Echocardiography is the cornerstone imaging technique in the diagnosis of infective endocarditis (IE) but is often misused in clinical practice. Recently, strict negative criteria have been proposed to avoid unnecessary follow-up echocardiograms. We aimed to evaluate the use of echocardiography in real-world clinical daily practice and the usefulness of these criteria in the diagnosis of IE. We retrospectively retrieved every echocardiogram performed in our center for suspected IE between 2014 and 2018, including 905 transthoracic echocardiograms (TTEs). Of these, 451 (49.8%) fulfilled the strict negative criteria (group 1). In this group, IE was seldom diagnosed (n = 4, 0.9%). In 338 patients (37.4%) no signs of IE were evident, but they did not fulfill the strict negative criteria (group 2). A follow-up echocardiogram and definitive diagnosis of IE were more frequent (n = 48, 14.2% and n = 20, 5.9%). Finally, in 116 patients (12.8%) the initial TTE showed typical or suggestive signs of IE, in whom the diagnosis was confirmed in 48 patients (41.4%). A definitive diagnosis of IE was established in a minority of the study population (n = 72, 8%). Only 1 readmission for underdiagnosis of IE was noted on group 2. We conclude that in a real-life setting only a minority of patients in whom IE was suspected had a definitive diagnosis. An initial TTE for suspected IE fulfilling the strict negative criteria predicts both a low probability of requesting a follow-up study and of a definitive diagnosis of IE.
Infective endocarditis (IE) is a rare but potentially lethal disease whose incidence appears to be increasing worldwide. The diagnosis is challenging and a high clinical suspicion and a low threshold for complementary tests that can confirm or exclude the diagnosis are mandatory. Echocardiography is the cornerstone imaging technique both in the diagnosis and management of IE, , and there is consensus that, whenever IE is suspected, an echocardiographic study is indicated as early as possible. , However, no clear criteria exist on the definition of “clinical suspicion of IE” and echocardiography is often overused and inappropriately ordered in clinical practice to rule out the disease in patients with a low pretest probability. , Many groups have proposed a more restrictive screening based on several predictors. It appears that an initial negative transthoracic echocardiogram (TTE) has a high negative predictive value for IE detection, , and the strict negative criteria have been recently described as an attempt to define a high-quality, initial negative TTE to avoid unnecessary follow-up echocardiograms. The purposes of our study are (1) to review the use of the initial and follow-up echocardiography for suspected IE in contemporary clinical daily practice; and (2) to confirm the usefulness of the echocardiographic strict negative criteria to potentially avoid unnecessary follow-up studies.
This is a retrospective analysis from a single-center prospective registry of every echocardiography performed in the Cardiology department of a tertiary-care teaching hospital in Valencia (Spain). We retrospectively retrieved every echocardiogram requested for suspected IE between January 1, 2014, and December 31, 2018, from our local echocardiographic database (EcocardioPlus V2.0, Neodata Business SL, Barcelona, Spain), which is a prospectively maintained digital database of all echocardiograms performed in our hospital. From a total of 32,572 studies, 1,251 echocardiograms were indicated for fever of unknown origin, bacteremia, IE suspicion, or showed findings suggestive of IE. Of note, 75 echocardiograms were excluded for inadequate selection (e.g., the inclusion of ambulatory echocardiograms or initial transesophageal echocardiogram [TEE]) and 271 were follow-up echocardiograms performed in 139 patients. The final study group was comprised 905 patients ( Figure 1 ). We registered the specialty which referred the patient for the echocardiogram, the time at which the test was requested and performed, the type of echocardiogram (TTE or TEE), and, if follow-up echocardiograms were performed, the type, temporality, and the number of these studies. We individually interpreted every echocardiogram and identified cases with signs of suspected IE.
Diagnosis of IE was established according to the Duke criteria, as recommended in clinical practice guidelines. Echocardiograms with equivocal signs of IE were also identified and registered. The following 2 commercially available systems were used throughout the study: iE33 and EPIQ 7 (Philips, Cambridge, Massachusetts), using an X5-1 xMATRIX array transducer or an X7-2t Live 3D xMATRIX array transducer for TTE or TEE, respectively.
Clinical outcomes during hospital admission and at 6 years follow-up were reviewed based on electronic medical records in our health care system. Researchers in charge of end point adjudications were all blinded to the study groups. This registry was performed in accordance with the Declaration of Helsinki and was approved by our local ethics committee.
With the echocardiographic data exported from the database, we applied the strict negative criteria defined by Sivak et al in their original publication ( Figure 2 ). Because these criteria only apply to patients in which an initial TTE was performed, we excluded the patients in whom a TEE was the first test performed (e.g., in the presence of prosthetic valves or intracardiac devices).
After performing an initial TTE for IE suspicion, we identified 3 risk groups in which the indication for a follow-up echocardiogram may differ. In group 1, the strict negative criteria were fulfilled, so a follow-up echocardiogram was generally not indicated. In group 2, neither the strict negative criteria were fulfilled, nor signs of IE were identified based on initial TTE. In this group, a follow-up echocardiogram could be indicated depending on the pretest probability and clinical and microbiologic likelihood of IE. Finally, in group 3 the strict negative criteria were not fulfilled, and typical or suggestive signs of IE were evident on initial TTE. Further study is generally indicated these cases.
The primary objective of the study was the percentage of TTEs ordered for IE clinical suspicion that fulfill the echocardiographic strict negative criteria, and the number of follow-up studies and definite IE diagnoses according to the study groups. As secondary analyses, we evaluated the echocardiographic predictors of a repeat echocardiogram in the follow-up, the clinical outcomes of the patients in whom IE was either diagnosed or ruled out by the initial echocardiogram, and the outcomes between a diagnosis of IE and another diagnosis at discharge.
Data were tested for normal distribution using the Kolmogorov-Smirnov test. Continuous normally distributed data were expressed as the mean ± SD of the mean and compared using the paired and unpaired Student’s t tests. Nonparametric data were expressed as the median with the interquartile range and compared using the Mann-Whitney U test. Group percentages were compared using the chi-square test or Fisher’s exact test where appropriate.
Using variables with p <0.1 in univariate analysis, a multivariate binary logistic regression model was performed to predict the occurrence of a follow-up echocardiogram after an initial TTE. Hazard ratios with the corresponding 95% confidence intervals were computed.
Survival analysis between the groups was performed using Kaplan-Meier curves and log-rank analysis.
Statistical significance was considered for a 2-tailed p <0.05. The statistical software SPSS (SPSS for Windows, version 15.0. SPSS Inc., Chicago, Illinois) was used.
Baseline clinical characteristics of the cohort are listed in Table 1 . The mean age was 64.8 ± 16 years and most patients (60.9%) were male. Comorbidities in the cohort were relatively high, especially hypertension, diabetes mellitus, dyslipidemia, chronic kidney disease, and cancer. A follow-up echocardiogram was more frequently requested in patients with valvular heart disease and cardiac valve prosthesis and less frequently in patients with cancer. A trend toward more frequent requests in males and patients with a history of parenteral drug abuse was also observed. Patients were mostly referred by the department of internal medicine (n = 377, 42%), followed by nephrology (n = 97, 11%) and cardiology (n = 81, 9%) ( Figure 1 ).
|Variable||All patients (n=905)||Follow-up echo||p|
|No (n=766)||Yes (n=139)|
|Age (years)||64.8 ± 16||64.77 ± 16.17||64.99 ± 15.08||0.879|
|Men||551 (60.9%)||457 (59.7%)||94 (67.6%)||0.089|
|Hypertension||505 (55.8%)||422 (55.1%)||83 (59.7%)||0.353|
|Diabetes mellitus||273 (30.2%)||226 (29.5%)||47 (33.8%)||0.316|
|Dyslipidemia||398 (44%)||330 (43.1%)||68 (48.9%)||0.227|
|Smoking habit||292 (32.3%)||249 (32.5%)||43 (30.9%)||0.768|
|Heart failure||135 (14.9%)||112 (14.6%)||23 (16.5%)||0.604|
|Coronary artery disease||111 (12.3%)||96 (12.5%)||15 (10.8%)||0.673|
|Valvular heart disease||123 (13.6%)||86 (11.2%)||37 (26.6%)||<0.001|
|Cardiac valve prosthesis||77 (8.5%)||54 (7%)||23 (16.5%)||0.001|
|Pacemaker||59 (6.5%)||47 (6.1%)||12 (8.6%)||0.265|
|Peripheral artery disease||76 (8.4%)||68 (8.9%)||8 (5.8%)||0.249|
|Chronic kidney disease||228 (25.2%)||187 (24.4%)||41 (29.5%)||0.204|
|Hemodialysis||72 (8%)||59 (7.7%)||13 (9.4%)||0.497|
|Chronic obstructive pulmonary disease||69 (7.6%)||61 (8%)||8 (5.8%)||0.487|
|Cancer||215 (23.8%)||194 (25.3%)||21 (15.1%)||0.009|
|Previous IE||16 (1.8%)||12 (1.6%)||4 (2.9%)||0.289|
|HIV infection||32 (3.5%)||26 (3.4%)||6 (4.3%)||0.616|
|Parenteral drug use||16 (1.8%)||11 (1.4%)||5 (3.6%)||0.084|
|Immunosuppressive therapy||96 (10.6%)||82 (10.7%)||14 (10.1%)||1|