The long and short of vegetations in infective endocarditis

Infective endocarditis (IE) is a multifaceted disease. Outcomes and prognosis are influenced by host-related factors (demographics, comorbidity, predisposing conditions or procedures and affected valve), causative pathogen, local or systemic complications and therapeutic intervention (medical or surgical). The incidence of community-acquired IE in the USA ranges from 5–7 cases per 100,000 person-years and does not appear to have changed over the last 30 years [1]. Tertiary referral centers around the globe are in the midst of a changing trend with a surge of Staphylococcus aureus IE [2,3] as a consequence of ‘medical progress’ involving increased utilization of cardiac devices, long-term intravascular catheters, hemodialysis, growing numbers of immune suppressed patients, injection drug use and the escalating prevalence of methicillin-resistant S. aureus in our hospitals. Despite the advent of effective antimicrobial therapy, advances in surgical techniques and intensive care, IE mortality remains unacceptably high [4,5].

Vegetations – the hallmark of IE – constitute one of the major Duke criteria for the diagnosis of IE [6]. Damage to the cardiac endothelium followed by platelet and fibrin deposition results in a sterile lesion known as nonbacterial thrombotic endocarditis (NBTE) or vegetation. Microbes entering the bloodstream via routine or invasive procedures adhere to the vegetation and are engulfed in an outer meshwork of fibrin and platelets. Subsequent rapid microbial multiplication in a protected area of impaired host defense leads to vegetation growth and continuous bacteremia [7]. Bacterial density in vegetations can reach 10⁹–10¹¹ per gram of tissue and organisms at the heart of the vegetation can be metabolically inert [7]. Ensuing complications can be local (e.g., congestive heart failure, cardiogenic shock, progressive valvular damage, septal rupture and intracardiac abscess) or distant (e.g., systemic embolization and persistent bacteremia) [8]. Congestive heart failure and systemic embolization (especially to the brain) are major contributors to substantial morbidity and mortality in IE [4,9,10].

Although integral to the pathophysiology and diagnosis of IE, the contribution of vegetation characteristics (size, length, mobility, density, location and number) to subsequent outcomes is the subject of much controversy. This uncertainty stems from the following:

•	Varying rates of utilization of transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) with inconsistent visualization of vegetations;
•	Timing of echocardiography with respect to onset of patient symptoms and infrequent follow-up echocardiograms to assess vegetation status;
•	Lack of consistency and uniformity in the reporting of size/length, mobility and density of vegetations;
•	Echocardiographers’ expertise in the accurate identification of structures that mimic vegetations (abnormal valve thickening in degenerative valve diseases, prior rheumatic valve disease, Libman–Sacks endocarditis or ruptured chordae tendineae);
•	Persistence of vegetations despite appropriate therapy;
•	Potential variability of vegetation size with microbial etiology of IE;
•	Contributions from host- and microbe-associated biological factors independent of macroscopic vegetation characteristics;
•	Paucity of carefully performed, prospective studies with adequate sample sizes dedicated to examining the aforementioned relationships.

Several investigators have attempted to explore the association between vegetation characteristics and subsequent complications in IE. Some studies have identified a higher risk of embolic events (EEs) with vegetations larger than 10 mm [10–14], while others have not [15]. The latter study was performed before the era of TEE and vegetations were visualized by TTE in only 38% of patients [15]. Perhaps the most elaborate study to date, evaluating vegetation characteristics on outcome in IE, was reported by Thuny and colleagues [10]. In this multicenter prospective study, 384 patients with definite IE by Duke criteria underwent TTE, TEE, cerebral and thoracoabdominal CT scans, blood cultures and serological studies within 48 h of hospital admission. Each echocardiogram was interpreted by two experienced echocardiographers who were unaware of the patients’ clinical status. They used predetermined criteria for the assessment of vegetation length, mobility and intracardiac abscess. Overall, 34% of patients had major EEs before or after diagnosis of IE (total EE) and 7.3% had EEs after the initiation of antibiotic therapy (new EE). Vegetation length of greater than 10 mm and severe vegetation mobility were predictors of new EEs, even after adjustment for S. aureus and Streptococcus bovis (which were independently associated with total EE). Vegetation length of greater than 15 mm was an independent predictor of 1-year mortality. It is unclear whether IE affecting a particular valve leads to a higher probability of embolization; mitral valve involvement has been suggested as conferring a higher embolic risk, especially with vegetations over 10 mm [12–14,16], while others have concluded that embolic risk is independent of vegetation location [10,11].

The influence of antimicrobial therapy on vegetations and subsequent risk of systemic embolization has also been addressed. There is general agreement that the risk of embolization diminishes rapidly following initiation of effective antimicrobial therapy; 65–70% of embolization occurred either prior to or during the first 2 weeks of antibiotic therapy [10,14,15]. Some investigators have performed serial echocardiograms to assess the state of vegetations during antibiotic therapy. Rohmann et al. reported that, in patients with increasing or unchanged vegetation size during antibiotic therapy, a higher rate of complications (valve surgery, EEs, perivalvular abscess and mortality) were observed compared with those with shrinking vegetations [17]. In a subsequent study from the same group, some antibiotics were believed to reduce vegetation size more than others [18]. Vilacosta et al. noted enlarging vegetations in 10.5% of their patients on a subsequent TEE; however, repeat echocardiograms were performed during a brief follow-up period of 14 ± 10 days [14]. Vegetations can persist despite effective antibiotic therapy. In a study in which serial TTE was performed, 71% of the vegetations were still detectable at the end of antibiotic therapy [19]. A total of 59% had no change in size, while 52% of vegetations were more dense. Vegetation persistence did not correlate with late complications. Hence, in the absence of enlarging vegetations, their continued detection during appropriate medical therapy may represent ongoing ‘healing’ and is not in itself a cause for concern [14,17].

The question of how the presence of vegetations affects the decision regarding valve surgery in IE is even more controversial. Recent large, carefully conducted, observational studies have tried to address the enduring question of mortality benefit from valve surgery in IE [20–24]. Propensity analysis was utilized to adjust for heterogeneity and confounding since randomization to medical therapy or valve surgery will be construed as unethical. Vegetation characteristics (other than presence and location) could not be elaborated in these trials as the logistics of obtaining such information retrospectively or from multiple sites even in prospective studies were conceivably prohibitive. Large vegetations (>10 mm), persistent vegetation(s) after systemic embolization and at least one EE during the first 2 weeks of antimicrobial therapy are consistently cited as ‘accepted’ indications for valve surgery in IE [8,25]. The reality and practicality of this intricate matter is more convoluted, to say the least. In the absence of additional complications necessitating operative intervention, physicians and surgeons are reluctant to subject patients with IE to valve surgery based on presence of vegetations alone. Moreover, with an acute cerebral EE, cardiac surgeons often opt to defer valve surgery because of the potential for cerebral bleed during intraoperative heparinization. Although ‘uncomplicated, nonhemorrhagic embolic stroke’ appears to have a favorable outcome after valve surgery for IE, evidence of cerebral hemorrhage prior to surgery enhances perioperative mortality [26,27].

Basic science research in the arena of microbe–vegetation–host interaction reminds us of several additional ‘invisible’ variables that might play a prominent role in the natural history and complications of IE. Some of these hypotheses have been studied clinically with varying results. Adherence of oral streptococci to NBTE lesions and ability to cause IE may depend on the production of dextran, a complex extracellular polysaccharide [28]. Ongoing synthesis of dextran during experimental IE correlates with vegetation size and resistance to antimicrobial therapy. Platelets appear to play a major role in the pathogenesis of IE and vegetation propagation. Interactions between certain pathogens, platelets and endothelium enhance their propensity to cause IE [7,28,29]. Strains of streptococci and staphylococci capable of producing IE aggregate platelets more actively than bacteria that seldom cause IE. Hence, platelet inhibition is an attractive concept to potentially prevent progressive vegetation growth and subsequent embolization. In animal models of IE, aspirin was shown to reduce vegetation size and bacterial density significantly [30,31]. A recent prospective, randomized, double-blinded, placebo-controlled trial of high-dose aspirin treatment for 4 weeks in patients with IE did not find any benefit of aspirin therapy in reducing the risk of EEs and vegetation resolution; in fact, there was a higher risk of bleeding [32]. Dosage of aspirin higher than necessary for platelet inhibition and initiation of aspirin after the diagnosis of IE with delayed antiplatelet activity have been cited as probable reasons for the lack of benefit of aspirin in this trial. A retrospective study assessing the impact of continuous daily antiplatelet therapy for at least 6 months prior to the onset of IE found a significant reduction in symptomatic EEs [33]. However, there was no mortality benefit, the impact on vegetation characteristics was not studied, and it is unclear whether antiplatelet therapy was continued following the diagnosis of IE [33]. Thrombin-induced platelet microbicidal protein (PMP) – a small, cationic platelet-derived peptide – is lethal to endovascular pathogens. PMP-resistant S. aureus is associated with enhanced progression and dissemination in experimental IE [34]. Isolation of PMP-resistant S. aureus in IE has been found to correlate with an intravascular device source for bacteremia [35]. Several other microbial and immunologic contributors to the pathogenesis of IE are under intense scrutiny.

Notwithstanding the previously discussed trials pertaining to vegetations in IE, several fundamental questions remain:

•	Does vegetation size truly matter? If so, what is the ‘critical’ size above which embolization tends to occur?
•	Is vegetation mobility an independent contributor to complications above and beyond vegetation size?
•	Should large and mobile vegetations in themselves be an indication for surgical therapy, especially if detected early in the course of IE? Since the risk of embolization diminishes during the first 2 weeks of effective antimicrobial therapy, it is extremely difficult to determine the timing of valve surgery prior to embolization. Moreover, the risks of open heart surgery, patients’ underlying comorbidities and the need for life-long anticoagulation (with metallic valves) will have to be considered carefully before valve replacement;
•	Does an initial ‘major’ embolization increase the chances of recurrence? If so, will valve surgery prevent additional emboli? As discussed earlier, most surgeons tend to refrain from immediate cardiac surgery in the setting of a recent embolic stroke (especially in the presence of hemorrhage). Also, it is unclear whether all patients with IE and major embolization benefit from urgent valve surgery or only those with persistent vegetations following EE;
•	Do ‘asymptomatic’ emboli enhance the risk of subsequent ‘major’ emboli? If so, should we be performing serial imaging studies (e.g., CT scans) to detect such emboli and consider surgical intervention if they are identified?
•	Is it necessary to perform serial echocardiograms to follow the evolution of vegetations? If so, what is the appropriate time interval for a repeat echo (1 week, 2 weeks or end of therapy)? On the one hand, enlarging vegetations appear to increase the risk of embolization; on the other, vegetations do persist following completion of effective antimicrobial therapy in a significant proportion of patients. Also, diminishing vegetation size during therapy could represent resolving infection or asymptomatic embolization of part of the vegetation;
•	Does the choice of antimicrobial therapy influence the vegetation size and density based on its ‘vegetational penetrability’ and alter the risk of subsequent complications?
•	Is there any role for antiplatelet therapy in patients with IE? If so, should it be commenced prophylactically in patients with an increased future risk of IE? Initiation of antiplatelet therapy after the diagnosis of IE may not be useful given the lack of sufficient time for such therapy to exert any influence on preventing embolization. Antiplatelet therapy may also increase the risk of bleeding if patients need urgent valve surgery. The number of patients who require antiplatelet agents for an unknown duration to prevent a single case of IE may not be cost effective and the bleeding risk may outweigh perceived benefits.

In summary, I believe we are confronted with more questions than answers regarding the perplexing issue of what vegetations represent and predict in patients with IE. Currently, our ability to forecast complications and outcomes based on macroscopic vegetation characteristics is a gamble at best. Validated and standardized parameters to estimate vegetation size, mobility and density should be utilized in future investigations. Contributions from pathogen-associated virulence factors and host-derived immunologic responses toward morphologic vegetation characteristics need to be elucidated further. Given the heterogeneity involved, crucial management decisions in IE will continue to occur on a case-by-case basis. Let us be optimistic and anticipate that the next generation of investigators will shed more light on the darker side of vegetations.

Financial disclosure

The author has no relevant financial interests related to this manuscript, including employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.