By Michael H. Crawford, MD

Professor of Medicine, Associate Chief for Education, Division of Cardiology, University of California, San Francisco

Dr. Crawford reports no financial relationships relevant to this field of study.

SYNOPSIS: Interrogation of the Danish National Patient Registry revealed bacteremia due to Enterococcus faecalis was most likely to be associated with infective endocarditis; thus, echocardiography is warranted in these patients.

SOURCE: Østergaard L, et al. Prevalence of infective endocarditis in patients with positive blood cultures: A Danish nationwide study. Eur Heart J 2019;40:3237-3244.

To decide which patients with bacteremia need an echocardiogram, knowledge of the risk of infective endocarditis (IE) with various blood stream infections is needed. Danish researchers interrogated the Danish National Patient Registry for patients with bacteremia typically associated with IE (Enterococcus faecalis, Staphylococcus aureusStreptococcus spp., and coagulase-negative staphylococci [CoNS]) from 2010 to 2017.

The study outcome was a diagnosis of IE and a hospitalization of at least 14 days (unless the patients died earlier). The 69,021 patients identified were collected into four groups of two contiguous years. The highest prevalence of IE was in patients with E. faecalis (17%), followed by S. aureus (10%), Streptococcus spp. (7%), and CoNS (2%). The prevalence of IE in E. faecalis patients significantly increased over time (12% in 2011 vs. 19% in 2015; P = 0.0005) and in those with Streptococcus spp. (6% in 2010 vs. 8% in 2017; P = 0.03). Overall, the rates of IE were higher in men with E. faecalis, Streptococcus spp., and CoNS (P < 0.0001), but not for S. aureus. Also, all but S. aureus showed a higher prevalence of IE with advancing age (P < 0.0001). The authors concluded that the overall prevalence of IE was one in six for E. faecalis bacteremia, one in 10 for S. aureus, and one in 14 for Streptococcus spp. These results suggest echocardiographic screening for bacteremia caused by these three organisms is clinically warranted.


Considering the high in-hospital mortality of IE (about 20%), early identification of patients at high risk for IE is desirable. The four bacteria species investigated in this study account for 75-85% of cases of IE in reported series. Thus, assessing the prevalence of IE in patients with bacteremia from these organisms makes sense. Interestingly, all four are gram-positive bacteria, which are known to be superior at adhering to the endothelium.

The most surprising result of the study was the higher IE rate for E. faecalis than S. aureus (17% vs. 10%). However, the study also showed an increase in E. faecalis IE with age, which could be attributed to colon cancer and other diseases increasing the prevalence of E. faecalis bacteremia. The higher overall prevalence of E. faecalis IE probably is due in part to the aging of the population. At age 70-80 years, the E. faecalis IE rate was 20% vs. 12% for S. aureus. At age 40-50 years, the authors observed a rate of 13% for both. E. faecalis IE also is much more prevalent in men for reasons that are poorly understood, but may be due more to underlying epidemiologic characteristics than biologic ones.

Current major organizational guidelines recommend consideration of echocardiography, especially transesophageal echocardiography (TEE) for S. aureus bacteremia (class IIa). The results of this study suggest this recommendation should be extended to E. faecalis. However, the systematic application of echocardiography, especially TEE, for a disease with a ≤ 20% prevalence in the at-risk population may not be feasible or cost-effective. In most series, the use of echo is about 50-65%. Many have suggested using a risk score such as NOVA, PREDICT, VIRSTA, or AANDOC to cull the highest-risk bacteremia patients for echoes. The Duke score is not recommended because studies have shown that it is largely driven by the echo results. These scores are highly sensitive and carry a negative predictive value of > 95% (but specificity is lower). This may be acceptable for such a high mortality disease.

There were several limitations to the Østergaard et al study. First, as it was an administrative database study, there was limited clinical information, such as echo results. Second, the authors used ICD-10 codes to diagnose IE. Prior validation studies revealed this approach carried a positive predictive value of 90%. Third, the increase in the prevalence of IE over time may have been because of an increased use of echo and nuclear imaging. Fourth, echoes were not performed systemically; the estimates of IE rates may be conservative. Finally, the differences in the incidence rates for various organisms may vary geographically; these results may not reflect all areas in the world. I believe the main message of this paper is the increased prevalence of E. faecalis IE and the corresponding need to consider echoes earlier in the course of E. faecalis bacteremia.