Infective endocarditis: an ongoing global challenge

This editorial refers to ‘Clinical presentation, aetiology, and outcome of infectious endocarditis. Results of the ESC-EORP EURO-ENDO (European infectious endocarditis) registry: a prospective cohort study’†, by G. Habib et al., on page 3222. Infective endocarditis (IE) is a life-threatening disease that consumes considerable healthcare resources and requires multidisciplinary management (‘endocarditis teams’). While still low, the incidence of IE has increased in recent years. In 2014, it was 3.49 episodes per 100 000 person-years1 (∼1600 cases) in Spain. In a single centre, such as the Hospital Clinic, Barcelona, this figure was reached after 40 years of recording cases of endocarditis. Therefore, multicentre and multinational cohort studies are necessary in order to know the epidemiological changes and clinical characteristics of IE, new diagnostic tools, and prognosis. Such studies can collect many cases over a very short period. Given that the number of endocarditis episodes diagnosed in Europe and worldwide in 1 year could be as high as 25 000 and 250 000 cases, respectively, these data can help us to improve patient outcome. IE is truly a global challenge! In this issue of the European Heart Journal, Habib et al.2 report initial data on the clinical presentation, aetiology, and outcome of 3116 IE episodes in Europe and elsewhere (156 hospitals, 40 countries) collected between 2016 and 2018 in the EURO-ENDO registry. The authors provide an updated overview of IE and a comparison of national and international endocarditis registries in the 21st century. Their findings highlight changes in IE in recent decades.3 As a multinational registry, EURO-ENDO can be compared with the Euro Heart Survey4 or the International Collaboration on Endocarditis (ICE) cohort.5 However, the first was a substudy of endocarditis from a valvular heart disease survey with only 159 cases performed in 2001. We compared these data with the ICE registry, which, in 2009, reported the clinical characteristics and outcome of 2781 cases of definite endocarditis collected at 58 sites in 25 countries during the years 2000 and 2005 (Supplementary material, Tables S1–S3). The main findings of EURO-ENDO2 are not surprising and include the increasing age of IE patients (63 vs. 58 years in ICE), a higher percentage of cases of endocarditis due to prosthetic valves and cardiovascular implantable electronic devices (41% vs. 28% in ICE), the increasing frequency of healthcare-associated endocarditis (33% vs. 23% in ICE), the high percentages of staphylococci (44%) and enterococci (16%), the increasingly important diagnostic role of techniques other than echocardiography (e.g. [18F]FDG PET/CT), and the high rates of surgery (51% vs. 48% in ICE). The authors identified a series of predictors of cardiac surgery (vegetation size, abscess, heart failure, ischaemic stroke, and management in a European country increased the probability and comorbidities, and female gender decreased the probability) and of death (vegetation size, abscess, comorbidities, renal failure, heart failure, stroke, and no cardiac surgery when indicated). While these findings were not unexpected, they do merit more in-depth evaluation, since they help explain the paradox of improved management not accompanied by a significant reduction in in-hospital mortality (17% in EURO-ENDO and 18% in ICE; Figure 1). Types of infective endocarditis and main outcomes in ICE (2000–2005)5 and EURO-ENDO (2016–2018)2 registries. The authors have categorized the clinical centres as ‘high-volume’ and ‘low-volume’, with an arbitrary limit of 20 cases/year, with high-volume centres (79.5%) providing 92% of the cases included. The registry has two important limitations: data on patients transferred from a non-cardiac centre (42% of cases in the ICE registry) were not recorded, and community hospitals (non-cardiac and low-volume) were under-represented. Therefore, it is impossible to know the real patient age, comorbidities, place of acquisition, type of endocarditis (native, prosthetic, or device-related), microbiological and surgical data, and mortality rates in non-referral centres. A population-based study of cases from non-cardiac hospitals in high- and low-income countries is needed to better depict epidemiological characteristics and outcomes. Also noteworthy is the distribution of cases in the reporting countries. The percentage of patients from Europe in the ICE and EURO-ENDO cohorts was 44% and 70%, respectively, and most cases were provided by France, Spain, and Italy.3 Furthermore, as the number of episodes from some European and non-European countries are under-represented, the conclusions of the ICE or EURO-ENDO studies cannot be extrapolated. The same is true of the USA6 and some Eastern European countries7, where intravenous drug use is epidemic and co-infection by HIV and/or HCV can modify the clinical characteristics, management, and outcome of IE. Complications in both cohorts were similar, although the rates of stroke and congestive heart failure were higher in EURO-ENDO (Supplementary material, Table S3). Central nervous system embolic events were recorded in 20% of patients. There is much room for improvement in this area. Although intravenous thrombolysis has been associated with an increased bleeding risk and is contraindicated in IE embolic events, mechanical thrombectomy is promising8 and can reduce the morbidity of this complication. Diagnosis of IE has traditionally been based on echocardiography (almost all patients in both cohorts). Transoesophageal echocardiography was performed in the same proportion of cases (58% vs. 59% in the ICE registry), although vegetations were less frequent in EURO-ENDO (73% vs. 87%). The 2015 European Guidelines were pioneering in that they included [18F]FDG PET/CT and other cardiac techniques (multi-slice CT and leucocyte scintigraphy) as new diagnostic tools in prosthetic valve endocarditis.3 , 9 It is not surprising, therefore, that [18F]FDG PET/CT was performed in almost 17% of patients in EURO-ENDO (mostly in Western Europe), with positive results in 67% of prosthetic valve endocarditis, 28% of native valve endocarditis, and 16% of lead endocarditis. We agree that, while widely available in cardiac centres, the technique was underused for diagnosing IE in clinical practice. EURO-ENDO reports increased rates of staphylococcal and enterococcal endocarditis in tertiary teaching hospitals that perform cardiac surgery. The authors mention that the small percentage of oral streptococci and orally acquired disease is reassuring, after the recent restrictions on antimicrobial prophylaxis in national and international guidelines. However, data extracted from EURO-ENDO should be interpreted with caution. As many cases (33%) were healthcare related, an oral portal of entry is much more unprovable. Moreover, as previously explained, 92% of cases were from high-volume centres, where prosthetic valve- and device-related cases are over-represented (higher prevalence of coagulase-negative staphylococci). Indeed, some reports have raised the issue of increased streptococcal IE in conditions not currently managed with IE prophylaxis, such as mitral prolapse or bicuspid aortic valve.10 EURO-ENDO can only provide a partial answer to the question of whether the frequency of cases due to oral streptococci has decreased or not in a community-based endocarditis registry. On the other hand, in the 21st century, it is difficult to understand the high percentage of negative blood cultures in EURO-ENDO (21% vs. 10% in the ICE registry, Supplementary material, Table S2). This discrepancy warrants clarification. Another important recommendation of the 2015 European guidelines was the creation of ‘endocarditis teams’.2 It would be interesting to know whether the prognosis of endocarditis would have differed between centres if these teams had been in place when the guidelines were published. Unfortunately, this information was not collected, although many tertiary referral centres presumably have such teams, which have improved outcomes of endocarditis11 and should be encouraged in both ‘high-volume’ and ‘low-volume’ centres. Cardiac surgery has become increasingly relevant, particularly when performed early in the course of IE. Cardiac surgery was performed in 48% and 51% in the ICE and the EURO-ENDO registries, respectively. However, an important added value of EURO-ENDO was to record the indication for surgery in 69% of patients, i.e. a change of 18%; not undergoing surgery when indicated was an independent factor for mortality in the multivariable analysis. This is an indirect sign of the complexity of cases, with high surgical risk being the most common contraindication for surgery (almost 60% of cases). The 18% of patients with an indication for surgery but who were finally not operated on seems difficult to reduce in the near future. In fact, it may even grow, since IE patients are constantly more complex and older, with more comorbidities. This is a key area for improvement regarding in-hospital and 1-year survival, because in-hospital mortality has not changed in the last 20 years. The EURO-ENDO study2 clearly shows that the prognosis of endocarditis patients without indications for cardiac surgery and those with indications who undergo cardiac surgery is similar and very good. Antimicrobial therapy and relapse rates were not reported in this initial paper of the EURO-ENDO registry.2 In a field that has evolved more slowly than other infectious diseases such as HIV/AIDS, hepatitis, and fungal infections, old regimens (vancomycin and aminoglycosides), which are frequently not adapted to an ageing population, are still recommended. Even clinicians involved in drafting the guidelines frequently do not follow them,12 thus indicating the extent to which they are not aware of the current situation of the disease. More clinical trials of antimicrobials for IE are needed, as is rapid translation of their results to practice guidelines. Creativity and originality are also needed to generate newer antimicrobials, newer regimens, combinations of available antimicrobials, and improved routes of administration (home oral or parenteral therapies). It will be very interesting to see in future EURO-ENDO studies how clinicians are using antimicrobials for IE and how frequently 2015 European guidelines are followed. Finally, from a European perspective, we encourage the authors to perform a substudy in order to determine whether the clinical characteristics and outcomes differed between European regions and between European and non-European countries. As several European sites also participated in the ICE registry, an analysis combining both registries will show epidemiological and clinical trends in IE in Europe during the last 20 years. In conclusion, the EURO-ENDO registry provides extremely valuable information regarding the current characteristics of IE and enables identification of areas where improvements are necessary to reduce the still very high morbidity and mortality of this disease. Ancillary studies (including 1-year outcomes) would be more than welcome for professionals working in the field. J.M.M. received a personal 80:20 research grant from Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain, during 2017–2019.