Author + information
- Published online April 1, 2019.
- Andrei D. Mărgulescu, MD, PhD∗ (, )
- Maria-Claudia-Berenice Şuran, MD and
- Dragoş Vinereanu, MD, PhD
- ↵∗Department of Cardiology, University and Emergency Hospital of Bucharest, University of Medicine and Pharmacy “Carol Davila” Bucharest, Spl. Independentei 169, Bucharest 050092, Romania
Two-dimensional speckle tracking echocardiography (2DSE) allows more accurate assessment of left ventricular (LV) volumes and ejection fraction (EF) compared with planimetric or visual assessments, and can assess myocardial deformation (strain and strain rate) both in the LV and the left atrium (LA). Introduction of 2DSE in clinical practice has been advocated based on good reported reproducibility, but the effect of image acquisitions and degree of LV systolic dysfunction on reproducibility has not been thoroughly assessed (1). We aimed to assess the reproducibility of combined acquisition and measurements of 2DSE indexes and to estimate the difference between repeated measurements that could be considered real and not due to reproducibility alone (measurement uncertainty) across the spectrum of EF.
We included 72 subjects (age 63 ± 14 years; 65% men) split into 4 equal-sized groups according to EF and presence of cardiovascular risk factors and/or overt cardiac pathology: group 1: EF ≥50%, structurally normal heart, no risk factors or cardiovascular disease; group 2: EF ≥50%, presence of hypertension and diabetes, or established cardiovascular disease; group 3: EF 30% to 49%; and group 4: EF <30%. We excluded subjects with acute cardiovascular conditions, irregular heart rhythms, and inappropriate speckle tracking in ≥2 LV segments in a single echocardiographic view. The study was approved by the University and Emergency Hospital of Bucharest local ethics committee, and subjects gave written informed consent.
Two-dimensional gray-scale loops, acquired at 50 to 80 frames/s from the parasternal short-axis (at mid-papillary muscle level), and apical 4-chamber, 2-chamber and long-axis views were stored for off-line analysis, as recommended (1). A single machine (Vivid E9, GE Healthcare, Horten, Norway) and software (EchoPAC, Version 113, GE Healthcare) were used for all studies. LV volumes (left ventricular end-diastolic volume [LVEDV]; left ventricular end-systolic volume [LVESV]) and EF were measured using the AutoEF package (GE Healthcare) (2). Global longitudinal LV systolic strain (GLS) and strain rate (GLSR) were calculated using the Automated Function Imaging (AFI) package (GE Healthcare) as the average from the 3 standard apical views, as recommended (1). Global circumferential (GCS) and radial systolic LV strain (GRS) were measured from the parasternal short-axis view. Global longitudinal LA strain during systole (GLASs) and LA contraction (GLASa) were measured as the average from the apical 4- and 2-chamber views, using the R-wave as reference (R-R gating) (3).
During the first echocardiography session, 2 sequential acquisitions were performed by 1 observer, followed by a third acquisition by a second observer. The first observer repeated image acquisition after a median of 1 day. Each observer performed measurements on their own acquisitions. The first observer also repeated measurements on the acquisition performed by the second observer. Thus, we assessed intraobserver, interobserver, and test-retest reproducibility of combined acquisition and measurements of 2DSE indexes, and the interobserver reproducibility of measurements only (on the same image dataset), respectively.
Coefficients of variation (CVs), and mean difference between measurements for all 2DSE indexes are summarized in Table 1. Test-retest CVs were worse than intraobserver CVs. Interobserver CVs were generally intermediate between intraobserver and test-retest CVs. EF, GLS, and GLSR had the best CVs of all 2DSE indexes, but their reproducibility decreased in patients with poor EFs. GRS and GCS had poor CVs, especially in patients with poor EFs, whereas GLASs and GLASa had only moderate CVs. The measurement uncertainty for a scenario in which the same echocardiographer followed individual patients is also provided in Table 1.
Our data may have direct clinical application, as follows. 1) EF (AutoEF), GLS, and GLSR might have acceptable reproducibility for clinical practice. 2) Whenever possible, the same echocardiographer should perform both acquisition and measurements in follow-up of the patients. 3) The EF spectrum of “heart failure with mildly reduced EF” (i.e., EF of 40% to 49%) was covered completely by the test-retest variability in subjects assigned to groups 2 and 3; this underlines the difficulty of defining this group in clinical practice based on echocardiography. 4) In patients who are receiving cardiotoxic cancer chemotherapy, a relative change of 15% in GLS from baseline (∼3% absolute change for low-normal GLS) might also be due to the test-retest variability of GLS, and not to cardiotoxicity; thus, diagnosing cardiotoxicity in individual patients should use an integrative approach and not mere cutoff values (4). 5) GRS, GCS, GLASs, and GLASa have limited usefulness for individual patient assessment, but might be used in clinical research with a large number of patients.
Please note: The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2019 American College of Cardiology Foundation
- Pathan F.,
- D'Elia N.,
- Nolan M.T.,
- Marwick T.H.,
- Negishi K.
- Plana J.C.,
- Galderisi M.,
- Barac A.,
- et al.