Background: The left ventricular ejection fraction (LVEF) measurement is a common tool for evaluating the LV systolic function. The application of the global longitudinal systolic strain (GLS) parameter in the assessment of the myocardial function has also received special attention recently.

Objectives: This study was aimed at comparing the LVEF and LV volumes obtained by the two methods of catheterization and two-dimensional (2D) echocardiography (available in our institution) and assessing the correlation between the LVEF and the GLS.

Patients and Methods: In this cross-sectional study, 45 patients were recruited from coronary angiography candidates. The patients underwent echocardiography immediately before catheterization. The LVEF and LV volumes were measured via echocardiography using the apical four- and two chamber-views. The GLS was calculated through the automated functional imaging algorithm. Left ventriculography was performed by calculating the LVEF in the right and left oblique views.

Results: The LVEF values obtained by the two methods of ventriculography and echocardiography were not significantly different. The highest correlation regarding the echocardiographic LVEF was obtained in the angiographic right anterior oblique view (P < 0.001, r = 0.95). There was a good agreement as regards the biplane LVEF between 2D echocardiography and ventriculography (-0.5 ± 13.27; CI of 95%). The GLS showed a significant correlation with the estimated EF in both methods, the highest being with the Biplane Simpson method (r = -0.84; P < 0.001). Linear regression was used to obtain the formula for estimating the 2D LVEF from the GLS [LVEF = 2.53 (GLS) + 10.48]. The GLS values ≥ -21.7% and≤ -11.7% were consistent with normal and severe global LV systolic dysfunction, respectively. The inter- and intra-observer agreement was more evident in the GLS measurement rather than in the LVEF.

Conclusions: Despite the widespread use of 2D LVEF and its good agreement with ventriculography, strain analysis seems to be more reliable as a quantitative tool for ventricular assessment.

The evaluation of the left ventricular (LV) systolic function is critical in the appropriate management of patients with different heart diseases. The estimation of the LV ejection fraction (EF) is the most common method for this purpose (1).

Multiple imaging modalities have been used for such assessment, including echocardiography, contrast ventriculography, radionuclide ventriculography, and cardiac magnetic resonance imaging. Tissue Doppler imaging (including deformation indices), three-dimensional (3D) echocardiography, and two-dimensional (2D) speckle-tracking echocardiography (2D-STE) have also been widely utilized in the evaluation of the LV function in recent decades (2, 3). There is, however, a dearth of data on the degree of agreement between the different techniques. Moreover, finding reference values for the normally functioning or dysfunctional ventricle would be a challenging issue.

The present study sought to compare the reliability of two imaging modalities available in our institution, i.e. noninvasive echocardiography and invasive cine-angiography, with respect to the LV volumes and LVEF. The global longitudinal strain (GLS) (derived by 2D-STE) was also calculated to determine reference values for the severely depressed and the normal systolic functions.

This cross-sectional study was conducted from March 2011 to 2012 on 54 consecutive patients with stable angina who were candidates for elective coronary angiography according to the American College of Cardiology/American Heart Association (ACC/AHA) guideline (4). Echocardiography was done in all the patients immediately before catheterization.

The exclusion criteria were defined as poor LV endocardial borders (on echocardiography or ventriculography), heart rate over 110 per minute, multiple ventricular extra systoles, atrial fibrillation, hemodynamic instability or significant aortic or mitral regurgitation, LV thrombus, decompensated heart failure, and extreme elevation of the LV end-diastolic pressure. Forty-five patients met the inclusion criteria and were enrolled in the study for further analysis. This study was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans and obtained the approval of the Research Ethics Committee of Mashhad University of Medical Sciences (code No:2513339). Informed written consent was obtained from each participant.

Forty-five patients, at a mean age of 57.45 ± 11.75 years (range: 32-84 years), were studied between the years 2011 and 2012. Ten (22.2%) patients had normal coronary arteries, whereas one had a history of coronary artery bypass graft surgery with patent grafts. Significant coronary artery stenosis (> 50%) was present in the remaining cases (75.58%) with the involvement of the left anterior descending, left circumflex, and right coronary arteries in 61.9%, 50%, and 47.6% of the patients, respectively. Three patients had a history of previous infarction.

The LVESV, LVEDV, and LVEF, which were recorded during catheterization and echocardiography, are presented in Table 1. The agreement between the two methods (ventriculography and 2D echocardiography) was tested considering the volumetric parameters and the EF (Table 2). There was a good agreement between these two (in both RAO and LAO projections) as regards the measurement of the LVESV and LVEF (considering CI 95%). The LVEDV measured in the LAO view had less agreement with 2D echocardiography compared with the RAO projection. Furthermore, while it seemed that the estimated volumes were larger than the data in 2D echocardiography in the RAO view, in the LAO projection, it was vice versa. The data regarding the ejection fraction were comparable.

Table 1. Measurements in Ventriculography and Echocardiography a, b

Table 2. Correlation between LVEF and GLS in Catheterization and Echocardiography a,b

There was a significant correlation between the GLS and the EF obtained from catheterism and echocardiography (Table 3), the highest correlation being between the GLS and the Biplane EF (P < 0.001, r = -0.848) (Figure 3).

Table 3. Inter- and intra-Observer Agreement in Echocardiography (n = 12) with the Correlation and Mean Difference between the Measurements a

The relationship between the GLS and the LVEF in the catheterization and echocardiography methods was studied by the linear regression test, separately. An appropriate and meaningful association was found between the constant coefficients (a) and the constant values (b) for each variable. According to the r square test and the significance of the constant values, a meaningful linear relationship can be defined by the following formula:

LVEF = a (GLS) + b

The values which stand for constant parameters (a and b) are presented in Table 3. They somehow differed based on the selected method (LVEF in catheterization in the RAO and LAO views versus LVEF in echocardiography in the four-chamber view or by the modified biplane Simpson method). Strain imaging (GLS) was used for estimating the severity of the LV systolic dysfunction. The reference values for the GLS in the LVEF estimation were -11.7% for severe LV systolic dysfunction and -21.7% for the normal systolic function.

The LVEF assessment has a critical role in the diagnosis, management, and prognosis of patients with different cardiovascular diseases. Echocardiography by itself plays a pivotal role in assessing the LV systolic function, mainly by calculating the LVEF, and is as such deemed the most widely employed method. Cardiac catheterism and radionucleotide ventriculography have been accepted as diagnostic methods, but they have such limitations as invasiveness, high cost, and high radiation exposure (6). There are conflicting data on the assessment of the LV volumes by conventional 2D echocardiography. In the study of Thomson et al. (6), 2D quantification was likely to underestimate the non-opacified LVEDV by 30-40%.

The agreement between the angiographic LVEDV and echocardiographic data was more apparent when comparing the RAO projection with the LAO view. The results (mean difference with 2D echocardiography) for the LVESV or the LVEF were comparable between the angiographic projections. This stands in contrast with the Habash-Bseiso et al. (7) study, in which the echocardiographic LVEF was lower than that determined by angiography (49% ± 1.0% vs. 54% ± 1.0%; P < 0.0001). That study also reported a significant correlation between the angiographic LVEF and the echocardiographic LVEF (r = 0.70, P < 0.0001). The geometric assumptions or angiographic magnification could be the source of errors in estimating the volumes by ventriculography (8).

Tissue Doppler-derived strain and strain rate imaging were introduced several years ago as a tool for the quantification of the myocardial mechanical function (9). Nonetheless, their angle dependency, noise interference, and substantial intra-observer and inter-observer variability resulted in a great tendency toward using the newer technique of 2D-STE (10).

2D-STE allows the evaluation of the regional strain values from all the LV segments, which can be averaged to global strain. Global strain is a relatively new parameter for the assessment of the LV function and has been introduced as the best echocardiographic parameter to predict the global infarct mass, which is superior to the more traditional echocardiographic parameters such as the LVEF and the wall motion score index (11, 12). The existing literature, however, contains few reports on the LVEF estimation via the GLS (13), and there is a paucity of information on a reference normal range for the LVEF. The values of -18.6% (with 0.1% of the limits of agreement) (14), -20.6% ± 2.6% (15), and 21.3 ± 2.1% (16) have been proposed as the normal range by three studies in which the GE system was used for assessment. In our study, -21.7% was considered the reference value for the normal group, which chimes in with the latter two studies. Takigiku et al. (16) reported that 2D image tracking ability in each ultrasound system was different, though less in longitudinal strain analysis compared with the radial or circumferential one.

In one study on patients with coronary artery diseases, the GLS assessed by the automated functional imaging algorithm was linearly related to the biplane LVEF (r = 0.83; P < 0.001), whereas the correlation was less in patients with myocardial infarction or heart failure (r = 0.42 and r = 0.62, both Ps < 0.001) (17). In another study on patients with previous infarction, the LVEF was measured by magnetic resonance imaging (MRI) and 2D and 3D echocardiography. The GLS was also determined. In regard to the EF, the correlation between the GLS and the MRI was significant and better than that of 2D echocardiography and MRI (13).

Infraction or heart failure cases were not specified in our study. There was a significant linear correlation between the LVEF estimated by both methods and the GLS in the overall study population. On the other hand, the inter- and intra-observer variability was less with the GLS than that with 2D echocardiography regarding the LVEF. The GLS can be used interchangeably with great confidence for evaluating the LV systolic function, especially in patients who need close surveillance and monitoring for changes in the ventricular function. Assuming that the endocardial borders could be well defined, the GLS would be a more reliable method inasmuch as it is less operator-dependent than 2D echocardiography when measuring the LVEF. The volumetric parameters and the EF obtained from noninvasive 2D echocardiography and invasive cine-angiography showed a desirable agreement with no significant difference. The RAO projection would be the preferred view for such assessment. The GLS parameter yielded a favorable correlation with the LVEF; accordingly, it can be suggested as a reliable method for the LV function assessment in routine practice, especially in patients monitored or tracked for the changes in the function.

This work was derived from a student thesis (No. 2078) and supported by a grant from the Research Council of Mashhad University of Medical Sciences. The authors of this manuscript certify that they comply with the Principles of Ethical Publishing in the Achieves of Cardiovascular Imaging.