Predicting Left Atrial Stasis in Patients With AF
Predicting Left Atrial Stasis in Patients With AF
During the pre-specified inclusion period, 133 patients were assessed during an AF episode of longer than 48 hour duration and with no effective anticoagulation in the preceding 3 weeks. Of these, the following were excluded from analysis due to the presence of exclusion criteria: inappropriate endocardial border definition of the left atrium (n = 11), previous closure of the LAA (n = 3), valve repair or presence of prosthetic heart valve (n = 10) and significant valvular disease (n = 27).
Out of the remaining 82 patients, most were male (65.9%) and the average CHADS2 and CHA2DS2-VASc scores were 2.0 ± 1.1 and 3.2 ± 1.6, respectively. Information on demographics, clinical data and anti-thrombotic medication are provided on Table 1.
Thrombus or sludge in the left atrium or LAA was found in 16 patients (19.5%) (group I). No differences were found for baseline variables when comparing these patients with the remaining (group II), except for the estimated AF duration, which was more often longer than 1 month in group I.
As regards standard transthoracic echocardiographic parameters, no differences were found between the two groups (Table 2). Also, no significant differences were found regarding left ventricular longitudinal strain and strain rate. On transesophageal echocardiogram lower flow velocities in the LAA and a higher prevalence of dense spontaneous echocardiographic contrast were observed in group I.
Adequate tracking of all 6 segments was possible in most patients (n = 58; 70.7%). The segments which were more frequently excluded from analysis were Seg6 (in 18 patients; 22.0%) and Seg11 (in 7 patients; 8.5%). In all the remaining segments, a satisfactory definition of the endocardial borders and tracking was possible (97.6% Seg2 2; 93.9% Seg3; 96.3% Seg4; 96.3% Seg5).
Bland-Altman analysis for inter-observer variability of speckle-tracking derived data is shown in Figure 3. Small differences were observed overall for PPS (−0.3%; 95%CI −5.2; 4.6), PNS (0.5%; 95%CI −3.5; 4.5), PPSR (−0.02 s; 95%CI −0.54; 0.51), PNSR (0.06 s; 95%CI −0.35; 0.48) and TPPS (6.7 ms; 95%CI −93.2; 106.6).
(Enlarge Image)
Figure 3.
Results of Bland-Altman analysis for interobserver variability regarding left atrial deformation. Legend: peak positive strain (PPS – A.), peak negative strain (PNS – B.), peak positive strain rate (PPSR – C.), peak negative strain rate (PNSR – D.) and time-to-peak systolic strain (TPPS – E.).
Lower values of peak positive and peak negative strain rate, as well as a lower peak-to-peak strain rate, were observed in group I patients. Patients with sludge or thrombi had a trend for higher indexed left atrial volume and left atrial dyssynchrony, as assessed through the standard deviation of time to peak positive strain (Table 3).
A moderate positive correlation was found between peak positive strain rate and maximum emptying velocity (r = 0.589; P < 0.001) and peak positive strain rate and maximum filling velocity of the LAA (r = 0.651; p < 0.001). Peak negative strain rate was also found to be associated both with maximum emptying velocity (r = −0.513; p < 0.001) and maximum filling velocity of the LAA (r = −0.552; p < 0.001). No significant correlation was observed between peak negative strain and LAA flow velocities and only a trend for a very slight association was observed between peak positive strain and both LAA maximum emptying velocity (r = 0.231; p = 0.066) and LAA maximum filling velocity (r = 0.222; p = 0.078).
On univariate analysis, body mass index, AF episode duration, indexed left atrial volume, peak positive strain rate, peak negative strain rate, peak-to-peak strain rate and time to peak positive strain were predictors of thrombus or sludge on transesophageal echocardiogram. However, only AF duration, peak negative strain rate and time-to-peak positive strain remained significant on multivariate analysis (Table 3). The area under the curve for the estimated probabilities using the obtained logistic regression model was 0.89 (95%CI 0.81–0.96; P < 0.001). The same logistic regression model also displayed a high discriminative capability for the prediction of dense spontaneous echocardiographic contrast: c-statistic = 0.81; 95%CI 0.71–0.91; P < 0.001).
Results
During the pre-specified inclusion period, 133 patients were assessed during an AF episode of longer than 48 hour duration and with no effective anticoagulation in the preceding 3 weeks. Of these, the following were excluded from analysis due to the presence of exclusion criteria: inappropriate endocardial border definition of the left atrium (n = 11), previous closure of the LAA (n = 3), valve repair or presence of prosthetic heart valve (n = 10) and significant valvular disease (n = 27).
Out of the remaining 82 patients, most were male (65.9%) and the average CHADS2 and CHA2DS2-VASc scores were 2.0 ± 1.1 and 3.2 ± 1.6, respectively. Information on demographics, clinical data and anti-thrombotic medication are provided on Table 1.
Echocardiographic Findings
Thrombus or sludge in the left atrium or LAA was found in 16 patients (19.5%) (group I). No differences were found for baseline variables when comparing these patients with the remaining (group II), except for the estimated AF duration, which was more often longer than 1 month in group I.
As regards standard transthoracic echocardiographic parameters, no differences were found between the two groups (Table 2). Also, no significant differences were found regarding left ventricular longitudinal strain and strain rate. On transesophageal echocardiogram lower flow velocities in the LAA and a higher prevalence of dense spontaneous echocardiographic contrast were observed in group I.
LA Deformation in Patients With LA Thrombus or Sludge
Adequate tracking of all 6 segments was possible in most patients (n = 58; 70.7%). The segments which were more frequently excluded from analysis were Seg6 (in 18 patients; 22.0%) and Seg11 (in 7 patients; 8.5%). In all the remaining segments, a satisfactory definition of the endocardial borders and tracking was possible (97.6% Seg2 2; 93.9% Seg3; 96.3% Seg4; 96.3% Seg5).
Bland-Altman analysis for inter-observer variability of speckle-tracking derived data is shown in Figure 3. Small differences were observed overall for PPS (−0.3%; 95%CI −5.2; 4.6), PNS (0.5%; 95%CI −3.5; 4.5), PPSR (−0.02 s; 95%CI −0.54; 0.51), PNSR (0.06 s; 95%CI −0.35; 0.48) and TPPS (6.7 ms; 95%CI −93.2; 106.6).
(Enlarge Image)
Figure 3.
Results of Bland-Altman analysis for interobserver variability regarding left atrial deformation. Legend: peak positive strain (PPS – A.), peak negative strain (PNS – B.), peak positive strain rate (PPSR – C.), peak negative strain rate (PNSR – D.) and time-to-peak systolic strain (TPPS – E.).
Lower values of peak positive and peak negative strain rate, as well as a lower peak-to-peak strain rate, were observed in group I patients. Patients with sludge or thrombi had a trend for higher indexed left atrial volume and left atrial dyssynchrony, as assessed through the standard deviation of time to peak positive strain (Table 3).
Predictors of LA Stasis
A moderate positive correlation was found between peak positive strain rate and maximum emptying velocity (r = 0.589; P < 0.001) and peak positive strain rate and maximum filling velocity of the LAA (r = 0.651; p < 0.001). Peak negative strain rate was also found to be associated both with maximum emptying velocity (r = −0.513; p < 0.001) and maximum filling velocity of the LAA (r = −0.552; p < 0.001). No significant correlation was observed between peak negative strain and LAA flow velocities and only a trend for a very slight association was observed between peak positive strain and both LAA maximum emptying velocity (r = 0.231; p = 0.066) and LAA maximum filling velocity (r = 0.222; p = 0.078).
On univariate analysis, body mass index, AF episode duration, indexed left atrial volume, peak positive strain rate, peak negative strain rate, peak-to-peak strain rate and time to peak positive strain were predictors of thrombus or sludge on transesophageal echocardiogram. However, only AF duration, peak negative strain rate and time-to-peak positive strain remained significant on multivariate analysis (Table 3). The area under the curve for the estimated probabilities using the obtained logistic regression model was 0.89 (95%CI 0.81–0.96; P < 0.001). The same logistic regression model also displayed a high discriminative capability for the prediction of dense spontaneous echocardiographic contrast: c-statistic = 0.81; 95%CI 0.71–0.91; P < 0.001).
Source...