In-hospital mortality for patients with infective endocarditis (managed medically and/or surgically) remains high, ranging from 10% to 20% [ ]. At 6 months, nearly 30% of patients will have died, and at roughly 7 years, mortality will be over 50% [ , ]. Various factors increase mortality, including older age, the occurrence of neurologic or pulmonary complications, embolic events, and renal injury at the time of admission [ , , ]. Hasbun et al. created a four-tiered scoring system for 6-month mortality in patients with left-sided native-valve endocarditis, which includes the patient’s mental status, Charlson comorbidity scale score, degree of congestive heart failure, causative microbiology, and treatment (medical or surgical) [ ].
The proportion of endocarditis patients requiring admission to an intensive care unit (ICU) prior to medical or surgical cure has not been specifically studied, but intensive care is typically necessary when end-organ injury or heart failure is present [ ]. An estimated 60% of patients with endocarditis experience at least one major complication, and surgical intervention is required in up to 50% [ , ]. Not surprisingly, a 50% in-hospital mortality is experienced by those patients who require ICU care [ , ]. Generally speaking, the complications from endocarditis result from direct damage to the heart, from vegetation embolization, or from systemic hypoperfusion due to septic or cardiogenic shock. Systemic embolization occurs in approximately 30% of cases [ , ]. Antimicrobial therapy is, of course, paramount to treatment and is covered in Chapter 7 .
This chapter will review the care of the critically ill endocarditis patient prior to definitive surgical repair. We will first discuss the management of heart-related complications of endocarditis including valvular dysfunction and conduction abnormalities. We will then discuss the treatment of neurologic, renal, respiratory, and visceral complications ( Table 8.1 ). The subsequent chapter will address the postoperative care and management of complications in this critically ill population.
|Heart||Valvular insufficiency||40%||Echocardiogram||Surgical intervention|
|Conduction abnormalities||25%||EKG, echocardiogram||Surgical intervention|
|Acute coronary syndrome||Rare||CT, coronary catheterization||Surgical intervention|
|Brain||Embolic stroke||20%–40%||CT brain, MRI||Consider anticoagulation|
|Cerebral hemorrhage||20%–40%||CT brain, MRI||Supportive care; ensure no concomitant mycotic aneurysm requiring intervention|
|Mycotic aneurysm||Uncommon||MRA||Neurosurgery, interventional neuroradiology for clipping versus coiling|
|Meningitis||1%–5%||Lumbar puncture||IV antimicrobial therapy|
|Cerebral abscess||1%–5%||MRA||Needle aspiration, drainage|
|Encephalopathy||25%||Clinical diagnosis, MRI||Supportive care, correct nutritional or metabolic deficiencies (B12, folate, thiamine)|
|Kidney||Acute kidney injury||30%–35%||Urinalysis, urine electrolytes, metabolic panel||IV antibiotics, maintaining kidney perfusion, fluid management|
|Septic emboli||Up to 30%||Urinalysis, CT abdomen/pelvis||IV antibiotics, maintaining kidney perfusion, fluid management|
|Glomerulonephritis||25%||Urinalysis (red cell casts), metabolic panel||IV antibiotics, maintaining kidney perfusion, fluid management|
|Acute interstitial nephritis||10%||Urinalysis (white cell casts), metabolic panel||IV antibiotics, maintaining kidney perfusion, fluid management|
|Lung||Septic emboli||5%–10%||Chest XR, CT chest||IV antimicrobial therapy |
Surgical resection for treatment failure
|Spleen||Septic emboli||30%–35%||CT abdomen, labs||Splenectomy|
|Bone||Osteomyelitis||5%||Radiography, CT, MRI, nuclear bone scan||Prolonged IV antibiotics|
Preoperative management of cardiac complications
Cardiac complications affect over half of all patients with infective endocarditis. The most common cardiac abnormalities associated with infective endocarditis are valvular insufficiency with subsequent heart failure, conduction abnormalities, and acute coronary syndrome (ACS).
Acute valvular insufficiency in the setting of infective endocarditis is generally the result of perivalvular abscesses, which occur in up to 40% of all patients. Perivalvular abscesses most commonly complicate prosthetic-valve endocarditis, followed by endocarditis involving the native aortic and then the mitral valve. Valvular insufficiency occurs as perivalvular abscesses disrupt normal coaptation during diastole, or when the inflammatory process leads to primary valve leaflet disruption such as perforations or tears. Multiple valvular involvement, which is present in roughly 15% of all endocarditis cases, is thought to be due to secondary infection of the additional valve from direct extension or from satellite infections that arise from regurgitant jets of infected blood. Regardless of the mechanism, acute valvular insufficiency is a medical emergency as failure to quickly treat the valvular disruption may result in acute heart failure, cardiopulmonary collapse, and death.
Acute valvular insufficiency of left-sided heart structures commonly presents with sudden onset shortness of breath and fatigue. The physical examination is often notable for tachycardia with a high-grade diastolic decrescendo or holodiastolic murmur, evidence of pulmonary congestion, including increased work of breathing and bilateral rales, and other exam findings suggestive of decreased cardiac output, e.g., hypotension, cool extremities, and pallor. Urgent workup should include cardiac enzymes and an echocardiogram, which will usually be diagnostic and may show valvular insufficiency in the presence of flail leaflets or calcified echodensities on the involved valves. In the setting of acute aortic valvular insufficiency resulting in acute pulmonary edema, the patient should undergo emergent surgery, as medical treatment of the pulmonary edema in this setting is of limited benefit. In patients who do not present with acute pulmonary edema or in extremis, initial care should include medical management, broad spectrum antibiotics, and optimization for surgery, with a focus on managing the heart failure with diuresis and decreasing afterload with the use of vasodilators. Definitive treatment remains surgical and will be discussed in subsequent chapters.
Conduction disorders are found in roughly 25% of patients with endocarditis and are independently associated with increased mortality and worse prognosis [ , ]. Typically atrioventricular or bundle branch blocks are usually the result of perivalvular abscesses that extend into the conduction system [ ]. Conduction disorders are most commonly seen with aortic root involvement and occur due to the anatomic proximity of the aortic valve to the interventricular septum and bundle of His. The resulting abnormalities vary but all fall under the category of heart block, i.e., a bundle branch, or first, second, or complete heart block.
Conduction abnormalities are commonly detected by EKG. Suspicion for conduction system involvement as a result of endocarditis should be followed by an echocardiogram [ ]. Echocardiographic findings of an annular or valvular mass with perivalvular extension in the setting of new onset atrioventricular block should prompt immediate surgical intervention, likely entailing both valve replacement and, at times, repair of the involved annulus or septum [ ].
Acute coronary syndrome
ACS is a rare complication of infective endocarditis and generally occurs in the setting of existing coronary artery disease [ ]. Coronary arteries may be directly involved due to mechanical distortion from aortic root abscesses or become directly infected via local extension through the ostia. Another potential mechanism involves infarction as a result of septic emboli or thrombosed mycotic aneurysms [ , ]. Lastly, and most uncommon, ACS may result from coronary compression secondary to periannular or root involvement [ , ]. Treatment in each of these cases is definitive surgical treatment of the infected valve or tissue, and coronary revascularization if due to infarction or direct ostial involvement.
Additional cardiac complications
Additional significant and potentially devastating cardiac complications of infective endocarditis result from extension of the perivalvular abscess leading to fibrinous (more common) and suppurative pericarditis, aortic root and myocardial abscesses, hemopericardium, cardiac tamponade, aortic aneurysms, dissections, and intracardiac fistulas to an adjacent cardiac chamber [ , ]. Fistulas are more common with prosthetic-valve endocarditis resulting from valvular extension eroding into adjacent myocardium [ ]. Structural involvement of endocarditis, such as in the examples listed above, is an absolute indication for surgery.
Preoperative management of noncardiac complications
Neurologic complications are the most common extracardiac sequelae of endocarditis, occurring in an estimated 20%–40% of cases [ , , ]. The most frequent neurologic complication is a transient ischemic attack or a stroke from embolization of (typically left-sided) vegetations. Embolic strokes tend to occur early in the course of illness [ , ]. Additional neurologic complications include cerebral hemorrhage, mycotic aneurysm, meningitis, cerebral abscess, and encephalopathy. Any of these neurologic complications increases the patient’s mortality risk [ , , ]. For example, Mourvillier et al. found a threefold increased risk of death for ICU endocarditis patients with neurologic complications [ ]. For any complication, neurologic or otherwise, early initiation of antibiotics for source control is critical. In fact, neurologic sequelae of endocarditis dramatically decrease once appropriate antibiotics are initiated [ , ].
Neurologic deficits including altered mental status should be investigated with cerebral imaging. While computed tomography (CT) scan is often faster and more easily accessible, magnetic resonance imaging (MRI) is more sensitive [ ]. In fact, MRI often reveals pathology even in the absence of neurologic symptoms. For instance, in a French study where all suspected or proven endocarditis patients underwent cerebral MRI with angiography regardless of neurologic symptoms, 82% had a cerebral lesion, and in 28% the findings led to a definitive diagnosis (i.e., upgraded from suspected to definitive endocarditis) or changed the therapeutic plan [ ].
Anticoagulation in patients requiring it prior to the diagnosis of endocarditis should be continued, but there is no evidence to support its use prophylactically to prevent emboli. Furthermore, in a patient who has suffered an ischemic stroke, the benefit of anticoagulation must be weighed against the risk of conversion to hemorrhagic stroke [ ]. The limited literature on rates of hemorrhagic transformation in infective endocarditis comes from prosthetic-valve endocarditis patients, with an estimated transformation rate from ischemic to hemorrhagic stroke of 10%–14% if on anticoagulation [ , ]. Decision-making can also be assisted by scoring systems created to predict the risk of hemorrhagic transformation in ischemic stroke patients. While these systems are not specific to endocarditis patients, they estimate risk of hemorrhagic transformation based on clinical parameters and thus can be used as guidance to help decide whether or not patients with an indication (e.g., venous thromboembolism or atrial fibrillation) should be anticoagulated. For instance, the hemorrhagic risk stratification score (HeRS) predicts increased risk of hemorrhagic transformation with increased age, larger infarct volume, and lower estimated glomerular filtration rate [ ]. The HeRS plus score additionally utilizes serum glucose, white blood cell count, and the use of coumadin prior to admission to predict the risk of hemorrhagic conversion [ ].
Antiplatelet agents have similarly been ineffective in preventing stroke and embolic complications, as determined by a randomized trial that found no benefit to full-dose aspirin in these patients [ ]. In that study, as well, there was a trend toward a higher risk of bleeding in treated patients. Major bleeding (defined as intracranial bleeding, bleeding resulting in a drop in hemoglobin of ≥2 g/dL or requiring transfusion, or bleeding into a confined space such as the pericardium) or minor bleeding (all other overt bleeding causing a drop in hemoglobin of <2 g/dL) occurred in close to 30% of patients on aspirin versus half that rate in control patients [ ].
In patients who have suffered a stroke and in whom surgical management is indicated, the timing of surgical intervention is frequently critical, as intraoperative or early postoperative conversion to a hemorrhagic stroke can be catastrophic. Unfortunately, this issue has not been settled and continues to be debated. Frequently, the decision is individualized with less than complete confidence in the chosen course of action. See Chapter 14 for more on the timing and indications for surgery [ ]. The benefit of removing the infected valve as a potential source of additional emboli must be weighed against the risk of further brain injury, and currently, most advocate delaying surgery for at least 2 weeks after an ischemic embolic infarct [ , ]. Use of the HeRS score might be useful in these cases to help determine timing, but this has not been definitively studied.
A mycotic aneurysm is an abnormal arterial dilation resulting from embolization of septic emboli which causes infection of the arterial wall, weakness, and subsequent dilation. They are one of the less common neurologic complications. When they occur, they often do not cause neurologic deficits until they grow large enough to cause a mass effect or rupture, but MRI or magnetic resonance angiography is sensitive enough to detect even those that are asymptomatic [ ]. There is no standardized management given the overall rarity of this entity, with individualized and varied practices reported in the literature [ ]. Generally speaking, management can be either with neurosurgery or with interventional neuroradiology, depending on the location of the aneurysm and the capabilities of a given institution. Surgical clipping is preferred when there is significant hematoma with mass effect, whereas endovascular coiling is utilized for patients with surgically inaccessible mycotic aneurysms, multiple aneurysms, or those considered high surgical risk [ ]. One particular benefit to endovascular management is the ability to systemically anticoagulate soon after intervention, whereas a craniotomy requires a longer waiting period off anticoagulation, i.e., mycotic aneurysms treated endovascularly can have operative intervention for their endocarditis sooner than those who have undergone neurosurgical management of their mycotic aneurysm.
Subarachnoid hemorrhage complicating endocarditis should lead one to image for mycotic aneurysms as they are associated with hemorrhage in approximately 1.5% of cases and would change management to include microsurgical clipping or endovascular coiling of the aneurysm as described above [ , ].
Meningitis and cerebral abscesses are also rare but recognized complications, occurring in approximately 6% and 1% of endocarditis patients, respectively [ ]. As with all cerebral complications, they primarily occur with left-sided valvular lesions. Overall, when addressing causes of meningitis, endocarditis is a rare cause. A nationwide study from the Netherlands identified endocarditis in just 2% of community-acquired bacterial meningitis patients over a 6-year period [ ]. That study found that among patients with endocarditis-associated meningitis, roughly 60% had cerebrospinal fluid (CSF) findings predictive of bacterial meningitis compared to 90% of patients with meningitis from other causes [ ]. Nonetheless, suspicion for meningitis, prompted by symptoms including fever, nuchal rigidity, change in mental status, headache, and/or nausea, should be investigated with a spinal tap, looking for glucose level <34 mg/dL, ratio of CSF glucose to blood glucose <0.23, protein level >220 mg/dL, or leukocyte count >2000 per 1 mm [ , ]. Intravenous antimicrobial therapy is the mainstay of treatment and should be guided by CSF cultures. Intravenous glucocorticoids, usually dexamethasone, should also be administered at the time of antibiotics. This Infectious Disease Society recommendation comes from several trials (but importantly, not all) that have shown dexamethasone to decrease mortality and neurologic sequelae such as hearing loss in bacterial meningitis patients (not specific to endocarditis), particularly in meningitis caused by Streptococcus pneumoniae [ ].
If a cerebral abscess is diagnosed, workup should also include imaging to assess for mycotic aneurysm (as described above), as they often coexist. Brain abscesses require drainage. Needle aspiration is generally preferred over surgical excision given its lower risk of neurologic sequelae. Obviously, aspirated fluid should be sent for culture to guide antimicrobial therapy [ ].
Any neurologic involvement may present with seizures. Treatment is the same as with any seizure in a critically ill patient, with intravenous loading of an antiseizure medication such as levetiracetam. Finally, if an endocarditis patient has encephalopathy, treatment should focus on supportive care and correction of nutritional or metabolic deficiencies such as B12, folate, and thiamine. Usually, these critically ill patients have multiple etiologies for encephalopathy, including, but not limited to, sepsis, multiorgan dysfunction, and delirium. Anything modifiable, such as electrolyte abnormalities, nutrition, and behavioral modifications to minimize delirium, should be addressed, with the expectation that the encephalopathy will improve.
Renal dysfunction and acute kidney injury (AKI) can result from localized or global infarcts (either due to septic emboli or to low-flow states in septic and/or cardiogenic shock), glomerulonephritis, and acute interstitial nephritis from antibiotics [ ]. AKI complicates nearly one-third of endocarditis cases [ ]. On histologic analysis of the kidneys of 62 endocarditis patients (biopsy or autopsy), Majumdar et al. found a 31% incidence of localized infarcts (over half due to septic emboli), a 26% incidence of acute glomerulonephritis, and a 10% incidence of acute interstitial nephritis. Septic emboli can also cause renal abscesses, and any renal complication increases mortality [ , ].
A history of flank pain may point toward a diagnosis of renal ischemia from emboli. Furthermore, urinalysis can help differentiate between the etiologies of AKI in endocarditis. Hematuria, frank or microscopic, is suggestive of renal ischemia, with proteinuria also present in an estimated 12% of patients with renal infarction [ ]. White cell casts in the urine sediment are suggestive of acute interstitial nephritis, and red cell casts are suggestive of glomerulonephritis. Initiation of antibiotics is paramount for source control, but it is critical to dose appropriately to prevent nephrotoxicity, particularly in the face of preexisting renal dysfunction or an acute renal injury. Maintenance of cardiac output and an adequate blood pressure both protect the kidney from global ischemic injury and acute tubular necrosis. Careful fluid management similarly plays a role. As in other diseases involving kidney injury, when indications exist, most commonly fluid overload, electrolyte abnormalities, or symptomatic uremia, there should be no hesitation to initiate hemodialysis.
Pulmonary embolic complications are less common than cerebral or renal involvement as they typically occur only with right-sided lesions which comprise only 5%–10% of endocarditis [ ]. Septic emboli typically result in bacterial pneumonia or lung abscesses which can be assessed by chest radiography, or with more sensitivity, chest CT. Characteristic findings on chest CT include subpleural peripheral nodules and wedge-shaped peripheral lesions [ ]. Again, both for prevention and therapy, immediate initiation of antimicrobial therapy is crucial. Effusions may require drainage with either thoracentesis or chest tube placement, whereas loculated effusions may require surgical drainage, just as pulmonary abscesses may require drainage or resection [ ]. Little data exist on the mortality associated specifically with pulmonary emboli due to endocarditis, given its overall rarity, but a mortality of 12%–30% is reported for all-comers with septic pulmonary emboli, e.g., those caused by endocarditis, Lemierre’s syndrome, infected peripheral deep venous thrombosis, etc. [ ].
Recurrent septic pulmonary emboli on appropriate antibiotic therapy are an indication for operative intervention for endocarditis as per both the European Society of Cardiology and the American Heart Association/American College of Cardiology guidelines [ , ], as detailed in Chapter 14 .
Embolism to the spleen can cause splenic infarction which can result in a splenic abscess either from the embolus itself being infected or from secondary infection of infarcted tissue. In an autopsy study of 68 endocarditis patients, approximately one-third had splenic infarcts [ ]. They are best diagnosed with abdominal CT. Splenic infarcts do not require specific treatment, but there is a risk of hemorrhage or splenic rupture if the patient is systemically anticoagulated for cardiopulmonary bypass. Hemorrhage and rupture require a splenectomy for treatment [ , ].
Splenic abscesses are typically symptomatic, causing fever, leukocytosis, and left upper quadrant abdominal pain [ ]. Antibiotics do not penetrate splenic abscesses well, and the mortality for medically managed splenic abscesses complicating infective endocarditis is as high as 80% [ , ]. Splenic abscesses should therefore be managed with splenectomy, despite reported mortality rates of 6%–14%, although this risk is ill-defined in the setting of endocarditis [ , ]. Robinson et al. reported their experience of 27 splenic abscesses in a total of 564 endocarditis patients (5%), approximately 50% of whom (13) died. None treated medically survived, whereas surgical splenectomy was associated with over 80% (14/17) survival [ ]. There is some evidence to suggest that, when possible, the splenectomy should occur prior to valve replacement surgery to avoid potentially contaminating the new prosthesis [ ]. Percutaneous aspiration and/or drain placement with interventional radiology has also been described [ ].
Osteomyelitis complicates as many as 6% of cases of infective endocarditis [ ]. Persistent musculoskeletal complaints (e.g., back pain, bone pain, stiffness, joint pain) should therefore be investigated. Workup can begin with plain radiography, although radiographs are normal in 50% of early-stage osteomyelitis cases [ ]. Abnormalities on radiography become more apparent after approximately 2 weeks of bony involvement. For increased sensitivity, CT, MRI, or nuclear bone scan should be performed. MRI is usually the modality of choice, with nuclear bone scan used in patients who have contraindications to either CT or MRI. Treatment is with a prolonged course of antibiotics—at least 6 weeks of intravenous antibiotics and at least 3 months of oral therapy. If blood cultures are positive for a likely pathogen, bone biopsy is not necessarily needed, but in the setting of diagnostic uncertainty, bone biopsy, either percutaneous or open, can provide the diagnosis and guide antimicrobial treatment [ ].
Complications and sequelae of infective endocarditis are common and associated with significantly increased morbidity and mortality. Cardiac or extracardiac involvement is generally the result of valvular disruption or embolic phenomena, which can affect nearly every organ system. A thorough knowledge of the range of complications and organs affected can inform workup and appropriate diagnoses and therapies, which in turn may impact the necessity and timing of surgery. Appropriate preoperative management of the complications of infective endocarditis should be a target of any quality initiative directed at improving overall care and outcomes in this critically ill population.