined as a serum creatinine of 2 mg/dL (177 µmol/L) or greater, has been reported in up to one-third of patients [38] . By contrast, chronic renal failure due to immune-complex mediated glomerulonephritis, which was a common contributing cause of death in patients who presented with classic IE in the preantibiotic era, is now rare. Immune complex-mediated renal disease is also uncommon in the antibiotic era, especially in patients whose infection is detected and treated early.

METASTATIC ABSCESSES — Rarely, metastatic abscesses develop in the kidneys, spleen, brain or soft tissues (eg, the psoas muscle) in the setting of IE. There is a strong association between IE and splenic abscess, even though otherwise splenic abscesses are less frequently observed than other types intraabdominal abscesses.

Patients with splenic abscesses usually do not have marked abdominal pain or splenomegaly; persistent fever during or after treatment for IE and occasionally recurrent bacteremia after cure of the valvular infection may be the only clue to the presence of this complication [39] . Splenic abscesses are often diagnosed only at autopsy and generally require splenectomy for cure. In one study of 27 patients with splenic abscesses, mortality was 100 percent in the patients who did not have a splenectomy compared to 18 percent in the patients who had the procedure [40] .

Discrete microabscesses or larger solitary brain abscesses can rarely occur in patients with IE. Abscess formation occurs as a sequela of septic embolization. Some patients with IE and brain abscesses also have purulent meningitis. In fact, the presence of meningitis due to S. aureus should suggest the possibility of concomitant S. aureus endocarditis. In one case series of 33 patients with S. aureus meningitis, seven (21 percent) also had endocarditis [41] .

Appropriate treatment, including drainage of such abscesses, is needed not only to control the local infection but also to prevent ongoing bacteremia, which is of particular concern among patients who may require surgical treatment of endocarditis with implantation of a prosthetic valve.

MUSCULOSKELETAL COMPLICATIONS — Vertebral osteomyelitis is a well known but relatively rare complication of IE. Although the majority of patients with IE and back pain do not have vertebral osteomyelitis, protracted, severe back pain in any patient with IE should alert the clinician to this possibility. Plain films are insensitive for diagnosing vertebral osteomyelitis, especially if taken early in the course of illness [42] . (See "Vertebral osteomyelitis"). Osteomyelitis more frequently complicates S. aureus endocarditis than IE due to other microorganisms [42] .

Acute septic arthritis, involving one or more joints, may be the first clue to the presence of IE in a small percentage of patients. IE should be strongly considered in selected cases of septic arthritis:

• When infections spontaneously arise in joints of the axial skeleton (eg, sacroiliac, pubic, or manubriosternal joints).
• When organisms with a known propensity to cause IE (eg, S. aureus, viridans streptococci or non-group A beta-hemolytic streptococci) grow from a joint aspirate, particularly in patients without a history of recent surgery, joint infection, or trauma.
• When multiple joints are infected.


COMPLICATIONS OF MEDICAL OR SURGICAL THERAPY — Patients with IE can develop a number of the complications associated with prolonged parenteral antimicrobial therapy or surgery:

• Aminoglycoside-induced ototoxicity or nephrotoxicity (See "Pathogenesis and prevention of aminoglycoside nephrotoxicity and ototoxicity")
• Secondary bacteremia due to central vascular lines (See "Pathogenesis of and risk factors for central venous catheter-related infections")
• Mediastinitis or early postoperative prosthetic valve endocarditis (See "Postoperative mediastinitis after cardiac surgery")
• Intravenous catheter-associated phlebitis
• Drug fever (See "Drug fever")
• Allergic or idiosyncratic reactions to various antimicrobial agents
• Bleeding due to disturbances in coagulation caused by anticoagulants (in prosthetic valve endocarditis)


MORTALITY — Multiple studies have evaluated death rates in patients with both native and prosthetic valve endocarditis [43-47] :

• The in-hospital mortality rate is between 18 and 23 percent
• The six month mortality rate is between 22 and 27 percent


The outcomes in patients with neurologic complications are described above. (See "Outcomes" above).

Predictors of death — Several studies have attempted to identify predictors of death in patients with IE. Each patient may have one or more of the following:

• Infection with S. aureus [44,46,48] , while mortality is lower with streptococcal infection (8 versus 33 percent with S. aureus in one series) [46]
• Heart failure [45,47]
• Diabetes mellitus [44]
• Embolic events [44,48]
• Perivalvular abscess [5,8,47]
• Vegetation size [29,48]
• Female gender [29]
• Contraindication to surgery [46]
• Low serum albumin [43]
• Persistent bacteremia [47]
• Abnormal mental status [46]
• Poor surgical candidacy [46]


All but two of the preceding studies [44,46] were retrospective. Because the clinical and echocardiographic features of patients with endocarditis change during the course of illness, some of the above findings should be interpreted with caution. In one of the reports using a prospective study design, neither heart failure as defined by the Framingham criteria nor cardiac surgery was independently associated with in-hospital mortality [44] . However, among patients with moderate to severe heart failure, cardiac surgery in other studies has been associated with a lower rate of long-term mortality compared to medical therapy alone. The data supporting this conclusion are presented separately. (See "Surgery for native valve endocarditis", section on Efficacy).

In view of the wide disparity in the methods used in the preceding studies, one should be cautious about making prognostic predictions in individual patients with IE.