Introduction

The first descriptions of endocarditis appeared in the literature in the 1600s at a time when the anatomy and physiology of the heart was still incompletely characterized [ ]. Since that time, significant progress in medicine, including advances in imaging techniques, antimicrobial therapy, and surgical interventions, have ushered in a new era of management of patients with infective endocarditis (IE). This disease that once perplexed notable figures in the history of medicine, including Giovanni Morgagni and Rudolf Virchow, continues to challenge clinicians in the current era, though undoubtedly outcomes have significantly improved [ , ]. In the following sections, we will outline current trends in IE that offer opportunities and challenges for diagnosis and management in the future. Additionally, we will highlight areas of continued work in an effort to better understand this complex disease process.

Trends in intravenous drug use

The opioid epidemic is an alarming trend and has been deemed a national public health crisis in the United States [ , ]. Consistent increases in abuse of prescription opiates as well as disturbing trends in the use of heroin and fentanyl have been observed over the past decade [ ]. The incidence of heroin-associated drug overdose deaths increased from 0.7 per 100,000 to 4.9 per 100,000 between 1999 and 2017, though more recent rates appear to have equilibrated [ ]. As anticipated, rates of intravenous drug use (IVDU)-associated IE have increased in parallel which has shifted the epidemiology of IE [ , ]. Patients with IVDU-associated IE are more likely to have right-sided IE, often have more virulent organisms, and may have complicated long-term clinical courses due to continued drug use [ ]. Practices including needle sharing, licking of needles, lack of skin cleaning prior to injection, and contaminants in the drug mixture may all contribute to bacteremia and increase the risk of IE. Thus, efforts such as needle exchange programs and treatment with opioid agonists, both of which aim to curb the spread of communicable diseases, may potentially have beneficial effects in reducing the risk of IE [ ].

Nevertheless, it is difficult to project future trends in IVDU and thus its impact on the future of IE remains unclear. What is certain, however, is that this public health crisis is not disappearing. Use of synthetic opioids such as fentanyl has risen precipitously since 2013 and furthermore, patients who abuse prescription opioids, such as oxycodone, are 19 times more likely to use heroin [ , ]. In the United States, a multifaceted approach to address the crisis has been outlined which targets addiction and treatment services as well as research to better quantify the burden of the disease and to develop innovative strategies for addiction treatment and opioid-free pain management. It remains to be seen what impact these public health measures will have on persons currently using intravenous drugs as well as on the distribution of prescription opioids, which is a critical factor given the potential for progression to IVDU. Both of these aspects will determine the future profile of IE patients.

These trends also underscore the valuable role of Addiction Medicine in the long-term management of patients with IVDU-associated IE. This patient population displays high rates of readmission and recurrent IE attributable to continued IVDU and furthermore, the predominant cause of death among patients with IVDU-associated IE is recidivism [ , ]. Additionally, reoperative valve surgery is associated with a significant increase in 30-day mortality when compared to patients undergoing an initial surgery for IE [ ]. At present, attempts to curb continued IVDU have fallen short yet this represents a critical component of long-term IE management that also stands to lessen healthcare costs [ ]. Involvement of Addiction Medicine specialists in the care of these patients has been associated with reduced readmissions [ ]. A multidisciplinary approach will be necessary to ensure abstinence from IVDU and thereby address the risks of recurrent IE in this patient population. This will need to be integrated into outpatient follow-up for patients with IVDU-associated IE following the initial hospitalization.

Risk stratification

IE occurs in only a minority of patients with a prosthetic valve or with predisposing conditions, thus prediction of patient populations at highest risk of IE is critical for determining prophylactic therapy and directing work-up. Various models have been proposed to not only estimate the risk of IE but also to predict the ultimate need for surgery as well as long-term outcomes following medical or surgical management. Leveraging large cohorts, these tools offer the promise of guiding patient expectations as well as guiding optimal management strategies.

First and foremost, many studies have focused on the risk of IE in patients with prosthetic valves undergoing high-risk procedures in order to guide antibiotic prophylaxis. The most current guidelines for the prevention of IE from the American Heart Association (AHA) date back to 2007 and specify both high-risk patient populations and prophylaxis recommendations for based on various procedure types [ ]. The authors acknowledged the difficulty in ascertaining well-supported data to guide their recommendations, given the low overall incidence of IE, and called for randomized trials comparing antibiotics to placebo to conclusively define the relative risks and benefits in various groups [ ]. Though dental procedures may incite a transient bacteremia, the association between this and the development of IE is still unclear and the overall lack of evidence has led to disparate guidelines with the AHA recommending antibiotics to only high-risk groups whereas in the United Kingdom no antibiotic prophylaxis is recommended [ ]. Large population studies with longitudinal follow-up and detailed data regarding the type of invasive procedure as well as the quantity and timing of any prophylactic antibiotic administration will be necessary. From a cost-effectiveness perspective, one study has suggested that antibiotic prophylaxis may be beneficial in high-risk patients; thus, cost effectiveness along with clinical outcomes should be incorporated into future published guidelines [ ].

Studies have also focused on estimation of postsurgical mortality in order to predict outcomes following repair. Contributors to mortality among patients with IE include older age, prosthetic valve-associated infections, the causative organism, and systemic manifestations [ ]. The EndoSCORE, for example, was developed specifically to predict postoperative mortality and factors in age, sex, comorbidities, the presence of a perivalvular abscess, preoperative cardiac and renal function, and the infectious species [ ]. The majority of these studies are derived from retrospective review of existing databases, many of which do not contain granular data specific to IE treatment, including the causative microorganism and exact details of more complex operations, which may be required in cases of perivalvular involvement. This is one important limitation of the use of databases such as the Society of Thoracic Surgeons (STS) database. More complex analyses with the incorporation of prospective data collected for the purposes of risk stratification for IE would help to refine estimation of patient outcomes under different treatment scenarios. Additionally, application of validated risk scores to the prospective risk stratification of patients presenting with IE and multiple risk factors for poor outcomes may provide objective guidance to suggest how to best manage these patients.

Device development to prevent infective endocarditis

Prosthetic valves contribute significantly to the burden of IE and therefore the infectious risks associated with the foreign body have increased interest in device engineering to develop prostheses with a lower risk of infection. Myriad investigations have attempted to thwart bacterial colonization and biofilm formation through device design, particularly for prosthetic valves. In vitro studies examining bacterial counts on valve prostheses following antibiotic coating demonstrated decreased bacterial counts, suggesting possible clinical utility in prevention of early IE [ , ]. This, however, has not yet translated to clinical use. Perhaps most notable clinical study examining a prosthetic device designed to reduce the risk of infection was the Artificial Valve Endocarditis Reduction Trial, a randomized trial aimed to determine whether use of a silver-impregnated sewing cuff (Silzone, St. Jude Medical, St. Paul, Minnesota) for a mechanical valve implanted in the aortic or mitral position reduced rates of prosthetic valve endocarditis as compared to a conventional sewing cuff [ ]. The trial was initiated in 1998 and suspended in 2000 due to concerns about increased rates of paravalvular leak requiring reoperation in patients receiving the Silzone valve [ , ]. Incorporation of new materials, including decellularized tissue, into prosthetic valve design may potentially thwart biofilm formation, though it is too early to determine how these valves will perform clinically, particularly with respect to long-term durability [ ].

Though much of the work on prevention of device-related IE has focused on prosthetic valves, cardiac implantable electronic devices including pacemakers and defibrillators also carry a small but not insignificant risk of IE and have received targeted investigation. With the growth in the population of elderly patients, the prevalence of these devices stands to increase in the coming years [ ]. In the Worldwide Randomized Antibiotic Envelope Infection Prevention trial exploring the benefits of an absorbable, antibiotic-eluting envelope for implantation of cardiac pacemakers and/or defibrillators, use of the envelope was associated with a significant reduction in major infections including wound complications and deep surgical site infections [ ]. However, when considering the rates of bacteremia and endocarditis, which were <1% in both the antibiotic envelope and control groups, there was no reduction observed with the use of the envelope [ ]. A subsequent meta-analysis supports the relative benefit of the use of antibiotic envelopes for the reduction of local complications, though effects on the rates of IE were not specifically examined in this study [ ]. Prevention of endocarditis may thus be more amenable to technologic advances in engineering of the device leads to enhance resistance to biofilm formation rather than addressing the risk of infection in the pocket itself.

Registry development

Given the complexity and nuances in management of patients with endocarditis, it is necessary to gather a large quantity of data on a heterogeneous cohort of patients in order to establish best practices. The European Infective Endocarditis Registry is a prospectively collective, multicenter cohort of adult patients with possible or confirmed IE undergoing medical or surgical management [ , ]. This collaborative provides a broad overview of the etiologies, management approaches, and outcomes among patients with IE and reflects current practice patterns [ ]. A Swedish registry and Italian registry of patients with IE, along with registries established in other countries, have also been implemented and queried to obtain granular data regarding IE management and outcomes [ , ]. In particular, the Italian registry began enrolling patients in 1979 which allows for a longitudinal assessment of key components of IE, such as imaging findings and culture results, which can uncover trends in the clinical approach to IE as well as shifts in the etiology, such as the increase in IVDU.

The majority of registry-based IE data comes from Europe despite the wealth of information collected in the STS) database. The STS database captures many patients undergoing surgery for IE, however, it is limited in that it only represents patients who ultimately underwent surgical management and therefore may miss those patients who were successfully managed medically or those in whom surgery was indicated but was not undertaken. Additionally, it may fail to account for nonindex cases or to adequately account for cases in which surgical repair is complex. Though admittedly a large undertaking, a prospective study of both medically and surgically managed patients IE within the United States stands to contribute significantly to clinical management, particularly with the increases in IVDU and evolving etiologies of IE.

Utilizing big data

Both prospectively collected and retrospective databases on patients with IE highlight the need for new approaches to data analysis and interpretation to inform clinical management. Though datasets, including the STS, contain a large quantity of variables and encompass a sizable population, the nuances of IE pathophysiology, diagnosis, and management demand not only a large population but also a great level of detail in order to elucidate the complexities of patient care. Thus, more advanced statistical and computational approaches may allow for improved interpretation and application of these data to inform clinical practice.

Machine learning leverages computational power to build upon what is known about a particular subject and apply this information prospectively. This approach overcomes some of the limitations of traditional statistical methodologies and may take into account previously unidentified factors which improve performance of a model. Machine learning has recently been applied to IE to explore biomarkers in order to stratify the risk of mortality. Using classification trees, the authors found that elevated C-reactive protein, C–C motif chemokine ligand, and interleukin-15 were correlated with in-hospital mortality [ ]. Further exploration of these serum markers, combined with known factors contributing to IE outcomes, will likely improve prognostication for both medical and surgical management.

Machine learning can also be applied to the interpretation of imaging data. A prior study utilized machine learning to differentiate between restrictive and constrictive pericarditis as visualized by routine echocardiography imaging [ ]. Others have utilized machine learning for discrimination between ventricular tissue properties on echocardiography [ ] as well as prediction of mortality utilizing metrics other than ejection fraction alone [ ]. It is plausible that a similar approach to characterize mitral valve lesions by transthoracic echocardiography, for example, may enhance our diagnostic capabilities and reduce the need for transesophageal echocardiography (TEE) in these cases. Similarly, machine learning may represent an avenue by which we can improve prediction of the complexity and extent of surgical repair using current imaging modalities and thereby refine estimation of postoperative outcomes.

Few other studies have specifically utilized machine learning for IE, however, this approach may enhance the ability to provide individualized patient management and set expectations based on unique presentations and pathophysiology. Furthermore, application of these artificial intelligence modalities may guide the decision for operative intervention in some patients and potentially overcome some limitations of currently available imaging and biochemical assays. Incorporation of machine learning into clinical practice, however, will require validation in large datasets and may only be applicable in certain patient populations. This is also an iterative process that will require reassessment with each new clinical advance, such as the introduction of new imaging methods, though the ability to leverage all available data has already shown promise in some studies in improving prediction accuracy as compared to existing statistical methods [ ].

Future of diagnosis and management

Clinical diagnosis of IE relies on evaluation of the modified Duke criteria which incorporates imaging, laboratory, physical exam, and patient history components. Despite this, the diagnosis of IE remains elusive in some patients and in many relies on invasive imaging, such as TEE, to confirm the diagnosis, evaluate for more complex disease, or monitor improvement. Thus, investigations have focused on enhancing the ability to diagnosis or better characterize IE. In addition to improvements in diagnosis, treatment advances in medical and surgical care together will contribute to reducing the mortality and impact of the disease on the healthcare system.

As a component of the clinical diagnosis utilizing the modified Duke criteria, imaging adjuncts are frequently utilized to examine the valve for vegetations, assess the hemodynamic significance of endocarditis lesions, and monitor for resolution of infection. In many cases, accurate assessment of valvular structures, in particular the mitral valve, has relied on TEE, thus recent studies have focused on noninvasive radiologic evaluation of disease burden. As an example, ¹⁸ F-fluorodeoxyglucose positron emission tomography (PET)/computed tomography (CT) was recently explored in a meta-analysis reviewing its diagnostic accuracy in both native valve and prosthetic valve endocarditis [ ]. PET/CT was found to improve the accuracy of diagnosis in patients with a cardiac implantable electronic device or a prosthetic valve when compared to the use of the modified Duke’s criteria alone [ ]. This imaging technique is not universally available but its increasing application in other disease processes may increase the availability for the diagnosis of IE.

A variety of other techniques have been applied to enhance the diagnosis of IE, particularly in the identification of the causative microorganism in cases of culture-negative endocarditis. This is a critical component in order to guide antimicrobial therapy. Routine blood cultures are the standard approach to microorganism identification, however, prior antibiotic administration may interfere with cultures and some organisms may not proliferate in routine culture leading to negative results. Utilization of polymerase chain reaction for identification of organisms from excised cardiac tissue holds promise in improving microorganism diagnosis [ ]. The use of a fluorescence in situ hybridization technique may also advance diagnostic capabilities with the added benefit of being able to differentiate between active biofilms and degraded bacteria [ ]. Observation of ongoing bacterial proliferation despite antimicrobial therapy may also help to guide the choice and duration of antimicrobial agent.

The administration of intravenous antibiotics, often for a six-week course, has been a critical component of IE treatment although recent work has suggested that enteral administration may be sufficient. A recent review of oral step-down therapy found that oral antibiotics with high bioavailability may be noninferior to intravenous antimicrobials when instituted following the resolution of bacteremia [ ]. Though this concept would benefit from large randomized controlled trials and may only applicable to certain causative organisms and in patients with stable disease, the potential to avoid the complications of long-term intravenous access and the associated costs may decrease the overall burden of IE on the healthcare system. This may also have a significant impact in patients with a history of IVDU who often require discharge to a nursing facility for a prolonged course of parenteral antibiotics. Newer antimicrobials also hold promise in lessening the burden of treatment while maintaining efficacy. The use of oritavancin, a newer intravenous antibiotic with activity against gram-positive organisms and a half-life of over two weeks, has been described for the treatment of Vancomycin-resistant Enterococcus prosthetic valve endocarditis using weekly dosing [ ]. This may represent another feasible approach to outpatient antibiotic treatment for IE patients.

In addition to advances in diagnosis and medical therapy, the surgical approach to IE has evolved, particularly as outcomes for other open heart surgeries have shown improvement over time. The increasing prevalence of IVDU-associated IE, which is commonly associated with more aggressive microorganisms, may lead an increased prevalence of paravalvular abscesses and fistulas, thereby increasing the complexity of surgery. These IVDU patients, however, are typically younger and have fewer comorbidities, making it reasonable to consider more extensive operations. With the trends in IVDU, these more complex surgeries may become commonplace for IE operations and with increasing experience, particularly at high volume centers, the mortality and postoperative outcomes for these operations will undoubtedly improve. Increasing experience with minimally invasive techniques may also make this a feasible approach in selected cases of IE [ , ]. Furthermore, percutaneous removal of vegetations through vacuum aspiration system has been studied and may represent an important therapeutic adjunct, particularly in patients who are unfit for surgery [ ]. The timing of surgery is also a topic of debate that could benefit from further refinement, particularly in cases with extracardiac involvement [ , ]. The difficulty in obtaining prospective data and the heterogeneous nature of the patient population limits the clinical translation of many of these studies. Regionalization of IE care at high volume centers may assist in research endeavors focused on surgical indications and conduct as well as improving patient care.

Preliminary work has been undertaken to target the systemic inflammatory response at the time of surgery for IE. The REMOVE study is a randomized controlled trial which will evaluate the efficacy of CytoSorb (CytoSorbents, Monmouth Junction, New Jersey), a nonselective cytokine-adsorbing column placed in the cardiopulmonary bypass circuit, on multiorgan dysfunction following surgery for IE [ ]. Though patients with IE have been found to have higher levels of circulating cytokines as compared to patients undergoing surgery for noninfectious valvular disease, a prior pilot study showed no difference between CytoSorb and no hemadsorption following cardiac surgery [ , ]. It remains to be seen what the impact of attempts at intraoperative immunomodulation will have both on objective levels of inflammatory markers as well as clinical outcomes.

Multidisciplinary approaches

Clinical successes with the use of a Heart Team for decision-making in complex patients with indications for cardiac surgery have spurred interest in the establishment of a dedicated multidisciplinary team to manage patients with IE. This is particularly critical in patients with IVDU-associated IE given the known high risk of recidivism. As highlighted in the European Society of Cardiology treatment of IE guidelines, the Endocarditis Team addresses the nuances in the pathophysiology of IE through a collaboration between providers from multiple specialties in order to optimize outcomes among these patients [ ]. The AHA also recommends a team-based approach to IE in their guidelines for valvular heart disease, particularly in the discussion of surgical intervention [ ]. The impact of this team spans all phases of the disease process including diagnosis, treatment approach, follow-up, and long-term outcomes. In addition to cardiologists, infectious disease specialists, and cardiac surgeons, providers in the field of neurology, neurosurgery, and rheumatology provide valuable input for optimized management of these complex patients [ ]. Patients with IVDU-associated IE would also benefit from early involvement of social workers, psychiatrists, and addiction medicine specialists to ensure abstinence thereby limiting repeated IE events.

Implementation of such a team has been well-received by participants and may help to advance both quality improvement and research initiatives [ , ]. Furthermore, in centers in which the rates of IE are relatively low, a multidisciplinary approach or even consultation with a team at a higher volume tertiary care center may promote adherence to societal guidelines and ensure evidence-based management. Regular meetings with this group of stakeholders help not only to triage and establish management goals for the initial diagnosis but also to coordinate with outpatient providers, to discuss cessation of antimicrobial therapy, and to determine long-term plans for follow-up, treatment, or delayed surgery where indicated. The successes and universal recognition of the value of a Heart Team approach may facilitate the establishment of a dedicated Endocarditis Team in many institutions in which the infrastructure for multidisciplinary management has already been established. Though not a new concept, the Endocarditis Team may be the critical component to comprehensively address and integrate all future developments in the care of patients with IE.

Conclusion

IE has presented diagnostic and management challenges since its first recognition centuries ago. Much of the improvement in the care of patients with this disease has relied on parallel advances in medicine in the realms of laboratory testing, imaging modalities, and surgical approaches. It is increasingly clear, however, that the complexity and heterogeneity of IE will require focused studies to lessen the significant mortality associated with its pathophysiology. Improved understanding of high-risk patients will serve to guide strategies for prevention. The ability to interpret and apply a wealth of data has the potential to compensate for limitations in imaging techniques. One key aspect of IE that will play a role in the future of the disease is the burden of IVDU-associated IE which has drastically altered the epidemiology of the disease and has implications in the treatment and long-term outcomes of IE patients. In this respect, clinical experience and the involvement of a multidisciplinary team hold the most promise in combatting this disease. Advances in all of these arenas will together serve to make a significant impact on the current high morbidity and mortality of IE.