Is Left Atrial Function Affected by Coronary Slow Flow? A Two-dimensional Speckle-Tracking Echocardiographic Study.

BACKGROUND: The coronary slow flow phenomenon (CSFP) is the slow passage of the angiographic contrast agent to the distal portion of the coronary artery in the absence of significant stenosis. We evaluated the left atrial (LA) function in patients with the CSFP using two-dimensional speckle-tracking echocardiography (2DSTE).
METHODS: The LA function was compared through 2DSTE between 36 patients with the CSFP and 36 participants with a normal coronary flow. The two groups were matched for age, sex, hypertension, diabetes mellitus, and the left ventricular function.
RESULTS: There were no statistically significant differences between the CSFP group and the control group regarding longitudinal systolic strain, early and late diastolic strains, and the strain rate of the LA myocardium.
CONCLUSIONS: The LA function as evaluated with 2DSTE was not different between the CSFP group and the normal coronary flow group when they were matched for age, sex, hypertension, diabetes, and the left ventricular function.

INTRODUCTION

The coronary slow flow phenomenon (CSFP) is delayed opacification of the coronary arteries during coronary angiography without significant coronary artery stenosis.[1] The incidence of the CSFP is about 1% of patients undergoing coronary angiography for the evaluation of chest pain.[2] The exact etiology of the CSFP is not clear. Autonomic nervous system, microvascular disease, endothelial dysfunction, atherosclerosis, inflammatory disease, and anatomic factor of vessels can be the contributing factors.[3]

Echocardiographic evaluation of the left atrial (LA) size and function is important and seems to have a prognostic role because LA malfunction or LA enlargement is associated with atrial fibrillation, stroke, and increased risk of cardiovascular morbidity and mortality.[45] The LA function has three phases: the reservoir phase at the ventricular systole time, when the LA receives the drainage of the pulmonary vein drainage; the conduit phase at early ventricular diastole, when the LA passes the blood to the left ventricle (LV); and the booster pump phase at late ventricular diastole, when the LA contracts actively and is responsible for 15%–30% of the LV stroke volume.[6]

Two-dimensional and Doppler echocardiography can assess the LA function, but the indices by these methods depend on hemodynamic loading and geometric measurements.[7] For these reasons, two-dimensional speckle-tracking echocardiography (2DSTE) was developed. 2DSTE enjoys better reproducibility and is angle independent by comparison with prior methods.[7] 2DSTE is the current technique for a better assessment of the LA function.[8] 2DSTE-derived indices of the LA function are associated with atrial fibrosis,[910] and the presence of myocardial fibrosis in the myocardial samples of participants with the CSFP has been demonstrated.[11] Furthermore, these indices are able to detect subclinical processes.[12]

Previous studies have assessed impaired LV diastolic function in the CSFP.[1314] The LV diastolic dysfunction is an important cause of heart failure. Diastolic dysfunction is associated with an elevated LV end-diastolic pressure and leads to increased LA wall tension, LA enlargement, and risk of atrial fibrillation. Advanced diastolic dysfunction is a significant cause of increased mortality.[815] The existing literature contains only one study on the LA function. In that study, the LV function as assessed with 2DSTE was decreased in patients with the CSFP, too.[16]

The aims of the present study were to evaluate the LA deformation by 2DSTE in patients with the CSFP by controlling the LV function in a 1-patient: 1-control participant design matched for age, sex, hypertension, and diabetes study.

METHODS

Patients

Thirty-six cases for the patient group and 36 cases for the control group – matched for age, sex, hypertension, and diabetes mellitus in a 1:1 patient:control design – were selected from a study by Narimani et al.,[17] who published their results on the 2DSTE assessment of the LV systolic and diastolic functions in the CSFP. Briefly, the authors evaluated 36 patients from 200 patients with the CSFP based on the recommendations of Gibson et al.[18] The exclusion criteria were an LV ejection fraction <55%; coronary ectasia; pulmonary hypertension; dilated or hypertrophic cardiomyopathy; congenital heart disease; pericardial disease; valvular heart stenosis with any degree; more than mild valvular heart regurgitation; myocardial infarction; history of inflammatory disease; history of chemotherapy or chest radiotherapy; liver, kidney, or thyroid disease; left bundle branch block; paced rhythm; history of cardiac surgery or percutaneous coronary intervention; and poor echocardiographic imaging. The control group was selected from the Coronary Angiography Data Bank of our hospital. We used the echocardiographic imaging obtained by Narimani et al.[17] (with permission) to evaluate the LA function. The research proposal was approved by our Institutional Review Board.

Echocardiography

Left atrial volumes

For an accurate evaluation of the LA size, we measured the LA volumes at three different times in the cardiac cycle: (1) largest LA volume at end systole, just before the opening of the mitral valve (maximal LA volume); (2) smallest LA volume at end diastole, just before the closing of the mitral valve (minimal LA volume); and (3) at mid-diastole, before the onset of the P wave on electrocardiogram (pre-A LA volume). We used the “Auto EF” option for the assessment of the LA volumes. Tracing the endocardial border of the LA demonstrated the curve of the LA volume during the cardiac cycle. Afterward, the LA maximum volume, LA minimum volume, and pre-A LA volume were measured in the apical 4-chamber and 2-chamber views and the mean of the values was recorded. Several volumetric indices were assessed using the following volume measurements:

Left atrial reservoir function:

Maximal LA volume − minimal LA volume = LA total emptying volume

100 × LA total emptying volume/maximal LA volume = LA total emptying fraction

100 × LA total emptying volume/maximal LA volume = LA expansion index

LA conduit function:

100 × (maximal LA volume − pre-A LA volume)/maximal LA volume = LA passive emptying fraction

100 × (maximal LA volume − pre-A LA volume)/LA total emptying volume = LA passive emptying percentage of total emptying

LA booster pump function:

100 × (pre-A LA volume − minimal LA volume)/LA total emptying volume = LA booster active emptying percentage of total emptying

100 × (pre-A LA volume − minimal LA volume)/LA pre-A volume = LA active emptying index.

Two-dimensional speckle-tracking echocardiography

Echocardiographic images were stored at 40–80 frames per se cond in three cardiac cycles in the apical 4- and 2-chamber views. In the first step, the endocardial border of the LA was manually traced at end ventricular systole. Then, the epicardial borders of the LA wall were automatically defined by software (Samsung Medison software for 2DSTE). After the adjustment and confirmation of these traced borders, each LA wall was divided into three segments by the software. In patients with adequate image quality, 12 segments were analyzed. The starting point of strain analysis was set as the initiation of the QRS wave. In the segments with poor tracking, the endocardial border was readjusted to achieve better tracking. Segments with poor tracking, despite multiple readjustments, were excluded from the study. The averaged value of the accepted segments was recorded as the global value of each index. In sum, 850 (98.4%) segments were analyzed. All the image analyses were done by a single expert echocardiologist. The longitudinal LA strain curve had a positive systolic peak (SS), an early diastolic plateau (SE), and a negative late diastolic peak (SA). The difference between SS and SE was defined as early diastolic strain (EDS), and the difference between SE and SA was defined as late diastolic strain (LDS). The longitudinal LA strain rate curve had a positive systolic peak (SRS), an early negative peak at early diastole (SRE), and a late negative peak at late diastole (SRA) [Figure 1]. SS and SRS represented the LA reservoir function, EDS and SRE reflected the LA conduit, and LDS and SRA were the indices of the LA booster function. Eleven randomly selected patients were reanalyzed to evaluate inter- and intra-observer variabilities, 1 month after the last analysis.

Figure 1

Longitudinal deformation indices of the left atrial myocardium. (a) Strain curve, SS = Systolic longitudinal strain, EDS = Early diastolic strain, LDS = Late diastolic strain, (b) Strain rate curve, SRS = Systolic strain rate, SRE = Early diastolic strain rate SRA = Late diastolic strain rate

Statistical analysis

The continuous data are demonstrated as means and standard deviations if normally distributed; otherwise, they are presented as medians and interquartile ranges. The normally distributed data were compared using the pairwise Student's t-test, and the skewed data were compared using the Wilcoxon signed-rank test. P ≤ 0.05 was considered statistically significant. The statistical analyses were done through IBM SPSS statistics for Windows (version 23.0) (IBM Corp., Armonk, NY, USA).

RESULTS

The demographic, clinical, and angiographic data of the study patients have been presented previously.[17] In summary, the age of the patients with the CSFP and the normal participants was 53.9 ± 8.3 years and 54.5 ± 9.4 years, respectively (P = 0.370). Eleven (31%) participants in each group were female, 15 (42%) participants in each group were hypertensive, and 7 (19%) participants in each group had diabetes. There were no statistically significant differences between the LV systolic strain and strain rate and early and late diastolic strain rates.[17] The other indices of the LV systolic and diastolic functions in our study population have been previously presented elsewhere.[17]

There were no statistically significant differences regarding the volumetric parameters of the LA function and the 2DSTE-derived indices of the LA function [Table 1].

Table 1

Echocardiographic data of the patients with the coronary slow flow phenomenon and the control subjects

Variables	Coronary slow flow group (n=36)	Control group (n=36)	P
Maximal LA volume index (mL/m2)	25.1±6.5	27.7±7.9	0.290
Minimal LA volume index (mL/m2)	7.9 (5.4-10.5)	8.0 (5.9-11.5)	0.342
Pre-A LA volume index (mL/m2)	14.4±5.2	15.3±5.8	0.420
LA total emptying volume (mL)	33.2±9.3	32.1±9.2	0.565
LA expansion index (%)	212.6 (171.4-298.7)	207.2 (162.7-236.2)	0.530
LA total emptying fraction (%)	68.5±8.2	66.2±8.3	0.308
LA passive emptying percentage of total emptying (%)	63.7±13.8	63.6±15.2	0.981
LA passive emptying fraction (%)	43.4±10.1	42.3±11.6	0.610
LA booster active emptying percentage of total emptying (%)	37.1 (26.4-45.2)	35.9 (24.4-47.6)	0.925
LA active emptying index (%)	43.7±13.9	40.5±13.4	0.365
SS (%)	41.5±13.5	44.6±13.8	0.392
EDS (%)	23.3 (16.5-30.6)	24.1 (17.6-29.0)	0.950
LDS (%)	21.1±6.7	23.4±7.8	0.273
SRS (1/s)	2.8±0.8	2.7±0.8	0.831
SRE (1/s)	−2.0 (−2.6-−1.7)	−2.2 (−2.5-−1.7)	0.838
SRA (1/s)	-3.0+/-1.0	-2.9+/-0.9	0.660

Data are presented as mean±SDs for the variables with a normal distribution and medians, and interquartile (25th-75th) ranges for the variables with a skewed distribution. EDS=Left atrial early diastolic longitudinal strain, LA=Left atrial, LDS=Left atrial late diastolic longitudinal strain, SRA=Left atrial late diastolic longitudinal strain rate, SRS=Left atrial systolic longitudinal strain rate, SRE=Left atrial early diastolic longitudinal strain rate, SS=Left atrial systolic longitudinal strain, SDs=Standard deviations

The interobserver variabilities as the coefficient variations for SS, EDS, LDS, SRS, SRE, and SRA were 8.5%, 9.3%, 7.2%, 9.2%, 9.7%, and 8.8%, correspondingly, and the intraobserver variabilities as the coefficient variations for SS, EDS, LDS, SRS, SRE, and SRA were 6.5%, 7.2%, 6.9%, 8.2%, 8.9%, and 7.8%, correspondingly.

DISCUSSION

The LA size and function have an important clinical and prognosis impact.[19] 2DSTE is a new technique for the analysis of the LV function, but many studies use 2DSTE for the other chambers such as the LA.[2021] Prior studies have demonstrated that strain imaging can detect the LA dysfunction before changes in the LA morphologic parameters.[22] 2DSTE is an easy technique for the evaluation of the LA function because it is semi-automated and confers offline processing.[19] 2DSTE, in contrast to the Doppler parameters of the LA function, is angle independent and is less affected by artifacts.[20]

In the present study – through LV function control and in a 1:1 patient-control matching design for age, sex, hypertension, and diabetes – we compared the LA function as evaluated through 2DSTE between patients with the CSFP and participants with a normal coronary flow. Our results showed no statistically significant differences in the LA deformation indices (i.e., SS, EDS, LDS, SRS, SRE, and SRA) between the patients with the CSFP and the control group. We found no statistically significant differences in all the LA functions as assessed by 2DSTE. It is not possible to make a thorough comparison between our results and those reported elsewhere because the study by Wang et al.[16] is the only investigation in the existing literature on the assessment of the LA function in patients with the CSFP. Evaluating the LA function in 82 patients with the CSFP and 55 participants with a normal coronary flow through 2DSTE and volumetric measurements, the authors observed that the patients with the CSFP had decreased EDS and SRE as the markers of the LA conduit function and increased LDS and SRA as the indices of the LA booster pump function compared with the control participants, whereas they found no statistically significant differences regarding the volumetric parameters of the LA function.[16] Furthermore, they found that systolic strain and the systolic and early diastolic strain rates of the LV were decreased in the CSFP group. It should be considered that unlike the study by Wang et al.,[16] our study drew upon a 1:1 patient-control matching design for age, sex, hypertension, and diabetes mellitus. This type of matching can be useful in the presence of interaction between confounding variables. For example, an interaction between hypertension and diabetes vis-à-vis the LA function as evaluated by 2DSTE has been shown previously.[2324]

Further, it has been demonstrated that the LV function is a major determinant of the LA function.[25] In the study by Wang et al.,[16] the LV function was reduced in the patients with the CSFP and the authors showed that the LA function was significantly correlated with the LV function; nonetheless, they failed to demonstrate whether or not the CSFP was a determinant of the LA function. In our study, there was no statistically significant difference between the two groups with respect to the LV function, and we controlled this confounding factor. The difference in the vendor is another matter which should be taken into consideration. Coronary artery ectasia is associated with the CSFP.[26] In a recent study, Aghajani et al.[27] showed that the LA function as assessed by 2DSTE was not impaired compared with that in normal participants, which is concordant with our findings.

Chiming in with the study by Wang et al.,[16] we found no significant differences in the volumetric indices of the LA function between our two study groups. These findings support our other findings obtained through the application of 2DSTE.

What should also be taken into account is that the difference in the LA function between our two study groups may have proven undetectable by methods such as 2DSTE. Accordingly, more meticulous methods such as cardiac magnetic resonance imaging and 3D echocardiography may be required.

Study limitations

The present study is a retrospective study and conducted in a single center with a low sample size. 2DSTE has its own shortcomings, most notably its potential error in the tracing of the endocardial and epicardial borders in suboptimal images and the risk of contamination by signals from the structures surrounding the LA. Another drawback of note is our assessment of the LA function using software designed for 2DSTE evaluation of the LV; we might have obtained more reliable results had we been able to evaluate the LA function with cardiac magnetic resonance imaging and 3D echocardiography. Normal values for the strain and strain rates of the LA as assessed by Samsung Medison software have yet to be published. The vendor employed in our study lacked the option for changing the starting point of strain analysis from the beginning of the QRS wave to the P wave.

CONCLUSIONS

In our patients with the CSFP, the functions of the LA (i.e., reservoir, conduit, and booster pump) as evaluated through 2DSTE were not different from those of our participants with a normal coronary flow when they were matched age, sex, hypertension, diabetes, and the LV function.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.