By Michael Crawford, MD, Editor
SOURCES: Showler A, et al. Use of transthoracic echocardiography in the management of low-risk Staphylococcus aureus bacteremia: Results from a retrospective multicenter cohort study. JACC Cardiovasc Imaging 2015;8:924-931.
Kaasch AJ, Michels G. Staphylococcus aureus bloodstream infection: When is transthoracic echocardiography sufficient? JACC Cardiovasc Imaging 2015;8:932-933.
Bacteremia due to Staphylococcus aureus (Staph) can be associated with infective endocarditis (IE). Transthoracic echo (TTE) may be falsely negative early in IE, so transesophageal echo (TEE) is often recommended. However, the role of TTE in diagnosing IE in patients with Staph bacteremia is unclear. Thus, investigators from Toronto conducted a retrospective cohort study of 833 hospitalized patients with Staph bacteremia from seven hospitals over a 3-year period between 2007 and 2010. Echoes were performed at the discretion of their primary physician. The primary outcome was diagnosis of IE within 90 days of Staph bacteremia. The 536 patients who received a TTE within 28 days of bacteremia were randomly assigned to derivation and validation cohorts. Multivariate analysis was used to identify high-risk criteria for developing IE in the derivation group, which were then applied to the validation group. Bacteremia was community-acquired in 28%, healthcare-associated in 37%, and nosocomial in 34%. Within 28 days of the first positive blood culture, 54% of the total population had a TTE, 11% TTE and TEE, and 3% had a TEE alone. TTE was normal in 69%, met criteria for IE in 22%, and was indeterminate in 9%. Four clinical criteria predicted IE: indeterminate or positive TTE; community-acquired bacteremia; IV drug use; and the presence of a high-risk cardiac condition. The presence of any one of these criteria in the validation group had a sensitivity of 97% and a specificity of 52%. The negative predictive value was 99% and the positive 25%. The authors concluded that in patients without community-acquired Staph bacteremia, a high-risk cardiac condition, or IV drug use, a negative TTE excluded IE.
Staph bacteremia is frequent in hospitalized patients and often raises concerns for IE. All guidelines recommend echocardiographic imaging in these cases, but differ in their recommendation for employing TEE. The Infectious Diseases Society of America recommends TEE in all, but various American and European cardiac societies recommend the selected use of TEE. This study addresses this issue and demonstrates that clinical plus TTE data can identify a low-risk group that does not need TEE because the incidence of subsequent IE is < 1%. In their series, using these criteria would decrease TEE use by half. The sensitivity of their criteria for IE was 97% with a negative predictive value of 99%. Specificity and positive predictor values were lower (52 and 25%), but this is probably because of the strict criteria they used for a negative TTE. Not only did the echo have to lack any major Duke Criteria (oscillating mass, perivalvular leak, or abscess), it also lacked nonspecific abnormalities, such as valve thickening, new regurgitation, or non-mobile masses. Also, their clinical criteria excluded patients with any cardiac-foreign material, congenital heart disease, cardiac transplant, valve disease, or a history of IE. Additionally, 15% of the patients had more than one TTE performed. Consequently, none of the patients who would not have needed TEE developed IE in the 90-day follow-up period. Finally, the study is biased toward patients more likely to have IE, since 32% of the total population did not get an echocardiogram. These patients were more likely on a surgical service and less likely community-acquired or associated with IV drug use. So despite being a retrospective observational study, the results are robust and agree with the recommendation of cardiac societies to use TEE selectively in Staph bacteremia.
In this study with an IE prevalence of 14%, TEE would have been indicated in 55% of the patients by their criteria. Interestingly, TEE was performed in only 21% of the 566 patients that had any echo performed, and only 5% had only a TEE performed. Thus, their criteria would actually increase the number of TEEs performed in these seven hospitals. It would appear that TEE is underutilized in practice and few are following the Infectious Disease Societies’ recommendation that all Staph bacteremia patients should have TEE. In the absence of better data, we should employ these new, more selective criteria, which are skewed toward higher sensitivity, since Staph IE is such a serious disease with a high incidence of morbidity and mortality.