Infective endocarditis

Article written and reviewed by Dr Luca Moretti

Infective endocarditis is an infection of the endocardium and intracardiac structures, predominantly native valves, prostheses and implanted material. The characteristic lesion is the vegetation, composed of microorganisms, fibrin, platelets and inflammatory components; however, infection may also be present when no vegetation is recognizable, particularly on prostheses or after embolization. Injury comprises three interconnected processes: persistence of an intravascular focus, destruction of cardiac structures, and embolic or metastatic dissemination. The disease must therefore be approached simultaneously as an infection, a mechanical heart disease and a potential cause of multiorgan lesions.

The course ranges from rapidly destructive sepsis to subacute forms with intermittent fever, fatigue, anemia and weight loss. Incidence varies across populations and surveillance systems; in high-income countries, ageing, degenerative valvular disease, hemodialysis, prostheses and devices have altered the profile traditionally dominated by rheumatic valvular disease. Healthcare exposure may be relevant even when symptoms begin outside hospital. Mortality remains high and depends mainly on the microorganism, heart failure, anatomical extent, neurological injury and the ability to control the focus with antibiotics and surgery when required.

The classification of clinical forms integrates substrate, site, organism and mode of acquisition. A prosthesis introduces biofilm and imaging problems that differ from those of a native valve; right-sided disease mainly produces pulmonary emboli, whereas left-sided disease exposes the systemic circulation. Negative blood cultures are a diagnostic condition that must be explained, not an etiologic agent. These axes should be reported together because no single label determines the entire treatment by itself.

Etiology, pathogenesis and pathophysiology

In the classic model, endothelial injury exposes matrix and promotes an initially sterile platelet-fibrin deposit. Bacteremia can colonize it when inoculum, pathogen adhesion capacity and host susceptibility converge. Turbulence and high-velocity jets contribute to injury, but known heart disease is not essential. Staphylococcus aureus can adhere to and invade apparently intact surfaces, explaining acute infections of valves without previously recognized abnormalities. Everyday bacteremia of oral, cutaneous or healthcare origin must be distinguished from procedural exposure alone: attributing every episode to a recent dental procedure can lead to overlooking a persistent source.

Prosthetic surfaces and transvenous leads become coated with host proteins that facilitate microbial adhesion. Biofilm contains microorganisms embedded in an extracellular matrix, with populations showing different metabolic activity and reduced susceptibility to eradication. Tolerance within the focus is not the same as resistance of the isolate in the laboratory. Even in native-valve vegetations, high bacterial density, limited vascularization and slow growth in inner regions contribute to the need for prolonged antimicrobial exposure. A therapy that is active in blood can therefore control bacteremia without resolving an abscess or sterilizing colonized material.

Staphylococci play a central role, with important differences among species. S. aureus is frequently associated with an aggressive course, valvular destruction and metastatic foci; coagulase-negative staphylococci are particularly relevant in prostheses and devices. Their isolation cannot automatically be dismissed as contamination when they grow in multiple sets or are associated with intracardiac material and concordant findings. Staphylococcus lugdunensis, although coagulase-negative, can cause highly destructive native-valve infection. Species-level identification and methicillin susceptibility therefore modify both diagnostic probability and therapeutic choice.

Oral streptococci include species with different susceptibility to beta-lactams and can cause subacute disease on predisposed valves. The anginosus group may be associated with suppurative processes, whereas Granulicatella and Abiotrophia pose specific microbiological and therapeutic difficulties. Streptococcus gallolyticus has a clinically relevant association with colorectal disease and warrants evaluation of a possible portal of entry. The presence of a colonic lesion, however, does not demonstrate that every intestinal bacteremia has already caused endocarditis: cardiac involvement must be documented through the appropriate pathway.

Enterococcus faecalis is particularly important in older adults and in patients with healthcare exposure or gastrointestinal and genitourinary disease. Persistent or recurrent bacteremia without a clear source warrants a low threshold for cardiac investigation. Relative antibiotic tolerance and possible high-level aminoglycoside resistance affect the ability to achieve synergy; susceptibility to ampicillin alone does not exhaust the therapeutic problem. E. faecium often has a different resistance profile and should not be treated as equivalent. Investigation of the portal of entry proceeds in parallel with definition of endocardial involvement.

HACEK bacteria are fastidious organisms of the oropharyngeal flora that are now generally detectable by automated blood-culture systems, without having to attribute every negative culture to slow growth. Non-HACEK Gram-negative organisms are less common but may occur in healthcare settings, immunosuppression and injection drug use, with problems of resistance and source control. Coxiella burnetii, Bartonella and Tropheryma whipplei require dedicated methods. Test selection depends on exposures, phenotype and previous treatment: the term blood culture-negative endocarditis encompasses biologically different situations.

Fungi, particularly Candida and Aspergillus, more often affect patients with prostheses, catheters, surgery, immunosuppression or other predisposing exposures. Candida may be isolated from blood, whereas Aspergillus is often not recovered by routine blood cultures. Large vegetations, embolization and infection of implanted material often make it necessary to combine antifungal therapy and surgery. Fungal disease is not simply a variant managed by prolonging an antibiotic: it requires identification, susceptibility testing when available and a specific pathway.

Valvular destruction results from invasion and the inflammatory response: leaflet perforation, chordal rupture, flail and loss of annular support produce regurgitation. In acute defects, the atrium and ventricle have not had time to adapt; pulmonary edema and shock may develop while chamber dimensions are still normal. Infection can extend beyond the annulus, forming an abscess, pseudoaneurysm or fistula. With aortic involvement, proximity to the conduction system explains why new atrioventricular block is a sign of possible local invasion rather than merely an isolated electrical complication.

Fragmentation of vegetations generates septic emboli. Risk is greatest before and early after effective treatment begins and depends on site, mobility, size, organism and previous emboli. Left-sided lesions reach the brain, spleen, kidneys, mesentery, coronary arteries and limbs; right-sided lesions predominantly affect the lungs. The transported inoculum can produce abscesses or infect the arterial wall, with formation of infectious aneurysms. In parallel, persistent immune stimulation can generate immune complexes and glomerulonephritis. Multiorgan injury is therefore not explained solely by vascular occlusion or sepsis.

Clinical manifestations

The history should reconstruct the onset and course of fever, chills, sweating, anorexia, weight loss and reduced functional capacity. Previous endocarditis, valvular disease, prostheses, repairs, devices, congenital heart disease, hemodialysis and immunosuppression should be documented, together with hospitalizations, procedures, catheters and injection drug use. Medications already taken and their dates are essential: a partially active antibiotic can attenuate fever and render blood cultures negative without eradicating the lesion. In older adults, delirium, falls or general deterioration may predominate over the classic febrile syndrome.

The onset of dyspnea, orthopnea or rapid worsening of oxygenation suggests heart failure and requires investigation for new regurgitation or a mechanical complication. Hypotension, oliguria, altered mental status and elevated lactate may reflect sepsis, low cardiac output or both mechanisms. Treatment of shock should therefore also be guided by anatomy: response to fluids and vasopressors does not repair a perforated valve. Syncope, palpitations or a new conduction disturbance require monitoring and reassessment of cardiac extension.

Extracardiac manifestations include neurological deficits, abdominal or flank pain, limb ischemia, persistent spinal pain and arthralgia. Low back pain with bacteremia may indicate spondylodiscitis, whereas splenic pain and persistent fever suggest infarction or abscess. In right-sided forms, pleuritic pain, hemoptysis and pulmonary infiltrates may precede signs of right heart failure and mimic primary pneumonia. Chronology helps distinguish embolization, metastatic localization and an extracardiac infectious source that preceded the endocarditis.

Physical examination first assesses respiratory and circulatory stability, then murmurs, congestion, peripheral perfusion and neurological status. Petechiae, splinter hemorrhages, Janeway lesions, Osler nodes and Roth spots belong to the classic clinical signs but have low sensitivity. Janeway lesions are usually painless, whereas Osler nodes are painful; these findings do not replace microbiology and imaging. An absent or unchanged murmur does not exclude infection, particularly in early stages, prosthetic valves and low-output states.

Examination of the skin, access sites, oral cavity and device pockets seeks a source or associated infection. Erythema, discharge or pocket erosion have a different significance from an isolated lead mass, which may be thrombotic. Examination also includes the abdomen, spine and joints according to symptoms. This search should continue even after a vegetation is identified: an uncontrolled extracardiac focus can sustain bacteremia and compromise the result of cardiac surgery.

Investigations and diagnosis

The work-up starts from clinical probability and a parallel search for the organism, lesion and complications. Blood cultures should precede antimicrobials whenever possible: three sets are usually obtained from separate peripheral venipunctures, each with aerobic and anaerobic bottles and a volume appropriate to the system instructions. In adults, correct bottle filling is a major determinant of yield. There is no need to wait for a fever peak; in sepsis or shock, samples are collected rapidly without delaying treatment. Catheter-drawn cultures may complement the evaluation of a vascular source but should not indiscriminately replace peripheral samples.

The laboratory should be informed of the suspicion, prosthetic material, exposures and previous antibiotics. Species identification, susceptibility testing and minimum inhibitory concentration when relevant guide de-escalation. Rapid systems can provide earlier useful information, but a negative molecular panel does not exclude organisms or mechanisms not included in the assay. Follow-up cultures document clearance of bacteremia; interpretation requires distinguishing the collection date from the reporting date and verifying when the patient began receiving a truly active regimen. Persistent fever and persistently positive blood cultures are not synonymous.

Transthoracic echocardiography is the first examination for vegetations, regurgitation, ventricular function and pressures. Transesophageal echocardiography is required when the result is negative or inconclusive despite high suspicion, for evaluation of prostheses and to define complications even after a positive TTE. An exception may be clearly documented native right-sided disease with high-quality images and no additional diagnostic questions. If suspicion persists, the study is repeated, generally within 5-7 days according to ESC guidance, and sooner in the event of deterioration, embolism or heart block. A vegetation may initially be small or may already have embolized.

The report should describe site, maximum length in multiple planes, mobility, attachment and consequences of the mass, together with perforations, chordal ruptures, regurgitation and perivalvular lesions. Comparison with previous studies distinguishes new dehiscence from a known leak. Three-dimensional acquisitions assist anatomical reconstruction and planning but should be integrated with 2D imaging and Doppler. Differences of a few millimeters may depend on the measurement plane: apparent growth should be credible before it is used as an indicator of uncontrolled infection or increased embolic risk.

Cardiac CT complements assessment of abscesses, pseudoaneurysms, fistulas and anatomical relationships, particularly in prosthetic valves. Echocardiography and CT are not interchangeable: the former better depicts movement and small dynamic defects, while the latter can clarify perianular extension obscured by artifacts. Preoperative coronary assessment is adapted to atherosclerotic risk, urgency and the presence of aortic vegetations; coronary CT may avoid an invasive procedure in selected patients. An acute coronary event may also result from embolism or compression by a local complication and requires a specific strategy.

18F-FDG PET/CT is particularly useful in prosthetic-valve and intracardiac-material infection when echocardiography leaves uncertainty. Metabolic preparation, focal or heterogeneous distribution, surgical materials and time from implantation affect interpretation. Postoperative inflammation and adhesives can produce sterile uptake, whereas antibiotics and small lesions may reduce sensitivity. An isolated SUV value does not define infection. Radiolabeled leukocyte SPECT/CT is a complementary option in selected settings; whole-body imaging may also identify foci and portals of entry that change management.

Complete blood count, inflammatory markers, renal and liver function, electrolytes, coagulation and urinalysis define severity and treatment safety. Hematuria, proteinuria and active urinary sediment may indicate glomerulonephritis, whereas increased troponin or natriuretic peptides do not by themselves constitute a diagnostic criterion. Initial and serial ECGs assess conduction disturbances. Brain MRI and vascular imaging are used according to the neurological picture and therapeutic decisions; thoracoabdominal CT and spinal MRI address suspicion of emboli and deep foci. Isolated cerebral microbleeds are not equivalent to an acute intracranial hemorrhage.

Negative blood cultures require differentiation between previous antibiotic exposure and organisms requiring dedicated methods. Serology for Coxiella and Bartonella, additional exposure-guided tests and molecular techniques on blood or tissue complete the work-up. For Bartonella, an IgG titer by immunofluorescence of at least 1:800 is included among major Duke-ISCVID evidence; for Coxiella, specific microbiological evidence and phase I IgG titers greater than 1:800 are relevant. Thresholds and cross-reactivity should be interpreted with the laboratory. Withholding antibiotics to repeat cultures is a selective decision in stable patients, not a strategy for sepsis or complications.

Operative tissue should be divided before fixation: appropriate material for culture and molecular diagnostics, and other material for histology and staining. Broad-range PCR, sequencing and specific tests may identify the organism after antibiotics, but DNA can persist after microorganisms are no longer viable. The result should be correlated with histology, blood cultures and anatomy. Endomyocardial biopsy is not the test for a valvular vegetation; tissue from an embolus or another focus may instead contribute in selected cases. Sample quality affects both positive and negative results.

The 2023 Duke-ISCVID criteria distinguish a pathological pathway and a clinical pathway. The former includes identification of microorganisms in appropriate cardiac, prosthetic or embolic material in the setting of active disease, or histological demonstration of active endocarditis. The clinical pathway combines major and minor categories. The categories organize the evidence and should be documented precisely: multiple methods demonstrating the same imaging lesion do not automatically become multiple independent major criteria. Direct surgical observation also has specific conditions for application and should not be added inappropriately to pathological confirmation.

Clinical classification according to Duke-ISCVID 2023:

• Definite endocarditis: two major criteria, or one major and three minor criteria, or five minor criteria.
• Possible endocarditis: one major and one minor criterion, or three minor criteria, without meeting the definite or rejected category.
• Rejected endocarditis: a firm alternative diagnosis; or resolution without recurrence after fewer than four days of antibiotics; or no pathological or gross evidence at surgery/autopsy after fewer than four days of antibiotics; or failure to meet the possible category.

Major Duke-ISCVID 2023 categories:

• Microbiology: at least two positive blood-culture sets for typical microorganisms, or at least three for organisms that cause endocarditis occasionally or rarely; specific serological or molecular evidence for Coxiella, Bartonella and Tropheryma whipplei is also included.
• Imaging: echocardiography or CT showing vegetation, leaflet perforation or aneurysm, abscess, pseudoaneurysm, fistula, new significant regurgitation or new partial prosthetic dehiscence; pathological PET findings under the specified conditions also contribute to this category.
• Surgery: direct evidence of endocarditis observed during an operation when no major imaging criterion is available and there is no subsequent histological or microbiological confirmation.

Minor Duke-ISCVID 2023 categories:

• Predisposition: cardiac conditions or exposures included in the definition, including previous endocarditis, prosthesis or repair, specific congenital heart diseases, valvular disease greater than mild, intracardiac endovascular device, hypertrophic cardiomyopathy and injection drug use.
• Documented fever greater than 38 °C.
• Vascular phenomena: arterial emboli, septic pulmonary infarcts, cerebral or splenic abscesses, infectious aneurysms, intracranial hemorrhage, conjunctival hemorrhages, Janeway lesions or purpura fulminans.
• Immunological phenomena: immune-complex glomerulonephritis, Osler nodes, Roth spots or positive rheumatoid factor.
• Compatible microbiological evidence that does not meet major-criterion requirements.
• Pathological FDG uptake within three months of implantation of prosthetic material when it does not meet the major criterion.
• New regurgitation identified on auscultation when echocardiography is unavailable; a change in a previously known murmur is not sufficient.

Qualification as a typical microorganism also depends on the substrate. S. aureus, S. lugdunensis, E. faecalis, streptococci excluding S. pneumoniae and S. pyogenes, Granulicatella, Abiotrophia, Gemella and HACEK organisms belong to the main group. In the presence of intracardiac prosthetic material, the definition includes additional organisms, including coagulase-negative staphylococci, Corynebacterium striatum and C. jeikeium, Cutibacterium acnes, Serratia marcescens, Pseudomonas aeruginosa, nontuberculous mycobacteria and Candida. For PET in surgical prosthetic valves, Duke-ISCVID distinguishes the three-month interval; ESC 2023 accepts pathological prosthetic uptake patterns as major evidence regardless of the interval, with expert interpretation. The two definitions should not be selectively combined.

Classification is updated throughout the diagnostic pathway: an initially possible form may become definite, while a finding may ultimately be attributed to degeneration, thrombus, fibroelastoma or non-infective endocarditis. Simple worsening of pre-existing regurgitation is not equivalent to the new major-regurgitation Duke-ISCVID criterion. No score eliminates clinical judgment, particularly after antibiotics or in the presence of prosthetic material. A residual vegetation after treatment may be organized and sterile; growth, new destruction, persistent bacteremia and the clinical course are more informative than echocardiographic persistence alone.

Treatment and prognosis

Management requires an Endocarditis Team integrating cardiology, infectious diseases, microbiology, imaging and cardiac surgery, with neurology and other specialties according to complications. The objectives are to eradicate the organism, control infectious foci, correct mechanical damage and prevent new events. Surgical assessment begins when a plausible indication emerges, not after antibiotics have been completed. In unstable patients, respiratory and circulatory support, specimen collection and antimicrobials proceed rapidly in parallel; transfer to an expert center should be planned before deterioration makes all options more hazardous.

Empirical therapy depends on the substrate, time since implantation, community or healthcare acquisition, previous antibiotics, colonization, allergies and organ function. In community-acquired native-valve or late prosthetic-valve forms, staphylococci, streptococci and enterococci are considered; in early prosthetic and healthcare-associated forms, resistance and Gram-negative organisms become more important. ESC guidance includes, in appropriate settings, ampicillin with ceftriaxone or ampicillin with an antistaphylococcal agent and gentamicin, but actual coverage must be verified against local epidemiology and presentation. A regimen suitable for a subacute form is not automatically adequate for healthcare-associated sepsis. Once the organism is identified, targeted therapy is instituted promptly.

Pharmacokinetics is part of treatment: weight, renal function, dialysis, volume of distribution and interactions modify dose and interval. Adult regimens reported in guidelines require individual adjustment, particularly in organ failure. Exposure monitoring is relevant for vancomycin and aminoglycosides and may be useful for other drugs in complex situations. Complete blood count, creatinine, electrolytes, transaminases and molecule-specific monitoring make it possible to maintain effective therapy without accumulating avoidable toxicity. Persistent bacteremia requires both verification of drug exposure and a search for the focus that makes medical therapy insufficient.

For fully susceptible streptococci, penicillin G, amoxicillin or ceftriaxone are reference options according to susceptibility and context. In adults, ceftriaxone 2 g intravenously once daily is a usual regimen; standard duration is four weeks for a native valve and six weeks for a prosthetic valve. The short two-week combination with an aminoglycoside applies only to selected uncomplicated native-valve forms with a susceptible strain and adequate renal function. Higher MICs, metastatic foci and organisms such as Granulicatella or Abiotrophia require distinct regimens. Eligibility for a short regimen depends on the complete case, not merely on membership in the streptococcal group.

For MSSA, an antistaphylococcal beta-lactam or cefazolin is preferred when usable. Adult regimens include oxacillin or (flu)cloxacillin at approximately 12 g/day in divided doses, or cefazolin generally 6 g/day, with necessary adjustments. Vancomycin is not an equivalent substitute merely because it is active in vitro. In native-valve staphylococcal endocarditis, routine addition of gentamicin increases toxicity without demonstrated clinical benefit. Duration is generally in the range of 4-6 weeks and is defined by site, complications and regimen; abbreviated regimens for selected right-sided forms do not apply to left-sided disease.

For MRSA, vancomycin is dosed using exposure monitoring consistent with renal function and susceptibility. Daptomycin is an alternative in specialist regimens; ESC guidance proposes high doses, generally 10 mg/kg/day, combined with an appropriate second agent to improve activity and limit resistance. Creatine kinase monitoring and toxicity surveillance are required. Persistent bacteremia requires investigation for abscess, infected prosthesis or catheter, thrombophlebitis and metastatic foci before attributing the problem solely to the antimicrobial. In staphylococcal prosthetic-valve endocarditis, rifampin and other combinations require separate evaluation because the clinical benefit of historical combinations is less certain than their microbiological plausibility.

In enterococcal endocarditis caused by E. faecalis, ampicillin with ceftriaxone achieves synergy through different cell-wall targets and reduces the need for aminoglycoside exposure. A usual adult regimen combines ampicillin 12 g/day in divided doses and ceftriaxone 2 g every 12 hours for six weeks. It can also be used in the presence of high-level aminoglycoside resistance, which instead abolishes the expected synergy with gentamicin. When gentamicin is chosen, susceptibility, duration of that component and monitoring should be explicit. E. faecium or resistant enterococci require choices guided by species and susceptibility testing, without automatically extending the E. faecalis regimen.

For HACEK organisms, ceftriaxone 2 g/day is a reference regimen for four weeks on a native valve and six weeks on a prosthetic valve. Alternatives depend on susceptibility, beta-lactamase production and tolerance. Non-HACEK Gram-negative organisms may require combinations and surgical source control, with choices determined by the organism and resistance profile. In blood culture-negative forms, empirical therapy after antibiotics should cover plausible usual organisms, whereas strong evidence for an intracellular pathogen leads to specific treatment. Negative cultures do not by themselves justify an indefinite combination of drugs directed against every possible microorganism.

Q fever endocarditis is treated with doxycycline and hydroxychloroquine in a prolonged regimen, generally for at least 18 months on a native valve and 24 months on a prosthetic valve, with clinical, serological, ophthalmological and pharmacological monitoring. Recommendations for Bartonella differ: ESC 2023 describes doxycycline with initial gentamicin; the 2025 AHA statement proposes regimens with doxycycline, or azithromycin, and rifampin, prioritizing reduced renal risk in a setting often complicated by glomerulonephritis. Durations and components should belong to one coherent regimen rather than an arbitrary combination of documents. Brucella and T. whipplei require dedicated pathways.

In fungal endocarditis, identification of the organism determines antifungal treatment, surgery and suppressive therapy. For Candida, regimens use high-dose echinocandins or lipid-formulation amphotericin B, possibly combined with flucytosine, with step-down to an azole only when susceptibility and response permit. Surgery is frequently part of treatment, and prolonged suppression may be required when the focus cannot be eradicated. For Aspergillus, voriconazole or lipid-formulation amphotericin B and early surgical assessment are specific reference strategies. Transient improvement does not eliminate the risk of late recurrence related to residual material or foci.

The duration of antibacterial therapy is counted from the first day of effective treatment, with clearance of blood cultures used as a reference when they were initially positive. Surgery does not automatically restart the count; a positive valve culture may instead require a new full course. PCR positivity alone does not establish viability. If a native valve is replaced during treatment, the presence of the new prosthesis does not retroactively transform the episode into prosthetic-valve endocarditis. Deep foci and particular organisms may alter duration and should be documented in the final prescription.

Oral step-down therapy is possible in selected patients after an adequate intravenous phase, clinical stability, microbiological control and absence of unresolved surgical indications. In the POET trial, 400 stable adults with left-sided endocarditis caused by selected organisms were assigned to continued intravenous treatment or oral combinations; the composite outcome occurred in 12.1% and 9.0%, respectively, meeting noninferiority. The result applies to a structured strategy, not to arbitrary oral monotherapy. Absorption, interactions, adherence, susceptibility and close follow-up are substantial conditions. OPAT is another possibility, while retaining vascular-access risks and the need for surveillance.

Surgery is considered for heart failure, uncontrolled infection and prevention of embolism. Refractory pulmonary edema or shock due to regurgitation, obstruction or fistula may require emergency surgery, generally within 24 hours. Urgent surgery, typically within 3-5 days, applies to many other conditions that do not permit waiting until antibiotics are completed. Response to diuretics can stabilize the patient but cannot repair the defect. Anatomy, haemodynamic tolerance and potential for recovery should be explicit in the decision and reassessed if conditions change.

Uncontrolled infection may be local, with abscess, fistula, pseudoaneurysm, dehiscence or progression of the lesion, or may manifest as persistent bacteremia or sepsis despite an effective regimen and control of other sources. New heart block in the setting of aortic disease strengthens suspicion of invasion. Fungi, difficult resistance patterns and infected material may reduce the probability of medical cure. Fever alone does not establish this scenario: drug reactions, emboli, access sites and extracardiac foci should be distinguished because changing antibiotics without identifying the reservoir does not solve the problem.

Surgical prevention of embolism integrates vegetation size and mobility, site, organism, previous events and operative risk. A persistent vegetation of at least 10 mm with embolization despite appropriate therapy constitutes an important indication; an isolated mass of the same size does not determine identical urgency in every patient. The Kang trial supports a benefit of early surgery in a selected population with severe valvular disease and large vegetations, mainly through reduction of embolic events. The results should not be indiscriminately extrapolated to extreme frailty, invasive prosthetic disease or devastating neurological injury.

The operative principle is debridement of infected and necrotic tissue with drainage and stable reconstruction. Mitral or tricuspid repair can preserve function when it allows complete source control; avoiding a prosthesis does not justify leaving infected tissue behind. In invasive aortic disease, annular reconstruction, root replacement or reconstruction of the intervalvular fibrous body may be required. The choice of biological, mechanical or homograft material depends on anatomy and the patient; no substitute by itself eliminates the risk of reinfection. Risk scores support comparison between surgical risk and the risk of leaving the infection uncorrected, without replacing that judgment.

Neurological complications modify timing and strategy. After TIA or ischemic stroke without hemorrhage, a strong cardiac indication may permit surgery without delay when the neurological prognosis is not unfavorable. Intracranial hemorrhage requires neurovascular assessment and often delay if the cardiac condition allows; instability may require a different balance. Intravenous thrombolysis is not recommended for stroke due to endocarditis, whereas thrombectomy may be considered in large-vessel occlusion. An infectious aneurysm may require endovascular or surgical treatment before cardiac surgery.

Anticoagulants and antiplatelet drugs are not initiated to prevent detachment of vegetations. A pre-existing antithrombotic indication, especially a mechanical prosthesis, should be reassessed in relation to hemorrhage, stroke, procedures and interactions; temporary management with heparin may be necessary. Sepsis does not justify ignoring the thrombotic risk of the prosthesis, just as the prosthesis does not justify ignoring a cerebral hemorrhage. Mechanical support for shock is considered as a bridge to a definitive plan and depends on anatomy: severe aortic regurgitation can make some support devices inappropriate.

Source control includes catheters, abscesses, spondylodiscitis, joint infections and portals of entry. A splenic abscess may require drainage or splenectomy, but sequencing relative to cardiac surgery depends on urgency and risk of reinfection: a valve causing refractory edema cannot always wait for complete eradication of every extracardiac focus. In infected devices, antibiotics without system removal often fail to achieve eradication; extraction and the need for reimplantation belong to a dedicated pathway. Every focus should have a documented objective and control plan before discharge.

Secondary prevention includes oral and skin hygiene, management of access sites and dental prophylaxis in high-risk categories for the indicated procedures. Isolated native valvular disease does not automatically require prophylaxis. In patients who inject drugs, treatment of substance-use disorder, harm reduction and continuity of care are part of infection management; recommending abstinence alone does not address the determinant of recurrence. Colorectal evaluation associated with S. gallolyticus and other investigations of the portal of entry should be completed according to urgency and clinical condition.

At the end of therapy, a reference echocardiogram documents regurgitation, function, prostheses and residual lesions. Follow-up monitors heart failure, arrhythmias, renal function, infectious foci and nutritional recovery; recurrent fever or chills require new cultures before antibiotics whenever possible. Relapse and reinfection are distinguished by chronology and microbiological comparison. Mortality depends mainly on shock, heart failure, organism, extent of disease, neurological injury and comorbidities; failure to perform surgery when indicated identifies a high-risk group, although observational comparisons with operated patients are strongly affected by selection.

Complications

Acute heart failure may rapidly progress to pulmonary edema and shock when valvular destruction exceeds compensatory capacity. Sepsis may add vasodilation and myocardial depression, producing a mixed picture in which pressure, perfusion and filling must be reassessed continuously. Pharmacological stabilization is a bridge to correction when a critical mechanical lesion exists. In the long term, residual regurgitation and ventricular injury may sustain heart-failure risk even after microbiological eradication.

Perivalvular complications include abscess, pseudoaneurysm, fistula and dehiscence. An abscess may evolve without a conspicuous vegetation; a fistula creates an abnormal shunt and dehiscence alters prosthetic support. Atrioventricular block and arrhythmias may signal septal extension, whereas a temporary pacing device treats only the electrical consequence. Anatomical definition with echocardiography and CT is essential for planning reconstruction that eliminates all infected cavities.

In the central nervous system, stroke and hemorrhage may result from embolization, hemorrhagic transformation, infection of the vascular wall or abscesses. Identification of an infectious aneurysm changes the risk of procedures and anticoagulation. Silent lesions may contribute to assessment but do not all have the same significance: isolated microbleeds are not equivalent to a recent major hemorrhage. Neurological outcome and potential for recovery should be discussed together with the risk of new emboli and the cardiac risk of waiting.

The spleen, kidneys, bowel and limbs may develop infarction, abscess or acute ischemia; the coronary arteries may be affected by emboli or adjacent anatomical complications. In right-sided forms, septic pulmonary emboli, cavitation, abscesses and empyema may dominate the respiratory course. A metastatic focus can sustain fever and bacteremia even after valvular improvement. A search guided by symptoms and clinical course helps avoid both overlooking a reservoir and attributing every radiological finding to active infection without clinical corroboration.

Renal injury may combine sepsis, hypoperfusion, congestion, emboli, glomerulonephritis and nephrotoxicity. Urinary sediment, haemodynamics, drug exposure and imaging help distinguish the mechanisms. An immunological picture may include hypocomplementemia and autoantibodies, sometimes mimicking vasculitis: starting immunosuppression without recognizing the infection can worsen the process. At the same time, renal and hepatic failure modify antimicrobial and antithrombotic therapy, creating a cycle that requires timely adjustments.

Recurrences may reflect persistence of infected material, an undrained abscess, inadequate antimicrobial exposure or a new bacteremia. Recovery of the same microorganism points toward relapse, but distinction may require typing and comparison of profiles. Late toxicities, vascular-access complications, bleeding and prolonged deconditioning are part of the overall outcome. Completion of the antimicrobial course should therefore be accompanied by a plan for residual lesions, prevention, rehabilitation and rapid access to reassessment, not merely a date for stopping medication.