Left Atrial Deformation And Risk of Transient Ischemic Attack And Stroke In Paroxysmal Atrial Fibrillation

Left atrial (LA) remodeling is closely related to cerebral stroke, but the relationship between impaired deformability of LA in early stages and stroke/TIA is not clear. The aim of this study was to evaluate the changes of LA deformability and its relationship with stroke/TIA events by using Speckle Tracking echocardiography. In 365 patients with paroxysmal atrial brillation (AF) (318, Non stroke/TIA; 47, stroke/TIA), comprehensive echocardiography was performed by using speckle tracking imaging to calculate mean LA longitudinal strain and strain rate values from apical four chamber view, apical two chamber view and apical three cavity view. The patients in stroke/TIA group had greater ages, a greater proportion of men and lower LA strain rate during left ventricular (LV) early diastole (SRE), and the difference was statistically signicant(p<0.05). In the univariate linear regression analysis, the following clinical and conventional echocardiographic parameters each had a signicant linear correlation with SRE(p<0.001), they were E/A ratio, LA volume index, body mass index, mean E/e′, LV ejection fraction, age, proportion of hypertension. Through a multiple linear regression analysis, the results show that there is a linear dependence between SRE and E/A ratio, LA volume index and Body mass index. The regression equation is y=-1.430-0.394X1+0.012X2+0.019X3(p<0.001) (y,SRE; X1,E/A ratio; X2,LA volume index; X3,Body mass index).In the multivariate logistic regression analyses, SRE and Sex ratio were independently risk factors stroke/TIA. (SRE, OR 2.945, 95% CI 1.092-7.943, p= 0.033; Sex, OR 0.462, 95% CI 0.230-0.930, p = 0.031)


Introduction
Stroke is a kind of serious complication in patients with atrial brillation (AF). The reason for this is that atrial brillation causes a rise in left atrial pressure, whereas long-term left atrial pressure causes progressive left atrial remodeling [1] . But more studies had con rmed that there was an obvious correlation between left atrial(LA) function changes and LA remodeling [2,3] .Because LA remodeling could lead to reduce atrial wall contraction force, prolong the time which blood ow through LA; thereby it would cause atrial wall brosis, atrial wall mechanical and electrical conduction abnormalities, and ultimately lead to LA enlargement, blood stasis and thrombosis formation [4,5] .
It was proved to be effective to assess LA remodeling and function by using LA longitudinal strain, the effect is better than that of conventional ultrasound echocardiography parameters such as LA volume index; and the former could more accurately evaluate LA reservoir function, conduit function and pump function. And as a non-invasive examination, its operation was simple and had better repeatability for an experienced echocardiographic operator, so it had a good application prospects [6,7] .
The present study was designed to measure LA conventional echocardiography and LA strain and strain rate parameters in patients with paroxysmal atrial brillation, then to analyze the difference between stroke/ transient ischemic attack (TIA) and non-stroke/TIA patients, and to explore relationship between the impaired deformability of LA and stroke/TIA events and its predictive value for stroke/TIA events.

Study Population
Of the 385 patients 365(94.8%) were con rmed with high quality speckle tracking LA strain imaging. The study enrolled 365 consecutive patients (222 were male, 58.95±10.52 years old) who underwent preprocedural transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) before radiofrequency catheter ablation (RFCA) for symptomatic drug-refractory paroxysmal AF in Beijing Anzhen Hospital Atrial brillation center, Capital Medical University from January 2017 to June 2020. 47 patients of them were found to have had a history of ischemic stroke or TIA. The exclusion criteria were as follows: (1) persistent or permanent AF; (2) AF with rheumatic valvular disease; (3) previous history of valve repair or replacement; (4) Echocardiography measurement of left ventricular (LV) EF less than 50%; (5) prior AF ablation; (6) any mitral valve disease, including mild-degree disease. All patients provided proper written informed consent.
We divided the patients into two groups according to the history of stroke/TIA in each patient to evaluate the difference of echocardiography data especially for LA strain and strain rate and try to lter out the best predictive parameters. Ischemic stroke was con rmed by a focal neurological de cit of sudden onset and magnetic resonance imaging or Computerized Tomography ndings and TIA was de ned as suddenonset focal neurologic symptoms or signs that resolved within 24 hours.
The study was approved by The Ethical Committee of Beijing Anzhen Hospital, Capital Medical University.

Conventional Transthoracic Echocardiography (TTE)
All subjects underwent comprehensive TTE study using a Vivid 7 or Vivid E9 cardiovascular ultrasound system (GE Medical Systems, Horten, Norway), equipped with a S51 phased-array sector probe (2.5 to 3.5MHz). Standard M-mode, 2-dimensional, and Doppler images were acquired in the parasternal and apical views. Left ventricular (LV) ejection fraction was calculated using the biplane Simpson's rule. The LV diastolic function was assessed by mean E/e′, E/A and LA volume index. Tissue Doppler imaging was performed to measure myocardial velocities. Pulsed sample volume was placed at both septal corner and lateral wall of mitral annulus; early diastolic (e′) and late diastolic (a′) myocardial velocities were recorded. The mean E/e′ ratio was then calculated. All measurements were recorded as average of three cardiac cycles according to the standards of American Society of Echocardiography guidelines [8,9] .

Speckle Tracking Echocardiography
Each patient was collected LA velocity vector imaging data, the standard gathering sections were apical four chamber view, apical two chamber view and apical three cavity view, each section was set to acquire three cardiac cycles and frame rate is adjusted to more than 60 frames per second. The analysis was performed offline using workstation software (EchoPAC PC; GE Medical Systems). LA strain was set to zero at the beginning of the P wave (P-triggered analysis). After manual tracing of the endocardial LA borders, an epicardial surface tracing was automatically generated by the system, creating a region of interest, which was automatically determined, and speckles were traced frame by frame. The region of interest was divided the into 6 segments and the averages of the values and curves of strain and strain rate in the longitudinal direction for each segment were generated automatically. The LA roof segments in each view were excluded in this study due to the discontinuity of LA wall because of the connection to the pulmonary veins. Segments with suboptimal image were rejected by the software and excluded from the analysis. By calculating the average strain and strain rate of left atrium in the four-chamber, two-chamber and three-chamber view, we got the mean longitudinal strain and mean longitudinal strain rate of LA.
The LA strain versus time curves were measured respectively, as the positive peak strain during LV systole is LA strain during LV systole (SS) and negative peak strain during LV diastole is LA strain during LV diastole (SD), the distance between the two peaks is the global longitudinal strain (GLS) of LA ( Figure 1); and for the LA strain rate versus time curves, the positive peak in the middle during LV systole indicates LA strain rate during LV systole (SRS), the negative peak on the right side indicates LA strain rate during LV early diastole (SRE) and the another negative peak on the left side indicates LA strain rate during LV late diastole (SRL) ( Figure 2).

Inter-and Intraobserver Variability
To detect the variation between the observer and the observer, we randomly selected 15 patients to measure the LA strain and strain rate; rst, measurements were repeated twice by one observer, then two experienced observers who were unaware of each other's measurements and of the study time point obtained VVI data from the same patients.

Statistical Analysis
Continuous data are expressed as mean ±SD. Categorical data are summarized as frequencies and percentages. Continuous variables were compared using the Student t test (for two groups of comparisons) and ANOVA (for multiple group comparisons). Categorical variables were compared using chi-square test. Univariate and multivariate logistic regression analyses to determine risk factors of stroke/TIA in patients with paroxysmal AF. Statistical analyses were performed using SPSS version 23.0 (SPSS, Inc, Chicago, IL). All P values reported are from two-sided tests, and P values < 0.05 were considered statistically signi cant. Table 1 lists the basic clinical and echocardiographic data of all patients according to the patient's history of stroke or TIA. We found that the differences in age, sex, and SRE were statistically signi cant between the two groups (p 0.05). Compared with the none stroke/TIA group, the patients in stroke/TIA group had greater ages, a greater proportion of men and lower SRE. For other clinical descriptive parameters and conventional two-dimensional echocardiographic parameters, there is no statistical difference between the two groups.

Association Between LA Strain Rate and Clinical Features, Cardiac Structure Deformability and Function Change
Through a univariate linear regression analysis, we found the following clinical and conventional echocardiographic parameters each had a signi cant linear correlation with SRE (p<0.001), these parameters include E/A ratio, LA volume index, body mass index, mean E/e′, LV ejection fraction, age, hypertension. After a parameters optimization, we performed multiple linear regression analysis, the results show that there is a linear dependence between SRE and multiple parameters such as E/A ratio, LA volume index and Body mass index. The regression equation is y=-3.487-0.83X1+0.024X2+0.05X3+0.064X4+0.028X5 (y, SRE; X1, E/A ratio; X2, LA volume index; X3, Body mass index; X4, Mean E/e′; X5, Age). The results of the regression analysis are shown in Table 2.
The predictive value of conventional ultrasound parameters and strain parameters for stroke events The results showed that LA volume index, mean LA systolic strain and mean LA SRE were highly sensitive to stroke/TIA events; and LV end-diastolic diameter, Mean E/e′ showed good speci city. However, on the whole, the AUC of these indicators is not good in this study (Table-3). The probable cause is the relatively small sample size of the stroke group, and although we strictly adhere to inclusion criteria, there is a greater likelihood of bias.
The value of various parameters for risk factors of Stroke/TIA Table 3 presents the logistic regression analyses for clinical and echocardiographic parameters associated with the presence of stroke/TIA. In the univariate logistic regression analyses, the presence of stroke/TIA was associated with old age (p < 0.05), more female gender (p < 0.05) and greater LA SRE (the absolute value was close to 0) (p < 0.001). In the multivariate logistic regression analyses, LA SRE and Sex ratio were independently risk factors stroke/TIA. (SRE, OR 2.945, 95% CI 1.092-7.943, p = 0.033; Sex, OR 0.462, 95% CI 0.230-0.930, p = 0.031).

Discussion
In his study we mainly analyzed the changes of LA strain and strain rate in patients with paroxysmal atrial brillation by using speckle tracking technique and explored its value in stroke/TIA risk prediction.
The main conclusions are: (1) the LA SRE change occurred earlier in patients with paroxysmal AF complicated stroke/TIA events; (2) the LA SRE and the clinical and conventional ultrasound echocardiography parameters signi cantly related; (3) LA SRE is an independent risk factor for stroke in paroxysmal atrial brillation patients.
LA Reservoir Function, Conduit Function, Pump Function and LA Deformability LA function in the whole cardiac cycle was divided into three phase, storing blood in LV systolic phase, most blood via LA into the LV in early diastolic phase and LA contraction pumping the remaining blood to LV in late LV diastolic phase [6,10] . Therefore, in the LA strain rate curve we recorded three different peaks: one positive peak in LV systolic phase and two negative peaks in LV early and late diastolic phase, respectively corresponding to LA reservoir function, conduit function and pump function. As we all know, with respect to strain, theoretically strain rate could re ect the earlier damage of LA deformability due to a combination of time change, which had a higher diagnostic value. First we should determine the L 0 of LA in longitudinal axis when analyzing LA strain rate, similar to LV strain and strain rate analysis, we believe that LA was in L 0 state at the beginning of LA contraction, when ECG is located on the P wave phase [11] .
Compared with the non-stroke group, LA volume index in the stroke/TIA group relative increased, but the difference was not statistically signi cant, but the LA SRE between two groups appeared obvious difference(p 0.05), which indicating there were no obvious morphological changes in LA, LA Deformability function decreased in the early stages. and LA longitudinal strain did not signi cantly change too.

Association Between LA SRE and Clinical, Conventional Echocardiographic Parameters
Previous studies have con rmed the presence of signi cant correlation between LA strain with clinical, conventional echocardiographic parameters [5,[12][13][14][15][16] . Our study also con rmed this, LA SRE had signi cant linear correlation with age, BMI, hypertension, LV EF, mean E/e′, LA volume index, E/A ratio respectively (P<0.001), especially showed the best correlation with age. LA SRE not only re ected LA remodeling, but more importantly, it re ected early damage of LA deformability function. It could become a potential ideal indicator for LA function changes in early stages [12,[17][18][19] .

LA Deformability Parameters and Risk Factors for Stroke/TIA
LA remodeling parameters such as LA volume index has been con rmed to be a risk factor for cardiovascular events, such as stroke, myocardial infarction, atrial thrombosis, recurrence of atrial brillation, etc [20][21][22] . Because the increased LA pressure and increased volume load would lead to the gradual brosis of LA wall, thus making LA reservoir, conduit dysfunction. But before the enlargement of LA and LA deformability capacity was likely to have been damaged, if at this stage we could nd an ideal ultrasound parameter for accurate evaluating LA deformability, so as to make an earlier prediction for the above mentioned cardiovascular events.
Previous studies had con rmed that it was feasible to evaluate LA remodeling and deformability using two-dimensional echocardiographic speckle tracking technology [5,7,23,24] . Small sample studies showed LA longitudinal strain was a risk factors of stroke/TIA in patients with AF [6,13] . In our study it was conformed that LA longitudinal strain rate had changed early compared to LA strain and re ect the damages of deformability of LA earlier.
In this study, we found that LA SRE was indeed an independent risk factor (p<0.05) in patients with paroxysmal atrial brillation incidence of stroke by multivariate Logistic regression analysis, the smaller the absolute value of SRE, the greater the risk of stroke/TIA. However, to gure out the speci c relationship between the index and stroke/TIA, we still need a long-term follow-up combining with clinical and conventional echocardiographic parameters for further analysis [25] .

Clinical Application
LA strain rate changes re ect the ability of the LA deformability, it changed early than LA remodeling [26][27][28] , and was more accurate than LA strain, so it was more suitable for screening of patients at an early stage. Because of the noninvasive and easy operation, the speckle tracking technique is becoming more and more mature, we hoped that it would be proved to be a reliable parameter in subsequent studies.

Limitation
First, this was a cross-sectional study, which had not been followed up to evaluate the effectiveness of the risk factors. Second, in this study, we aimed to study the patients with paroxysmal AF, and in further study we need to incorporate non atrial brillation patients with or without stroke/TIA, then compare the changes of LA strain rate. Third, our patients all had preserved LV ejection fractions.

Conclusions
LA SRE re ects the deformability changes of LA in early stages. It was an independent risk factor for stroke/TIA in patients with paroxysmal AF, and further studies are needed to include more control and long-term follow-up and evaluate the clinical utility of LA strain rate change in prospective studies.