Background
It is traditionally difficult to estimate left ventricular (LV) systolic function in atrial fibrillation (AF). The aim of this study was to validate the use of an index beat, the beat after the nearly equal preceding (RR1) and pre-preceding (RR2) intervals, for the measurement of LV peak longitudinal systolic strain (PLSS). The difference between RR1 and RR2 intervals of the index beat must be <60 msec. LV PLSS measured from the index beat (PLSSindex) was compared with LV PLSS measured from the conventional but time-consuming method of averaging multiple cardiac cycles (PLSSavg).
Methods
Ninety-eight patients with persistent or permanent AF and resting ventricular rates ≤ 105 beats/min were prospectively included. LV PLSSindex and LV PLSSavg were obtained from two-dimensional speckle-tracking echocardiography.
Results
LV PLSSindex had a highly significant correlation with LV PLSSavg ( r = 0.970, P < .001). Bland-Altman analysis showed only small bias of 0.01%, and the 95% limits of agreement were +1.64% to −1.62%. Compared with those with lower risk scores of stroke indicated by CHADS 2 scores < 2 or CHA 2 DS 2 -VASc scores < 2, patients with higher risk scores of stroke indicated by CHADS 2 scores ≥ 2 or CHA 2 DS 2 -VASc scores ≥ 2 had lower PLSSavg and PLSSindex ( P ≤ .012).
Conclusions
LV PLSSindex was a good alternative to LV PLSSavg in patients with AF. Use of the index beat to measure LV longitudinal systolic strain in patients with AF was as accurate as the time-consuming method of averaging multiple cardiac cycles.
Because of beat-to-beat variation, it has traditionally been difficult to estimate left ventricular (LV) systolic function in atrial fibrillation (AF). Increased cardiovascular morbidity and mortality have been reported in patients with AF, and assessment of LV systolic function may be important in these patients. To assess LV systolic function in AF accurately and reproducibly, the widely accepted method is the average of measurements obtained from multiple consecutive cardiac cycles. The optimal number of beats required to evaluate LV systolic function in a reproducible manner in AF has been suggested to be ≥13 beats for <5% variability in the measured values, or approximately three times that necessary in sinus rhythm. Such measurement is not only time consuming but also dependent on the window of cardiac cycles selected. To overcome this tedious method of averaging multiple cardiac cycles of LV systolic function, an index beat, the beat after the nearly equal preceding (RR1) and pre-preceding (RR2) intervals, has been used to replace the averaging technique in evaluation of LV systolic function in patients with AF.
In addition, AF is a common arrhythmia that represents an independent risk factor for stroke. The CHADS 2 score is reported to be able to predict risk for stroke, vascular events, and heart failure. Recently, a newly developed scoring system, the CHA 2 DS 2 -VASc score, which extends the CHADS 2 score by considering additional stroke risk factors, has been reported to provide some improvement in predictive value for thromboembolism over the CHADS 2 scheme.
Two-dimensional speckle-tracking is an echocardiographic technique that has been used increasingly to analyze LV systolic and diastolic function and assess ischemia, dyssynchrony, left atrial function, and other cardiac conditions. However, validation of using an index beat in the assessment of LV systolic strain function in patients with AF has not been performed. The primary aim of this study was to evaluate whether LV longitudinal systolic strain measured from an index beat could represent that measured from the average of measurements taken over multiple cardiac cycles, in the hope of providing a simple and accurate alternative in the assessment of LV longitudinal systolic strain function in patients with AF. The second aim of this study was to compare the echocardiographic parameters in study patients according to the CHADS 2 and CHA 2 DS 2 -VASc scores.
Methods
Patients
Ninety-eight patients with persistent or permanent AF referred for echocardiographic examinations were prospectively included. Patients with fewer than 5 segments visualized in the apical four-chamber view ( n = 6), significant mitral stenosis ( n = 4), resting ventricular rates > 105 beats/min during AF ( n = 8), and no beat fulfilling the requirements of the index beat in the stored image ( n = 3) were excluded. Patients were further classified according to CHADS 2 and CHA 2 DS 2 -VASc scores. The CHADS 2 score was calculated for each patient on the basis of a point system in which 2 points were assigned for a history of stroke or transient ischemic attack, and 1 point each was assigned for age ≥ 75 years and the presence of congestive heart failure, hypertension, and diabetes. The CHA 2 DS 2 -VASc score was calculated for each patient on the basis of a point system in which 2 points each were assigned for a history of stroke or transient ischemic attack and age ≥ 75 years, and 1 point each was assigned for age between 65 and 74 years, the presence of congestive heart failure, hypertension, diabetes, vascular disease (myocardial infarction, aortic plaque, or peripheral artery disease), and female sex. The protocol was approved by our institutional review board, and all enrolled patients gave written informed consent.
Echocardiography
The echocardiographic examination was performed using a Vivid 7 (GE Vingmed Ultrasound AS, Horten, Norway), with the participant respiring quietly in the left decubitus position. Two-dimensional and two-dimensionally guided M-mode images were recorded from standardized views. The Doppler sample volume was placed at the tips of the mitral leaflets to obtain the LV inflow waveforms from the apical four-chamber view. All sample volumes were positioned with the ultrasound beam aligned to flow. Transmitral E-wave velocity (E) and E-wave deceleration time were obtained from five beats, and the data were averaged to give the mean value for later analysis. Pulsed Doppler tissue imaging was performed with the sample volume placed at the lateral and septal corners of the mitral annulus from the apical four-chamber view, and early diastolic mitral annular velocity (Ea) was measured from five cardiac cycles, and the data of lateral and septal Ea were averaged to give the mean value for later analysis. LV ejection fraction (LVEF) was measured using the modified Simpson’s method from the index beat. Left atrial volume was measured using the biplane area-length method from the index beat. Left atrial volume index was calculated by dividing left atrial volume by body surface area.
We acquired LV apical four-chamber views using a high frame rate (50–90 frames/sec), and 15 consecutive cardiac cycles were obtained and digitally stored for offline analysis. Image analysis was performed offline using custom analysis software (EchoPAC version 08; GE-Vingmed Ultrasound AS) by a cardiologist blinded to the other data. The endocardial border was manually defined using a point-and-click technique. An epicardial surface tracing was automatically generated by the system, creating a region of interest, which was manually adjusted to cover the full thickness of the myocardium in the systolic frame. The width of the smallest region of interest was 8 mm. A time-strain plot was produced automatically by the software. Peak longitudinal systolic strain (PLSS) was assessed in the six LV segments in the apical four-chamber view. The values of PLSS in the six LV segments were averaged to give the mean value for later analysis. We used cine loops to determine which beat would be calculated. LV PLSS was calculated beat by beat, and then data from 15 cardiac cycles were averaged (PLSSavg) for later analysis. Simultaneous electrocardiograms were also recorded to measure the corresponding RR intervals, and the data from 15 cardiac cycles were averaged to give the mean value for later analysis.
Index Beat Selection
The index beat taken after two nearly equal preceding cardiac cycles was selected from the 15 stored cardiac cycles. The RR1 and RR2 intervals of the index beat must be >500 ms, and the difference between them must be <60 ms ( Figure 1 ). The cardiac cycle of the index beat was also >500 msec. If no beat fulfilled the requirements of the index beat in the 15 stored cardiac cycles, we excluded such patients. If several beats fulfilled the requirements of the index beat in the 15 stored cardiac cycles, we chose the beat with the smallest difference between the RR1 and RR2 intervals.
To determine whether the degree of RR match influenced how well the index beat works, we used differences between RR1 and RR2 of 0 to 39, 40 to 70, and 71 to 100 msec to categorize the difference between RR intervals as an excellent match, a good match, and a modest match. LV PLSS was measured from different categories of RR match. The average values in any categories of RR match were used for later analysis.
Reproducibility
Fifteen patients were randomly selected for evaluation of the interobserver variability of PLSS measurement by two independent observers. To obtain intraobserver variability, the same measurements were repeated 1 week apart. Mean percentage error was calculated as the absolute difference divided by the average of the two observations.
Statistical Analysis
All data are expressed as mean ± SD. SPSS version 18.0 (SPSS, Inc., Chicago, IL) was used for statistical analysis. Differences between groups were checked using χ 2 tests for categorical variables and independent t tests for continuous variables. Differences between LV PLSSavg and the values of LV PLSS measured from excellent match, good match, and modest match of RR intervals were compared using paired t tests. The relationship between two continuous variables was assessed using a bivariate correlation method (Pearson’s correlation).
For comparison of agreement in measurements by the two different methods (multiple cardiac cycles vs index beat), Bland-Altman analysis was applied to evaluate systematic differences and limits of agreement between them. Linear regression analysis was used to assess the relationship between the index beat and the average of multiple cardiac cycles. All tests were two sided, and the level of significance was established as P < .05.
Results
The baseline and echocardiographic characteristics of 98 study patients are shown in Table 1 . There were 37 patients (38%) with LVEFs < 50%. There were eight patients (8%) with regional wall motion abnormalities. Pharmacologic therapy, including β-blockers (48%), calcium channel blockers (24%), angiotensin-converting enzyme inhibitors (9%), angiotensin II antagonists (47%), and diuretics (34%), was continued without interruption. The values of PLSSavg and PLSS measured from the index beat (PLSSindex) were −11.5 ± 3.4% and −11.4 ± 3.6%, respectively, in all patients. The PLSSindex values of basoseptal, midseptal, apicoseptal, apicolateral, midlateral, and basolateral segments in all patients were −11.3 ± 4.9%, −13.4 ± 4.7%, −14.7 ± 5.2%, −11.8 ± 5.0%, −9.3 ± 4.8%, and −8.5 ± 4.8%, respectively. The value of the RR interval was 784 ± 168 msec (range, 573−1,370 msec).
| Characteristic | All patients | CHADS 2 score ≥ 2 ( n = 52) | CHADS 2 score < 2 ( n = 46) | P |
|---|---|---|---|---|
| Age (y) | 71 ± 11 | 76 ± 9 | 65 ± 9 | <.001 |
| Men/women | 68/30 | 34/18 | 34/12 | .361 |
| RR interval (msec) | 784 ± 168 | 790 ± 174 | 776 ± 163 | .688 |
| Systolic blood pressure (mm Hg) | 135 ± 24 | 139 ± 28 | 130 ± 19 | .094 |
| Diastolic blood pressure (mm Hg) | 79 ± 11 | 80 ± 13 | 78 ± 8 | .526 |
| Body surface area (m 2 ) | 1.75 ± 0.20 | 1.72 ± 0.21 | 1.79 ± 0.19 | .132 |
| Diabetes mellitus | 24 (24%) | 23 (44%) | 1 (2%) | <.001 |
| Hypertension | 51 (52%) | 39 (75%) | 12 (26%) | <.001 |
| Cerebrovascular accident | 9 (9%) | 9 (17%) | 0 (0%) | .006 |
| Congestive heart failure | 34 (35%) | 25 (48%) | 9 (20%) | .003 |
| CHADS 2 score | 1.72 ± 1.34 | 2.77 ± 0.90 | 0.54 ± 0.50 | <.001 |
| CHA 2 DS 2 -VASc score | 3.04 ± 1.88 | 4.37 ± 1.44 | 1.54 ± 0.96 | <.001 |
| LV end-diastolic volume (mL) | 82 ± 37 | 83 ± 44 | 79 ± 28 | .609 |
| LV end-systolic volume (mL) | 42 ± 29 | 46 ± 35 | 37 ± 20 | .154 |
| Left atrial volume index (mL/m 2 ) | 45 ± 19 | 48 ± 20 | 44 ± 17 | .318 |
| LVEF (%) | 52 ± 15 | 50 ± 16 | 55 ± 14 | .118 |
| E (cm/sec) | 92 ± 21 | 93 ± 21 | 91 ± 21 | .581 |
| E-wave deceleration time (msec) | 150 ± 45 | 163 ± 56 | 136 ± 21 | .003 |
| Ea (cm/sec) | 9.4 ± 2.5 | 8.1 ± 1.9 | 10.7 ± 2.5 | <.001 |
| E/Ea ratio | 10.6 ± 4.1 | 12.0 ± 4.2 | 9.0 ± 3.4 | <.001 |
| LV PLSSavg (%) | −11.5 ± 3.4 | −10.7 ± 3.3 | −12.4 ± 3.3 | .012 |
| LV PLSSindex (%) | −11.4 ± 3.6 | −10.6 ± 3.3 | −12.4 ± 3.6 | .010 |
LV PLSSindex had a highly significant correlation with LV PLSSavg in all patients ( r = 0.970, P < .001; Figure 2 A). Bland-Altman analysis showed only small bias of 0.01%, and the 95% limits of agreement were +1.64% to −1.62% ( Figure 2 B). LV PLSSindex had also a highly significant correlation with LV PLSSavg both in patients with standard deviations above ( r = 0.968, P < .001; Figure 3 A) and below ( r = 0.978, P < .001; Figure 3 B) the median standard deviation over the 15 RR intervals. Bland-Altman analysis also showed only small biases of 0.01% and 0%, and the 95% limits of agreement were +1.85% to −1.80% and +1.42% to −1.42% in patients with standard deviations above and below the median standard deviation over the 15 RR intervals, respectively ( Figures 3 C and 3 D). The value of LV PLSSindex minus LV PLSSavg between the patients with standard deviations above and below the median standard deviation over the 15 RR intervals was not different ( P = .907). Similarly, Bland-Altman analysis showed only small biases of 0.07% and −0.05%, and the 95% limits of agreement were +1.55% to −1.42% and +1.74% to −1.84% in patients with CHADS 2 scores ≥ 2 and < 2, respectively. The value of LV PLSSindex minus LV PLSSavg between patients with CHADS 2 scores ≥ 2 and < 2 was not different ( P = .487). In addition, Bland-Altman analysis also revealed only small biases of −0.01% and 0.06%, and the 95% limits of agreement were +1.71% to −1.73% and +1.53% to −1.41% in patients with LVEFs ≥ 50% and < 50%, respectively. The value of LV PLSSindex minus LV PLSSavg between patients with LVEFs ≥ 50% and < 50% was also not different ( P = .688). In addition, we could acquire the values of LV PLSS measured from excellent match, good match, and modest match of RR intervals in 86, 78, and 67 patients, respectively. The differences between LV PLSSavg and the values of LV PLSS measured from excellent match, good match, and modest match of RR intervals were 0.63 ± 0.60%, 0.69 ± 0.63%, and 0.76 ± 0.56%, respectively. There was no significant difference between any two of them ( P ≥ .247).
The comparison of baseline and echocardiographic characteristics between patients with CHADS 2 scores ≥ 2 and < 2 is also shown in Table 1 . Compared with patients with CHADS 2 scores < 2, patients with CHADS 2 scores ≥ 2 had older age; higher prevalence of histories of diabetes mellitus, hypertension, cerebrovascular accident, and congestive heart failure; longer E-wave deceleration times; lower Ea; higher E/Ea ratios; lower LV PLSSavg; and lower LV PLSSindex. However, the left atrial volume index and LVEF were comparable between the two groups. In addition, PLSSavg and PLSSindex had significant correlations with Ea ( r = −0.659, P < .001, and r = −0.653, P < .001, respectively) and the E/Ea ratio ( r = 0.357, P < .001, and r = 0.348, P < .001, respectively).
Table 2 shows the comparison of baseline and echocardiographic characteristics between patients with CHA 2 DS 2 -VASc scores ≥ 2 and < 2. Compared with patients with CHA 2 DS 2 -VASc scores < 2, those with CHA 2 DS 2 -VASc scores ≥ 2 had older age; higher systolic blood pressure; lower body surface area; higher prevalence of histories of diabetes mellitus, hypertension, and congestive heart failure; lower Ea; higher E/Ea ratios; lower LV PLSSavg; and lower LV PLSSindex. However, the left atrial volume index and LVEF were comparable between the two groups. Table 2 also shows the comparison of baseline and echocardiographic characteristics between patients with CHA 2 DS 2 -VASc scores ≥ 2 and < 2 in patients with CHADS 2 scores < 2. In this subgroup analysis, compared with patients with CHA 2 DS 2 -VASc scores < 2, those with CHA 2 DS 2 -VASc scores ≥ 2 had older age, shorter RR intervals, lower Ea, higher E/Ea ratios, lower LV PLSSavg, and lower LV PLSSindex. However, the left atrial volume index and LVEF were comparable between the two groups.
| Characteristic | All patients | P | CHADS 2 score < 2 | P | ||
|---|---|---|---|---|---|---|
| CHA 2 DS 2 -VASc score ≥ 2 ( n = 73) | CHA 2 DS 2 -VASc score < 2 ( n = 25) | CHA 2 DS 2 -VASc score ≥ 2 ( n = 21) | CHA 2 DS 2 -VASc score < 2 ( n = 25) | |||
| Age (y) | 74 ± 10 | 61 ± 7 | <.001 | 71 ± 9 | 61 ± 7 | <.001 |
| Men/women | 47/26 | 21/4 | .081 | 13/8 | 21/4 | .107 |
| RR interval (msec) | 769 ± 171 | 827 ± 157 | .137 | 716 ± 153 | 827 ± 157 | .020 |
| Systolic blood pressure (mm Hg) | 138 ± 25 | 125 ± 17 | .027 | 136 ± 19 | 125 ± 17 | .059 |
| Diastolic blood pressure (mm Hg) | 79 ± 12 | 79 ± 8 | .939 | 77 ± 9 | 79 ± 8 | .618 |
| Body surface area (m 2 ) | 1.73 ± 0.20 | 1.84 ± 0.18 | .018 | 1.73 ± 0.18 | 1.84 ± 0.18 | .050 |
| Diabetes mellitus | 24 (33%) | 0 (0%) | <.001 | 1 (5%) | 0 (0%) | .457 |
| Hypertension | 43 (59%) | 8 (32%) | .020 | 4 (19%) | 8 (32%) | .502 |
| Cerebrovascular accident | 9 (12%) | 0 (0%) | .106 | 0 (0%) | 0 (0%) | – |
| Congestive heart failure | 32 (44%) | 2 (8%) | .001 | 7 (21%) | 2 (8%) | .059 |
| CHADS 2 score | 2.18 ± 1.23 | 0.40 ± 0.50 | <.001 | 0.71 ± 0.46 | 0.40 ± 0.50 | .033 |
| CHA 2 DS 2 -VASc score | 3.81 ± 1.53 | 0.80 ± 0.41 | <.001 | 2.43 ± 0.60 | 0.80 ± 0.41 | <.001 |
| LV end-diastolic volume (mL) | 81 ± 40 | 82 ± 29 | .946 | 76 ± 27 | 82 ± 29 | .515 |
| LV end-systolic volume (mL) | 44 ± 32 | 37 ± 19 | .297 | 38 ± 22 | 37 ± 19 | .771 |
| Left atrial volume index (mL/m 2 ) | 47 ± 20 | 43 ± 14 | .315 | 45 ± 20 | 43 ± 14 | .585 |
| LVEF (%) | 51 ± 16 | 57 ± 13 | .088 | 52 ± 15 | 57 ± 13 | .312 |
| E (cm/sec) | 93 ± 22 | 88 ± 15 | .275 | 94 ± 26 | 88 ± 15 | .346 |
| E-wave deceleration time (ms) | 155 ± 51 | 138 ± 17 | .102 | 134 ± 26 | 138 ± 17 | .628 |
| Ea (cm/sec) | 8.6 ± 2.0 | 11.6 ± 2.6 | <.001 | 9.7 ± 1.9 | 11.6 ± 2.6 | .008 |
| E/Ea ratio | 11.5 ± 4.2 | 8.0 ± 2.4 | <.001 | 10.2 ± 4.0 | 8.0 ± 2.4 | .029 |
| LV PLSSavg (%) | −10.8 ± 3.2 | −13.4 ± 3.4 | .001 | −11.2 ± 2.9 | −13.4 ± 3.4 | .029 |
| LV PLSSindex (%) | −10.8 ± 3.2 | −13.5 ± 3.7 | .001 | −11.2 ± 3.2 | −13.5 ± 3.7 | .031 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
