Pathogenesis of vegetation formation in infective endocarditis

INTRODUCTION — Infective endocarditis (IE) arises when an adherent platelet-fibrin nidus becomes secondarily infected and produces vegetations, which in turn may directly damage the endocardial tissue and/or valves. The pathogenesis of infective endocarditis will be reviewed here.

Other aspects of infective endocarditis, including clinical consequences of vegetation formation, are discussed separately. (See "Native valve endocarditis: Epidemiology, risk factors, and microbiology" and "Prosthetic valve endocarditis: Epidemiology, clinical manifestations, and diagnosis" and "Right-sided native valve infective endocarditis" and "Infective endocarditis in children" and "Clinical manifestations and evaluation of adults with suspected left-sided native valve endocarditis" and "Antimicrobial therapy of left-sided native valve endocarditis" and "Antimicrobial therapy of prosthetic valve endocarditis" and "Surgery for left-sided native valve infective endocarditis" and "Complications and outcome of infective endocarditis".)

PATHOGENESIS

Vegetation formation — The endothelial lining of the heart and its valves is normally resistant to infection with bacteria and fungi. Experiments in animal models have demonstrated that a sequence of interrelated events must occur before microbes can establish an infective nidus or vegetation on the endocardium.

It is likely that the pathogenesis of vegetation formation varies depending on the infecting organism. In most instances, the initial step in the establishment of a vegetation is endocardial injury, followed by focal adherence of platelets and fibrin. The initially sterile platelet-fibrin nidus becomes secondarily infected by microorganisms circulating in the blood, either from a distant infection or as a result of transient bacteremia from a mucosal or skin source [1,2]. Subsequent microbial growth results in further activation of the coagulation system via the extrinsic clotting pathway, adherent monocytes release a variety of cytokines, and activated endothelial cells continue to lead to further local deposition of fibronectin. These processes culminate in a macroscopic excrescence or vegetation. Bacterial growth occurs within cells and within the matrix of fibronectin inside vegetations, making it difficult for host immune responses to control or eradicate the ongoing infection (figure 1 and figure 2) [3].

Some organisms with high virulence are capable of infecting normal human heart valves, such as Staphylococcus aureus. One proposed mechanism involves binding of S. aureus bacteria to endothelial cells through extracellular matrix binding proteins and then subsequent internalization into the endothelial cell. Once internalized, S. aureus can evade the host immune response, further activate endothelial cells, and induce tissue destruction by the release of exoproteins (figure 1 and figure 2) [3].

A mouse model of IE illustrates the two pathways of vegetation formation by S. aureus. One pathway, the “damage-induced” IE pathway, supports earlier models which demonstrate binding of bacteria to the platelet-fibrin layer on the damaged endothelium. S. aureus adhesins such as ClfA are heavily involved in this process. The other pathway, the “inflammation-induced” IE pathway does not require a damaged endothelium. In the “inflammation-induced” IE model, inflammation (induced by histamine infusion) causes the release of von Willebrand factor and extensive platelet recruitment to the valve surface. Within these platelet aggregates, S. aureus bacterial cells become “entrapped” and activated platelets and neutrophils amplify the growing vegetation. ClfA does not appear to be important in this latter pathway. The “inflammation-induced” IE model demonstrates one way that S. aureus IE can develop in the setting of bacteremia-induced inflammation and widespread endothelial activation in the absence of pre-existing heart valve damage. Von Willebrand factor, which allows bacteria to withstand the shear forces of blood flow and thus adhere to the valve, is important for the early stages of bacterial adhesion in both pathways, through slightly different mechanisms [4].

Bacterial-platelet and neutrophil-platelet interactions also appear to be important in the development and enlargement of a vegetation. For example, vegetations may contain layers of neutrophil extracellular traps that interconnect bacteria-platelet aggregates within vegetations. These extracellular traps may serve as "scaffolds," or meshes, that both further entrap bacteria-platelet aggregates and enhance the growth of vegetations [5]. Some pathogens such as streptococci may interact with platelets to form biofilms with multilayer architecture that are embedded within mats or tiny nidi within vegetations [6].

The significance of vegetation size is discussed further separately. (See "Surgery for left-sided native valve infective endocarditis", section on 'Vegetation characteristics and risk of embolization'.)

Histopathologic features — A mature vegetation is an amorphous collection of fibrin, platelets, leukocytes, red blood cell debris, and dense clusters of bacteria. The surface generally consists of fibrin and varying numbers of leukocytes. Clumps of bacteria, histiocytes, and monocytes are generally observed deep within the vegetation; giant cells containing phagocytosed bacteria may also be seen. Extremely high concentrations of bacteria (eg, 10⁹ to 10¹¹ colony forming units/gram of tissue) may accumulate deep within vegetations; some of these organisms exist in a state of reduced metabolic activity. Vegetations are avascular structures; following therapy and during the healing process, capillaries and fibroblasts appear.

Transmission electron-microscopy (TEM) allows for further discrimination of vegetation structure. Using TEM to analyze a vegetation from a clinical S. aureus IE case, investigators described three distinct layers: a bacteria-rich region with necrotic tissue and cellular debris, a second bacteria-rich region in a homogeneous matrix resembling a biofilm, and an outermost region composed of platelets, other blood cells, fibrin, and connective tissue [7]. These findings support the concept of a vegetation as a protected niche where bacteria can persist. In addition, these findings, at least in the setting of S. aureus, suggest a process of “inside-out necrosis” within vegetations.

In one study comparing vegetation histology of infected prosthetic valves and sterile (but dysfunctional) native valves, the former demonstrated microorganisms and inflammatory infiltrates containing neutrophils, while the latter demonstrated inflammatory adherent thrombi containing mostly lymphocytes and macrophages; microorganisms were not present [8].

Effect of antibiotic therapy — Following treatment of endocarditis, vegetations may disappear, persist, or, uncommonly, increase in size [9-11]. In an observational study including 134 patients with IE who did not undergo surgery, residual vegetation at the end of antibiotic therapy was present in 49 percent of cases; the vegetation size was increased in 7 percent of cases [10]. In another observational study including 30 patients with native valve IE who did not undergo surgery, the mean vegetation area decreased from 1.70 cm² (SD 2.12) to 0.78 cm² (SD 1.46) following antibiotic therapy [11].

A retrospective histopathology study among 480 patients with endocarditis who underwent valve replacement or repair noted the following findings [12]:

●Valvular specimens from patients undergoing surgery prior to completion of antimicrobial therapy demonstrated microorganisms by Gram stain and/or histologic examination in most cases (81 and 67 percent, respectively).

●Detection of microorganisms does not imply treatment failure in patients undergoing surgery prior to completion of antimicrobial therapy. The only reliable method to determine the efficacy of treatment in such cases is culture of the excised tissue; clearance of dead organisms via lysis or phagocytosis may take months.

Evaluation of vegetations with polymerase chain reaction may demonstrate the presence of bacterial DNA months to years after completion of appropriate therapy; such findings likely represent persistence of nonviable bacterial debris within the vegetation, although, in some cases, such results may reflect persistence of viable organisms deep within the vegetation [13].

Host factors

Endocardial injury — It is virtually impossible to induce endocarditis in experimental animals unless the valvular endocardium is first traumatized with a polyethylene catheter inserted into the right or left side of the heart [1]. A corollary to this observation is that animals with prior catheter-induced valvular lesions that are injected with bacteria such as viridans streptococci (which have the inherent ability to adhere to fibrin-platelet matrices on damaged endocardium) predictably develop a vegetation at the site of valve injury within a few days. Similar lesions probably occur naturally in bacteremic humans with congenital or acquired cardiac lesions that induce continuous endocardial trauma via regurgitant flow or high-pressure jets of blood through stenotic lesions. S. aureus, on the other hand, is uniquely able to adhere to structurally normal heart valves, as noted above.

Vegetation sites — Vegetations tend to develop at sites where blood travels from an area of high pressure through a narrow orifice into an area of lower pressure. The explanation for this phenomenon can be deduced from in vitro experiments demonstrating the physics of turbulent flow. As an example, it has been observed that, following injection of nebulized bacteria into an air stream passing through an agar-coated tube, the highest concentration of bacteria is found in the low pressure area immediately distal to the narrowing [14].

The propensity for vegetations to form at specific sites may be correlated with a decrease in lateral pressure downstream from the regurgitant flow, which causes a decrease in the perfusion of the intimal lining at these sites [15]:

●In the setting of pre-existing valvular disease, vegetations are usually located on the atrial surface of incompetent atrioventricular valves or the ventricular surface of incompetent semilunar valves.

●In the setting of ventricular septal defect, vegetations tend to develop on the orifice of the defect, on the right ventricular side of the opening, and/or secondarily on the tricuspid and pulmonic valves [16].

●In the setting of aortic insufficiency, vegetations may localize to the chordae tendineae of the anterior leaflet of the mitral valve.

●In the setting of mitral regurgitation, vegetations may develop on the wall of the left atrium, where the regurgitant jet strikes the atrial wall and results in endocardial thickening (MacCallum's patch).

In addition, presence of mitral annular calcification (a chronic inflammatory and degenerative process) may serve as a nidus for S. aureus vegetations [17].

Pathogen factors

Source of infection — The bacterial source for endocarditis may be readily discernible (such as a dental abscess, infected skin lesion, or infected vascular catheter) or there may be no clear history of antecedent infection. In such cases, the source is frequently attributed to minor trauma of the oropharyngeal, gastrointestinal, or genitourinary mucosa.

Endocarditis in injection drug users is presumed to be a consequence of bacterial contamination of material injected, of injection equipment, and/or of the skin surface at the injection site.

Microbial adherence — Microbial adherence is a crucial event in the pathogenesis of endocarditis. Organisms typically associated with endocarditis have the capacity to adhere avidly to valve tissue; these include S. aureus, viridans streptococci, enterococci, and Pseudomonas aeruginosa [18]. Bacterial adherence to a platelet-fibrin nidus involves a complex interaction of microbial cell wall components. The mechanism of adherence is poorly understood and may vary among different organisms.

In vitro models have provided some data regarding the pathogenesis of adherence:

●The relative frequency with which different species of bacteria cause endocarditis in humans has been correlated with their ability to cause endocarditis in rabbits [15]. In this experimental model, the amount of dextran produced by the cell wall of streptococci grown in broth is proportional to the intrinsic ability of various Streptococcus species to cause endocarditis, suggesting the importance of the interaction between dextran and the valvular endothelium [19].

●Intrinsic binding affinity to fibronectin appears to be a contributing pathogenic factor; in rats, S. aureus mutants with diminished capacity to bind fibronectin had reduced ability to cause endocarditis compared with parent strains with higher fibronectin-binding affinity [20]. In the mouse model of IE, the fibrinogen binding protein, ClfA, is an important mediator of S. aureus binding to damaged valves [4].

●Glucosyltransferases, proteins made by viridans streptococci, have the ability to convert sucrose into polysaccharides which in turn may act as "modulins" to induce various cytokines that in turn recruit leukocytes to vegetations [21].

Other potential mediators of microbial adherence include fibrinogen, laminin, and type 4 collagen. These mediators may be particularly important for specific strains of bacteria that have the propensity to cause endocarditis. As an example, strains of Streptococcus mutans that express certain cell collagen binding proteins have high fibrinogen binding rates in vitro and have a propensity to cause endocarditis in vivo [22].

Microbial virulence — Proteome analysis of IE vegetations has demonstrated that a variety of species-specific proteins, including virulence factors, are found in vegetations. For example, S. aureus vegetations contain virulence factors such as Panton-Valentine Leukocidin and proteins involved in iron binding; vegetations associated with other microbes have other specific-specific proteins [23]. It has been hypothesized that proteolysis of vegetation components may arise from proteases within vegetations and contribute to destabilization of vegetations and subsequent embolism [23].

SUMMARY

●In general, the initial step in the establishment of a vegetation is endocardial injury, followed by focal adherence of platelets and fibrin. The initially sterile platelet-fibrin nidus becomes secondarily infected by microorganisms circulating in the blood. Microbial growth results in the secondary accumulation of more platelets and fibrin and further activation of the coagulation system via the extrinsic clotting pathway. Staphylococcus aureus vegetations can be induced by either endocardial injury or inflammation. (See 'Vegetation formation' above.)

●A mature vegetation is an amorphous collection of fibrin, platelets, leukocytes, red blood cell debris, and dense clusters of bacteria. The surface generally consists of fibrin and varying numbers of leukocytes. Clumps of bacteria, histiocytes, and monocytes are generally observed deep within the vegetation; giant cells containing phagocytosed bacteria may also be seen. (See 'Histopathologic features' above.)

●Detection of microorganisms on histopathology examination of an excised valve does not imply treatment failure; clearance of dead organisms via lysis or phagocytosis may take months. Similarly, detection of bacterial DNA via polymerase chain reaction months to years after completion of appropriate therapy likely reflects persistence of nonviable bacterial debris within the vegetation. (See 'Effect of antibiotic therapy' above.)

●Vegetations tend to develop at sites where blood travels from an area of high pressure through a narrow orifice into an area of lower pressure. (See 'Vegetation sites' above.)

●The bacterial source for endocarditis may be readily discernible (such as a dental abscess, infected skin lesion, or infected vascular catheter) or there may be no clear history of antecedent infection. In such cases, the source is frequently attributed to minor trauma of the oropharyngeal, gastrointestinal, or genitourinary mucosa. (See 'Source of infection' above.)