Cardiovascular conditions predisposing to infective endocarditis: time to reconsider the current risk classification system?

This editorial refers to ‘Quantifying infective endocarditis risk in patients with predisposing cardiac conditions’†, by M.H. Thornhill et al., on page 586. Infective endocarditis (IE) remains a devastating disease despite continued advances in diagnosis, anti-microbial agents, and surgical therapy.1 Treatment typically involves a several week course of parenteral antibiotics, and nearly half of cases will require cardiac surgery during the acute to subacute phase.1 The epidemiology of IE has changed significantly in recent years, a trend characterized by a decrease in the prevalence of rheumatic heart disease, the advancing age of the population, patient co-morbidities, increases in the use of prosthetic heart valve substitutes and cardiac implanted electronic devices (CIEDs), and the worldwide surge in injection drug use.1–3 Staphylococcal species are now the most common bacterial pathogens, and many instances of IE occur in the absence of antecedent valvular heart disease.1–3 Patient outcomes are highly variable, depending on prompt recognition, early complications, and the timely application of surgery when indicated. Development of effective strategies to prevent IE has been challenged by its relatively low incidence at the population level and the increasing heterogeneity of predisposing conditions that confer risk.2 Recent guideline recommendations for the use of prophylactic antibiotics to prevent IE have been focused on their delivery to at-risk patients prior to dental/oral surgical procedures associated with transient bacteraemia.4–6 As a prevention strategy, they are thus quite narrow in scope and not without controversy. Inability to prove causality, concerns regarding the emergence of antibiotic resistance, and wariness about cost-effectiveness led professional societies and health organizations to re-examine this practice. It has been >10 years since the American Heart Association (AHA), followed in turn by the National Institute for Health and Care Excellence (NICE) and the European Society of Cardiology (ESC), issued restricted recommendations for antibiotic prophylaxis.4–6 Two of the three organizations (AHA and ESC) limited antibiotic prophylaxis to ‘high-risk’ patient subsets [e.g. those with prior IE, prosthetic heart valves, or cyanotic congenital heart disease (CHD)], whereas the third (NICE) recommended against the routine use of antibiotics across all patient risk categories.4–6 The results of epidemiological studies examining the effects of a reduction in antibiotic prescribing for dental prophylaxis on the incidence of IE have been mixed.7–9 It is reasonable to question whether IE risk in the current era can be reliably assigned using outdated historical benchmarks.10 Accurate risk assessment is essential not only for targeting prevention strategies but also for ensuring best practices and informing the design of future clinical trials. In this issue of the journal, Thornhill et al.11 provide a contemporary assessment of the relative risks of incident IE from all causes and hospital-related mortality among patients with cardiovascular disease identified as ‘high’ or ‘moderate’ risk by AHA and ESC criteria (see Supplementary table S1 in Thornhill et al. in for definitions) during admission to all English hospitals between January 2000 and March 2008.11 Individuals were followed for a minimum of 5 years through March 2013. Data were extracted from the Hospital Episode Statistics (HES) registry. The registry, using National Health Service unique patient numbers, captured cases of IE admitted to English hospitals reliably and accurately. The English population in 2008 served as the referent cohort. The study confirmed that the 5-year relative risks of incident IE and in-hospital death following admission with IE are very high among select patient cohorts traditionally considered ‘high risk’, namely those with a previous history of IE or prosthetic or repaired heart valves, CHD, or CHD with palliative shunts/conduits. Five-year risks of IE and related mortality in other patient subgroups, however, varied somewhat from what was expected. For example, patients with CHD repaired with prosthetic material (traditionally considered a ‘high-risk’ group for the first six post-operative months) had lower risks of incident IE or IE-related mortality than other patients considered at only ‘moderate risk’. Surprisingly, several ‘moderate-risk’ patient subgroups, including those with rheumatic fever, non-rheumatic valve disease, or congenital valve anomalies, had risk estimates comparable with those for many ‘high-risk’ conditions. Patients with CIEDs or hypertrophic cardiomyopathy were also at increased risk. The markedly elevated odds ratios reported (ranging up to 266 with a 95% confidence interval of 244–288 for the development of IE among patients with a history of previous IE) have to be placed in the context of low absolute event rates. Other acknowledged study limitations include the reliance on administrative claims data, the lack of microbiological and echocardiographic data, the absence of information regarding antibiotic course or surgical treatment, and the inability to account for patient co-morbidities, such as diabetes mellitus or end-stage renal disease. Factors predisposing to infective endocarditis (IE). Risk assessment for the development of IE should take into account patient-related (both cardiac and non-cardiac), organism-related, and health system-related factors. As suggested by Thornhill et al.,11 patient cardiac-related risk stratification algorithms may need to be revised based on their analysis of contemporary data from patients admitted to English hospitals in the period from 2000 to 2008. Examples of patient co-morbidities include immunosuppression, diabetes mellitus, and end-stage renal disease. High-risk predisposing patient cardiac factors would include a history of previous IE, prosthetic heart valve replacement, and cyanotic congenital heart disease. Thornhill and colleagues have identified other patient subsets (e.g. acquired heart valve disease, cardiac implanted electronic devices) that may need to be viewed as high risk (see text).11 When IE risk is viewed from a more global perspective, it is clear that antibiotic prophylaxis prior to dental procedures targets a limited segment of the at-risk population. The figure is not meant to be inclusive of all factors contributing to risk. CA, community acquired; HCA, healthcare associated. Although the authors have appropriately avoided interpreting their data through the narrow prism of antibiotic prophylaxis for dental procedures, they propose that we may need to re-examine traditional risk stratification algorithms for the development of IE and associated hospital mortality, thereby to broaden approaches to prevention and to facilitate more timely diagnosis and intervention. The clinical and financial implications of reclassifying risk from ‘moderate’ to ‘high’ for many patients, including those with certain types of congenital or acquired valvular heart disease (VHD) or CIEDs, for example, are potentially enormous. In this regard, it would be of interest to learn whether IE risks vary across subsets of patients with different types of VHD, such as bicuspid aortic valve disease, mitral valve prolapse, and calcific aortic stenosis. There remains a great deal of work to be done to reduce the expanding reach of IE and its many complications. Antibiotic prophylaxis, when indicated, and a continued emphasis on dental and cutaneous hygiene remain the cornerstones of prevention,6 yet their benefits can extend only so far and to a limited segment of the at-risk population. The development of more contemporary risk stratification tools would be a welcome addition to the continuous quality improvement efforts needed to design novel prevention strategies and optimize patient management. Thornhill and colleagues herein provide a fresh look at the evolving epidemiology of IE.11 In the process, they remind us of the benefits of asking critical questions, using the data at hand to revisit past assumptions, and raising the notion that in the current era perhaps we understand less about IE than we thought we did. Conflict of interest: none declared.