Paroxysmal atrial fibrillation (AF) frequently, but not always, progresses to persistent/permanent AF. The aim of this study was to evaluate the echocardiographic predictors of AF progression in patients with paroxysmal AF.
A multicenter, prospective, observational study was conducted that included 313 patients with paroxysmal AF who underwent two-dimensional speckle-tracking echocardiography. The diameter, volume, and mechanical function of the left atrium, including global strain (ε) and ε rate, were measured.
Progression to persistent or permanent AF occurred in 52 patients (16.6%) during a median follow-up period of 26 months. Echocardiographic measure of left atrial (LA) diameter, volume, and function (E velocity, E/A and E/e′ ratio, LA expansion index, active emptying fraction, global longitudinal ε and ε rate) were associated with AF progression. LA ε ≤ 30.9% was the strongest predictor of AF progression, which was associated with a more than fourfold hazard increase for AF progression (hazard ratio, 4.224; P = .001). LA diameter > 39 mm and maximal LA volume index > 34.2 mL/m 2 were associated with about a twofold hazard increase for AF progression (hazard ratios, 1.994 and 2.649; P = .016 and P = .001, respectively). When adjusted for a model combining maximal LA volume index, E velocity, LA expansion index, and active emptying fraction, LA ε ≤ 30.9% maintained a more than threefold hazard increase for AF progression (adjusted hazard ratio, 3.970; P = .003).
Echocardiographic measures of LA diameter, volume, and mechanical function, including LA ε, were associated with AF progression. LA ε was the strongest independent predictor of AF progression and is expected to serve as a valuable predictor of AF progression.
Atrial fibrillation (AF) is the most common arrhythmia and is associated with an increased risk for stroke, heart failure, and premature death. AF is classified as paroxysmal, persistent, or permanent, depending on the duration of symptoms and its propensity to terminate by itself, terminate with electrical or pharmacologic intervention, or not terminate at all. Progression of paroxysmal AF to persistent or permanent AF is frequently seen and is associated with increased morbidity and mortality. However, not all patients with paroxysmal AF progress to persistent or permanent AF. A better understanding of AF progression is required to implement effective preventive strategies and to reduce the burden of AF. De Vos et al . demonstrated that heart failure, stroke, chronic obstructive pulmonary disease, hypertension, and age were independent predictors of AF progression.
Atrial remodeling is recognized as a key feature in the genesis and perpetuation of AF. There has been increasing interest in the noninvasive evaluation of left atrial (LA) size and mechanical function using two-dimensional (2D) echocardiography and Doppler analysis of transmitral flow and tissue Doppler measurement of LA myocardial velocities. Recently, strain (ε) analysis with 2D speckle-tracking echocardiography has been applied to the left atrium to quantify the magnitude of atrial deformation. A previous study reported that LA global longitudinal ε is decreased in patients with paroxysmal AF compared with normal control subjects. And decreased LA ε is associated with AF recurrence after catheter ablation. However, to date, little is known about the role of echocardiographic evaluation of LA remodeling for the prediction of AF progression. Although a few studies have demonstrated that increased LA dimension is associated with AF progression, the role of echocardiographic measures of LA volume and function, including ε, has not been evaluated. We attempted to identify the echocardiographic predictors of progression to persistent or permanent AF in patients with paroxysmal AF and to evaluate whether LA ε measurement can provide additive information for the prediction of AF progression.
The Echocardiographic Predictors of Progression to Persistent or Permanent Atrial Fibrillation in Patients with Paroxysmal Atrial Fibrillation (E6P) study was designed as a multicenter, prospective, observational study and was conducted at four sites in Korea (Seoul National University Bundang Hospital, Seongnam; Hallym University Medical Center, Anyang; Korea University Medical Center, Seoul; and Chungnam University Hospital, Daejeon). The study protocol was approved by the institutional review board of each center.
Patients with paroxysmal AF who underwent echocardiography for the evaluation of cardiac structure and function between 2006 and 2011 were assessed for eligibility. The inclusion criteria were (1) documented first AF on electrocardiography or 24-h electrocardiographic monitoring in the previous 2 months and (2) sinus rhythm at the time of echocardiography. Patients were excluded from the study if any of the following was present: (1) history of AF, (2) restoration of sinus rhythm after pharmacologic or electrical cardioversion, (3) dyspnea ≥ New York Heart Association functional class II; (4) left ventricular (LV) ejection fraction < 50%; (5) valvular stenosis or regurgitation ≥ moderate; (6) history of ischemic heart disease; (7) history of transient ischemic attack or ischemic stroke; and (8) disease that can predispose to AF, including hyperthyroidism, alcoholism, postoperative status, and septic shock. A total of 363 patients with paroxysmal AF were enrolled.
At the time of enrollment, clinical information was obtained from hospital records. We determined the CHADS 2 and CHA 2 DS 2 -VASc scores, which allow instant classification of the relative thromboembolic risk in patients with AF. We also determined the HATCH score, which was previously proposed to assess the clinical risk for AF progression, calculated by assigning 1 point each for hypertension, age ≥ 75 years, and chronic obstructive pulmonary disease and 2 points each for transient ischemic attack or stroke and heart failure.
A Vivid 7 ultrasound system (GE Vingmed Ultrasound AS, Horten, Norway) was used for the transthoracic echocardiographic examinations. All images and measurements were acquired from the standard views, according to the guidelines of the American Society of Echocardiography, and were digitally stored for offline analysis. All echocardiograms were analyzed in blinded fashion by an independent core laboratory (Seoul National University Bundang Hospital). In the parasternal long-axis views, the maximal LA anterior-posterior (A-P) diameter was measured. LA volumes were measured using a biplane area-length method from the apical four- and two-chamber views and were indexed to body surface area: maximal LA volume index (before mitral valve opening) (LAVI max ), pre-A LA volume index (before atrial contraction) (LAVI pre-A ), and minimal LA volume index (after atrial contraction, LAVI min ). LA filling volume index and active emptying volume index were calculated as LAVI max − LAVI min and LAVI pre-A − LAVI min , respectively. LA expansion index (%) and active emptying fraction (%) were calculated as (LA filling volume index/LAVI min ) × 100% and (LA active emptying volume index/LAVI pre-A ) × 100%, respectively.
Global LA myocardial longitudinal ε during ventricular systole was measured by 2D speckle-tracking echocardiography as previously described. Grayscale images of apical four-chamber views were obtained with frame rates of 50 to 80 Hz. Recordings were processed with speckle-tracking software (EchoPAC; GE Vingmed Ultrasound AS), allowing offline semiautomated speckle-based ε analyses. Briefly, the lines were manually traced along the LA endocardium at the time of the end-systolic phase. An additional epicardial line was automatically generated by the software, which created a region of interest. After manually adjusting the shape of the region of interest, LA ε and ε rate during the whole cardiac cycle were calculated. In the present study, LA ε was simply measured from the apical four-chamber view. However, we also calculated the average mean value for LA ε obtained from the four- and two-chamber views (ε avg ) for the comparison with ε from the apical four-chamber view only. Usually, measurement of LA ε from the apical two-chamber view is thought to be challenging because of the LA appendage. In fact, we could not reliably measure ε from the apical two-chamber view in three patients with large LA appendages.
The primary end point of the study was assessment of AF progression. Patents were followed for ≥1 year: 12-lead electrocardiography was repeated at every 6-month visit, and 24-h electrocardiographic monitoring was repeated once annually. Of the 363 patients, 50 (13.8%) who were not followed for ≥1 year were excluded. Progression of AF was defined as paroxysmal AF at baseline that had become persistent or permanent. Persistent AF was defined as recurrent AF that was not self-terminating with episode durations of >7 days, and permanent AF was defined as AF that had been present for ≥6 months without intervening spontaneous episodes of sinus rhythm for which cardioversion was unsuccessful and subsequently not attempted.
Intraobserver and interobserver variability for LA volume index and ε were assessed in 17 randomly selected subjects. For intraobserver variability, data sets were analyzed again 1 month after performing the initial measurements. Interobserver variability was determined by analyzing data from two separate observers. The mean bias, the lower and upper limits of agreement, and the coefficient of variation were calculated. Reproducibility was assessed using Bland-Altman analysis and the intraclass correlation coefficient.
Continuous variables are expressed as mean ± SD and categorical variables as numbers and percentages. Continuous variables were compared using the Student t test or the Mann-Whitney test. Categorical variables were compared using the χ 2 test or the Fisher exact test. Cox regression analysis was performed to evaluate the association between clinical and echocardiographic variables and AF progression. Results were reported as hazard ratios (HRs) with 95% CIs. The HR defines the relative likelihood of an event per unit change, and the magnitude of change in the HR depends on the unit used in the analysis. Accordingly, we describe HRs with the units that were used to express changes. Receiver operating characteristic (ROC) curve analysis was performed to determine the predictive accuracy and optimal cutoff values of echocardiographic parameters which demonstrated significant association with AF progression. Areas under the ROC curves (AUCs) were compared using the DeLong method. Then, we fitted Cox proportional hazards models again to estimate the unadjusted and adjusted HRs of echocardiographic variables as categorical variables by the optimal cutoff values. Kaplan-Meier curves were used to estimate the survival function of time to events according to the echocardiographic parameters by the optimal cutoff value. Differences between time-to-event curves were compared using the log-rank test. Annual event rates were calculated by dividing the 4-year Kaplan-Meier event rate by 4. Statistical significance was determined by a P value ≤ .05. All analyses were performed using SPSS version 20.0 (IBM, Chicago, IL).
A total of 313 patients (194 men; mean age, 57 ± 14 years) were included. During follow-up (median, 26 months; interquartile range, 16–37 months), progression to either persistent or permanent AF (28 and 24, respectively) was observed in 52 patients (16.6%), resulting in an annualized AF progression rate of 5.7% ( Figure 1 ). Clinical characteristics, including the CHADS 2 , CHA 2 DS 2 -VASc, and HATCH scores, were similar between patients with and without AF progression ( Table 1 ). However, patients with AF progression used oral anticoagulation more frequently; patients who used oral anticoagulation showed higher CHADS 2 scores (1.1 ± 0.8 vs 0.6 ± 0.8, P < .001), CHA 2 DS 2 -VASc scores (2.0 ± 1.3 vs 1.3 ± 1.2, P < .001), and LAVI max (39.5 ± 12.7 vs 32.7 ± 11.5 mL/m 2 , P < .001). There were no significant differences in LV volume, mass index, and ejection fraction between patients with and without AF progression. However, patients with AF progression were found to have increased LA A-P diameters, LAVI min , LAVI pre-A , LAVI max , E velocities, and E/A and E/e′ ratios and to have decreased LA expansion index values, active emptying fractions, ε, and ε rates compared with those without progression.
|Variable||All patients||Patients without AF progression||Patients with AF progression||P ∗|
|( n = 313)||( n = 261)||( n = 52)|
|Age (y)||57 ± 14||57 ± 14||59 ± 11||.265|
|Men||194 (62%)||159 (61%)||35 (67%)||.386|
|Hypertension||149 (48%)||123 (47%)||26 (50%)||.705|
|Diabetes||45 (14%)||38 (15%)||7 (13%)||.837|
|BMI (kg/m 2 )||24 ± 3||24 ± 3||24 ± 3||.593|
|CHADS 2 score||0.7 ± 0.8||0.7 ± 0.8||0.7 ± 0.7||.774|
|CHA 2 DS 2 -VASc score||1.5 ± 1.3||1.5 ± 1.3||1.4 ± 1.2||.630|
|HATCH score||0.6 ± 0.7||0.6 ± 0.7||0.6 ± 0.6||.806|
|Any antiarrhythmic drug||116 (37%)||95 (36%)||21 (40%)||.531|
|Angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker||110 (35%)||94 (36%)||16 (31%)||.469|
|β-blocker||136 (43%)||108 (41%)||28 (54%)||.098|
|Statin||91 (29%)||74 (28%)||17 (33%)||.509|
|Antiplatelet agent||200 (64%)||168 (64%)||32 (62%)||.698|
|Oral anticoagulant||70 (22%)||46 (18%)||24 (46%)||<.001|
|LV EDV (mL)||74.9 ± 20.4||74.3 ± 19.9||78.3 ± 22.6||.200|
|LV ESV (mL)||28.5 ± 9.4||28.1 ± 8.9||30.5 ± 11.5||.082|
|LV SV (mL)||46.5 ± 13.4||46.2 ± 13.2||47.7 ± 14.1||.464|
|LV ejection fraction (%)||62.8 ± 6.0||62.9 ± 5.7||62.1 ± 7.0||.385|
|LV mass index (mL/m 2 )||92.2 ± 22.7||91.6 ± 22.5||95.5 ± 23.6||.260|
|LA A-P diameter (mm)||39 ± 6||39 ± 06||41 ± 6||.032|
|LAVI min (mL/m 2 )||17.2 ± 9.0||16.1 ± 7.6||22.3 ± 12.8||.001|
|LAVI pre-A (mL/m 2 )||24.0 ± 10.2||23.0 ± 9.0||29.0 ± 13.8||.004|
|LAVI max (mL/m 2 )||34.2 ± 12.1||33.0 ± 11.1||40.4 ± 14.6||<.001|
|E (cm/sec)||66.8 ± 16.4||65.5 ± 15.6||73.6 ± 18.3||.001|
|A (cm/sec)||64.2 ± 20.7||64.5 ± 21.1||62.5 ± 19.0||.528|
|E/A ratio||1.1 ± 0.5||1.1 ± 0.5||1.3 ± 0.5||.045|
|Deceleration time (msec)||209 ± 52||210 ± 52||201 ± 48||.252|
|e′ (cm/sec)||7.6 ± 2.5||7.6 ± 2.6||7.5 ± 2.0||.659|
|a′ (cm/sec)||8.3 ± 2.3||8.4 ± 2.2||7.8 ± 2.3||.079|
|s′ (cm/sec)||7.5 ± 1.9||7.6 ± 1.9||7.1 ± 1.4||.059|
|E/e′ ratio||9.5 ± 3.1||9.3 ± 3.0||10.5 ± 3.9||.011|
|LA filling volume index (mL/m 2 )||17.0 ± 5.9||16.8 ± 5.9||18.1 ± 5.8||.158|
|LA expansion index (%)||118 ± 57||122 ± 58||101 ± 48||.011|
|LA active emptying volume index (mL/m 2 )||6.8 ± 3.6||6.3 ± 3.6||6.7 ± 3.7||.764|
|LA active emptying fraction (%)||29.7 ± 13.0||30.6 ± 12.9||24.8 ± 12.4||.003|
|LA ε (%)||27.8 ± 9.2||28.4 ± 9.3||24.6 ± 8.0||.006|
|LA ε rate (sec −1 )||1.2 ± 0.4||1.2 ± 0.4||1.1 ± 0.4||.049|
The results of univariate analyses for AF progression are summarized in Table 2 . The clinical parameters, including the CHADS 2 , CHA 2 DS 2 -VASc, and HATCH scores, were not associated with AF progression. LV volume, mass, and ejection fraction were also not associated with AF progression. However, LA A-P diameter, LAVI min , LAVI pre-A , and LAVI max were significantly associated with AF progression: a 10-mm increase in LA A-P diameter and a 10 mL/m 2 increase in LAVI max were associated with 67% and 50% increases in AF progression risk, respectively. E velocity, E/A ratio, and E/e′ ratio were also associated with increased hazard for AF progression, whereas A or a′ velocity was not: a 10 cm/sec increase in E velocity was associated with a 29% increase in AF progression risk. LA expansion index, active emptying fraction, ε, and ε rate were also significant predictors of AF progression: a 10% increase in LA ε was associated with a 39% decrease in AF progression risk.
|Variable||Hazard ratio (95% CI)||P|
|Age, per 10 y||1.142 (0.927–1.405)||.211|
|Male gender||0.794 (0.445–1.417)||.435|
|BMI, per 5 kg/m 2||1.120 (0.744–1.687)||.587|
|CHADS 2 score||0.949 (0.677–1.331)||.762|
|CHA 2 DS 2 -VASc score||0.962 (0.775–1.194)||.724|
|HATCH score||0.962 (0.630–1.468)||.857|
|LV EDV, per 10 mL||1.100 (0.966–1.253)||.149|
|LV ESV, per 10 mL||1.258 (0.980–1.616)||.072|
|LV SV, per 10 mL||1.100 (0.898–1.349)||.357|
|LV ejection fraction, per 10%||0.797 (0.501–1.266)||.336|
|LV mass index, per 10 mg/m 2||1.050 (0.937–1.177)||.401|
|LA A-P diameter, per 10 mm||1.674 (1.083–2.588)||.020|
|LAVI min , per 10 mL/m 2||1.664 (1.355–2.044)||<.001|
|LAVI pre-A , per 10 mL/m 2||1.532 (1.249–1.879)||<.001|
|LAVI max , per 10 mL/m 2||1.498 (1.240–1.809)||<.001|
|E, per 10 cm/sec||1.288 (1.110–1.493)||.001|
|A, per 10 cm/sec||0.958 (0.833–1.101)||.545|
|E/A ratio||1.602 (1.017–2.525)||.042|
|Deceleration time, per 10 msec||0.966 (0.911–1.024)||.244|
|E/e′ ratio||1.114 (1.027–1.207)||.009|
|LA filling volume index||1.032 (0.988–1.078)||.162|
|LA expansion index||0.993 (0.987–0.998)||.012|
|LA active emptying volume index||0.792 (0.918–1.067)||.792|
|LA active emptying fraction||0.970 (0.950–0.990)||.004|
|LA ε, per 10%||0.612 (0.433–0.863)||.005|
|LA ε rate||0.467 (0.219–0.995)||.048|
The results of ROC curve analysis of echocardiographic parameters that demonstrated significant associations with AF progression are presented in Table 3 . The AUC for LA A-P diameter was 0.59 (95% CI, 0.53–0.64; P = .039), and the optimal cutoff value was >39 mm. The AUC for LAVI max was 0.66 (95% CI, 0.60–0.71; P < .001), which was not significantly different from those for LAVI min (AUC, 0.66; 95% CI, 0.60–0.71; P < .001; difference, 0.00; P = .993) and LAVI pre-A (AUC, 0.62; 95% CI, 0.57–0.68; P = .003; difference, 0.03; P = .082). The optimal cutoff value of LAVI max was >34.2 mL/m 2 . The AUC for LA ε was 0.64 (95% CI, 0.58–0.69; P < .001), with an optimal cutoff value of ≤30.9%. The AUC for LA ε avg was not significantly different from that for LA ε (AUC, 0.65; 95% CI, 0.59–0.70; difference, 0.01; P = .628).
|Variable||Optimal cutoff value||AUC (95% CI)||P|
|LA A-P diameter (mm)||>39||0.59 (0.53–0.64)||.039|
|LAVI min (mL/m 2 )||>16.9||0.66 (0.60–0.71)||<.001|
|LAVI pre-A (mL/m 2 )||>19.3||0.62 (0.57–0.68)||.003|
|LAVI max (mL/m 2 )||>34.2||0.66 (0.60–0.71)||<.001|
|E (cm/sec)||>62||0.63 (0.57–0.68)||.002|
|E/A ratio||>0.87||0.60 (0.55–0.66)||.012|
|E/e′ ratio||>7.8||0.59 (0.53–0.64)||.039|
|LA expansion index (%)||≤106||0.61 (0.55–0.66)||.012|
|LA active emptying fraction (%)||≤26||0.64 (0.58–0.69)||.001|
|LA ε (%)||≤30.9||0.64 (0.58–0.69)||<.001|
|LA ε rate (sec −1 )||≤0.9||0.58 (0.53–0.64)||.055|