Risk of infective endocarditis after left-sided surgical valve replacement

This editorial refers to ‘Incidence and factors associated with infective endocarditis in patients undergoing left-sided heart valve replacement’†, by L. Østergaard et al., on page 2668. Infective endocarditis is a complex and challenging disease, in which complications, recurrence, and mortality are common.1 The incidence in the developed world is rising and is intrinsically linked to healthcare provision for an ageing and susceptible population.2 Since William Osler’s eloquent description of infective endocarditis in adults with rheumatic heart disease, it has been clear that specific groups of patients are at increased risk.3 While the incidence of infective endocarditis in the overall population is low (∼35 per million per year), it is >100 times higher in patients at highest risk, i.e. those with a prosthetic valve replacement, previous infective endocarditis, and some types of congenital heart disease.4 Identification and stratification of at-risk groups is important. First, it allows doctors to lower their threshold of suspicion in specific patients, which is helpful given the variable and non-specific clinical presentation. Secondly, it facilitates education and empowerment of at-risk patients in their diagnosis and management. Thirdly, it allows directed use of strategies for disease prevention (which may have risks or cost implications) to groups at highest risk. Both European Society of Cardiology (ESC) and American Heart Association/American College of Cardiology (AHA/ACC) guidelines advise risk stratification to target the use of oral antibiotic prophylaxis before dental extraction to the highest risk groups, given the uncertainty and lack of evidence concerning efficacy.5–7 Finally, understanding specific risk factors may inform both clinical and translational research, and lead to future strategies for disease prevention. Patients with a prosthetic valve replacement are the largest high-risk group for infective endocarditis and, given the burden of valvular heart disease, this number is set to increase significantly in future years.8 Prosthetic valve endocarditis (PVE) is a life-threatening complication which is challenging to diagnose and treat, and is characterized by higher rates of indeterminate echocardiographic findings, staphylococcal infection, abscess formation and valve dehiscence, complex surgical intervention, and disease recurrence.9 , 10 The incidence and specific risk factors for PVE are not well defined in contemporary series. In this context, the study by Ostergaard et al. in this issue of the European Heart Journal is informative.11 Using the Danish National Patient Registry, the authors identified a cohort of 18 041 patients who underwent left-sided surgical prosthetic valve replacement from 1996 to 2015 [16 018 patients with an aortic valve replacement (AVR), 1751 patients with a mitral valve replacement (MVR), and 272 who underwent double valve replacement]. Patients were followed from the time of surgery for up to 12 years or until the end of the study, with a primary endpoint of rehospitalization due to PVE. Cox proportional hazards analysis was used to identify baseline factors predictive of PVE, and propensity scoring was performed to generate matched cohorts of biological and mechanical valve recipients for comparison. The headline finding from this study is that the cumulative incidence of PVE was 5.2% at 10 years, with an overall incidence of PVE of 69.8/10 000 person-years. The incidence was equivalent in patients following AVR and MVR. In patients with an AVR, specific clinical factors associated with an increased risk of PVE were male gender [hazard ratio (HR) 1.59, 95% confidence interval (CI) 1.33–1.89], presence of a cardiac implantable electronic device (HR 1.57, 95% CI 1.19–2.06), diabetes (HR 1.45, 95% CI 1.15–1.84), atrial fibrillation (HR 1.34, 95% CI 1.13–1.59), and cancer (HR 1.28, 95% CI 1.01–1.62). In patients with an MVR, male gender (HR 1.68, 95% CI 1.06–2.68) and heart failure (HR 1.69, 95% 1.06–2.68) were associated with an increased risk of PVE. In multivariate and propensity-matched analyses, there was a significantly higher risk of PVE in patients receiving a bioprosthetic AVR, in comparison with those with a mechanical valve (HR from matched cohort 1.94, 95% CI 1.37–2.75). In patients receiving an MVR, a bioprosthetic valve was also a risk factor for PVE (HR from multivariate analysis 1.91, 95% CI 1.08–3.37), but this could not be confirmed in the small matched cohort due to no events in the mechanical MVR group. The strengths of this study are that it provides an accurate estimate of PVE incidence in a large, contemporary, surgical prosthetic valve population over a prolonged period with limited loss to follow-up. The incidence identified is similar to that recently reported in large nationwide data sets from the UK (0.46% per person-year)4, Sweden (0.57% per person-year),12 and the authors’ previous work on the overall incidence of PVE in Denmark (0.6% per person-year).13 There are, however, some limitations. The diagnosis of PVE was reliant on ICD-10 (International Classiification of Diseases 10th Revision) coding, and data on the valve affected and underlying microbiology were not available. The finding that bioprostheses are associated with a higher risk of PVE is subject to major confounding by indication. Biological prostheses are not only used in those who are older and co-morbid, but also in those who are more frail and clinically unsuitable for anticoagulation. Despite rigorous propensity matching, this is difficult to adjust for and there is a likelihood of residual confounding. Finally, there was no differentiation between early and late PVE, which might have revealed specific risk factors in the post-operative phase compared with late follow-up. The fact that 1 in 20 patients will experience PVE at 10 years despite current efforts at disease prevention carries an important message for patients and clinicians. Patient knowledge of infective endocarditis is generally poor, and a specific educational programme after valve replacement with yearly recall, follow-up Emails, and dental review might improve awareness, oral hygiene, and earlier diagnosis.14 For clinicians, a particular focus on the group of elderly male patients with bioprosthetic valves is warranted, as they appear to be at particularly elevated risk, although the mechanism remains unclear.12 , 15 Finally, this study elegantly demonstrates the importance of large, national or international data sets to yield insights into rare and challenging diseases such as PVE. In this context, the results of the ongoing EuroENDO registry, co-ordinated by the ESC EURObservational Research Programme and enrolling >2000 patients with native valve endocarditis and PVE across the world, are keenly awaited.16 Conflict of interest: B.D.P. has received unrestricted research grants from Edwards Lifesciences and lecture fees from Edwards Lifesciences, Symetis, and Boston Scientific. T.J.C. has no conflicts to declare.