Background
The aims of this study were to evaluate the relationship of the CHA 2 DS 2 -VASc score and risk categories with transesophageal echocardiographic (TEE) risk factors for thromboembolism and to compare the CHA 2 DS 2 -VASc and CHADS 2 risk stratification schemes with respect to their ability to predict these risk factors in a multiethnic US population with nonvalvular atrial fibrillation.
Methods
Transesophageal echocardiograms of 167 patients (mean age, 66.3 ± 11.6 years; 146 men [87%]; 100 whites [60%]; 40 Hispanics [24%]; 27 blacks [16%]) with nonvalvular atrial fibrillation were retrospectively reviewed for smoke, sludge, thrombus, and left atrial appendage (LAA) emptying velocity ≤20 cm/sec. The patients’ CHA 2 DS 2 -VASc and CHADS 2 risk scores and categories were also calculated.
Results
Any LAA abnormality, smoke, sludge, thrombus, and abnormal LAA emptying velocity were present in 45%, 38%, 13%, 3%, and 22% of patients, respectively. Heart failure ( P < .001), age ( P < .001 for age ≥75 vs ≤64 years, P = .013 for age 65–74 vs ≤64 years), and diabetes ( P = .019) were independent predictors of LAA abnormalities, while ethnicity was not. The prevalence of TEE risk factors for thromboembolism increased with increasing CHA 2 DS 2 -VASc score and risk category. The CHADS 2 risk categories of 35 patients (21%) were upgraded by the CHA 2 DS 2 -VASc scheme. Using the latter scheme, fewer patients were classified as at intermediate risk compared with the CHADS 2 system (21 [13%] vs 46 [28%]). Patients classified as at low risk by either scheme had almost no TEE risk factors. Of 30 intermediate-risk patients by CHADS 2 score upgraded to high risk using CHA 2 DS 2 -VASc score, eight (27%) had at least one TEE risk factor for thromboembolism. C-statistics, sensitivity, and specificity for predicting any LAA abnormality were 0.607 (95% confidence interval, 0.549–0.665), 92.0%, and 28.9% for CHA 2 DS 2 -VASc score and 0.685 (95% confidence interval, 0.615–0.755), 81.3%, and 54.2% for CHADS 2 score.
Conclusions
CHA 2 DS 2 -VASc score is associated with TEE risk factors for thromboembolism in a multiethnic US population. Compared with CHADS 2 score, it has increased sensitivity, decreased specificity, and lower ability for predicting TEE risk factors in this population.
Anticoagulation reduces the risk for thromboembolism in patients with atrial fibrillation but is also associated with significant hemorrhagic complications. Several risk stratification schemes have been developed to identify patients with nonvalvular atrial fibrillation (NVAF) who are at higher thromboembolic risk and therefore are more likely to achieve net clinical benefit from anticoagulation. The CHADS 2 scheme (congestive heart failure, hypertension, age ≥ 75 years, diabetes mellitus, and stroke or transient ischemic attack [double weight]) is the most popular of these schemes but has been criticized because it can classify too many patients with NVAF as at intermediate risk, for which optimal antithrombotic therapy is uncertain. Because of this, the CHA 2 DS 2 -VASc scheme (congestive heart failure, hypertension, age ≥ 75 years [double weight], diabetes mellitus, stroke or transient ischemic attack [double weight], vascular disease, age 65 to 74 years, and sex category) has been introduced with the goal of refining risk stratification by adding vascular disease and female gender as risk factors to the CHADS 2 scheme and assigning double weight to advanced age. In studies of predominantly European populations, the CHA 2 DS 2 -VASc risk stratification scheme has been reported to identify patients at “truly low risk” for stroke and peripheral embolism better than CHADS 2 and to assign fewer patients to the intermediate risk category without sacrificing overall diagnostic performance. In its guidelines for the management of atrial fibrillation, the European Society of Cardiology has endorsed the use of the CHA 2 DS 2 -VASc scheme in stratifying thromboembolic risk in patients with NVAF and has recommended anticoagulation for patients with a CHA 2 DS 2 -VASc scores ≥2 (one major or two nonmajor risk factors). The American College of Cardiology and American Heart Association focused update on atrial fibrillation did not incorporate the CHA 2 DS 2 -VASc scheme into management strategies.
NVAF risk factors, prevalence, comorbidities, antithrombotic treatments, stroke subtypes, and other adverse outcomes vary by ethnic group. Although risk factors for NVAF, such as hypertension and diabetes, are more common among Hispanics and blacks, the prevalence of NVAF is paradoxically higher in Caucasians. This paradox has been observed in the general population and in patients with stroke and heart failure. Blacks with NVAF have a higher incidence of stroke than Caucasians, but the strokes they experience are more likely to be due to noncardioembolic etiologies. Recurrent strokes are more common in Mexican Americans than in Caucasians with NVAF and prior strokes. Because these ethnic variations in NVAF involve stroke and other components of the CHA 2 DS 2 -VASc score, there may be differences in the relationship of the score to stroke and other thromboembolic events between multiethnic US and European populations. Therefore, the observation of superior clinical value of the CHA 2 DS 2 -VASc scheme for predicting outcomes relative to CHADS 2 made in European populations with NVAF may not extend to multiethnic US populations.
Left atrial appendage (LAA) thrombogenic milieu is important in the pathogenesis of NVAF-related stroke and other thromboembolic events. Several transesophageal echocardiographic (TEE) markers of LAA thrombogenic milieu, including thrombus, smoke, sludge, and decreased LAA emptying velocity, predict stroke and other systemic embolism and have been used as stroke surrogates in some studies evaluating clinical risk scores. Individual components of the CHADS 2 score are associated with these TEE predictors of thromboembolism, and a higher CHADS 2 score, representing the cumulative effects of these clinical elements, is associated with increased prevalence of TEE risk factors. Recently, the CHA 2 DS 2 -VASc score has also been shown to be associated with these TEE risk factors in European patients with NVAF, but this relationship has not been well studied in US populations. On the basis of the assumption that TEE predictors of thromboembolism are reasonable surrogates for cardioembolic stroke, an analysis of the relationship of CHA 2 DS 2 -VASc score to these risk factors may provide insight into its clinical value for stratifying risk in a multiethnic US population with NVAF. The aims of the present study were to (1) evaluate the relationship of the CHA 2 DS 2 -VASc score and risk category with TEE risk factors for thromboembolism in a multiethnic US population and (2) compare the ability of the CHA 2 DS 2 -VASc and the CHADS 2 risk stratification schemes to predict these TEE risk factors in this population.
Methods
This study was a retrospective review of transesophageal echocardiograms and electronic clinical records. We searched the TEE databases of our university-affiliated hospital and the adjacent Veterans Affairs medical center for consecutive patients undergoing TEE imaging between 2008 and 2012 while in atrial fibrillation and identified 193 patients. Of these, 12 with mitral valve replacement or annuloplasty rings, five with severe organic mitral regurgitation, seven with mitral stenosis, and two Asian patients were excluded. The remaining 167 patients, including 77 (46.1%) from the academic hospital and 90 (53.9%) from the Veterans Affairs facility, were considered to have NVAF and were included in our analysis. Demographic characteristics, self-reported ethnicity, and clinical variables were obtained by review of electronic medical records. Clinical variables of interest included history of hypertension, diabetes mellitus, heart failure, transient ischemic attack, stroke, systemic or pulmonary embolism, myocardial infarction, amputation, and lower extremity revascularization. Atrial fibrillation was classified as paroxysmal, persistent, or permanent according to definitions in the American College of Cardiology and American Heart Association guidelines, and classification could be determined in 165 patients. Hypertension was defined as systolic blood pressure > 140 mm Hg or diastolic blood pressure > 90 mm Hg on two separate occasions or treatment with one or more antihypertensive agents if not being used for rate control. Diabetes mellitus was defined as the use of diabetic medications or fasting blood glucose > 126 mg/dl. Left ventricular ejection fraction was obtained from the report of the transthoracic echocardiographic study performed closest in time to the TEE study. For the purpose of comparing the two schemes, we used the CHA 2 DS 2 -VASc definition of heart failure (any history of systolic heart failure or ejection fraction < 40%) as the heart failure criterion for both schemes rather than the original CHADS 2 criterion of a history of recent systolic or diastolic heart failure. Vascular disease was defined as intermittent claudication, amputation, lower extremity revascularization, history of myocardial infarction, or complex aortic plaques on TEE imaging. From these variables, we calculated each patient’s CHADS 2 score (range, 0–6) and CHA 2 DS 2 -VASc score (range, 0–9). For both schemes, we also classified patients with scores of 0, 1, and ≥2 in low-risk, intermediate-risk, and high-risk categories, respectively.
We reviewed the digitally stored transesophageal echocardiograms of these patients for smoke, sludge, thrombus, and abnormal mean LAA emptying velocity blinded to the clinical data and risk scores. We also assessed aortic atherosclerosis as part of the vascular disease assessment for CHA 2 DS 2 -VASc, considering plaque thickness > 4 mm or mobile components as complex plaque. We defined smoke as dynamic swirling smokelike echo signals imaged with optimal gain settings. We diagnosed sludge as more organized echodensities with a gelatinous appearance. We defined thrombus as a discrete mass seen from multiple windows and separate from the endocardium and pectinate muscles. For assessment of LAA emptying velocity, 10 consecutive pulsed-wave Doppler outflow velocity signals during diastole were measured 1 cm below the orifice of the LAA over at least three cardiac cycles and averaged. LAA emptying velocity < 20 cm/sec was considered abnormal. LAA emptying velocity could be measured in 155 patients but was not performed or was unsuitable for measurement in the remaining 12. Patients with at least one of the aforementioned TEE findings were considered to have any LAA abnormality associated with an increased risk for thromboembolism. Of the 12 patients in whom LAA emptying velocity could not be measured, three had other LAA abnormalities and were included in the analyses of any LAA abnormality ( n = 158), while the remaining nine were excluded from such analysis.
Statistical Analysis
Demographic and clinical characteristics and LAA parameters on TEE imaging are reported as mean ± SD for continuous variables and as numbers and proportions for categorical variables. Continuous variables were compared using analysis of variance or Kruskal-Wallis tests, as appropriate, and categorical variables using χ 2 tests. The relationships of the individual components of the CHA 2 DS 2 -VASc score with LAA thrombogenic milieu were assessed using univariate and multivariate stepwise logistic regression, with the presence of any LAA abnormality on TEE imaging as the dependent variable. All models that were considered included ethnicity and three age groups: <64, 65 to 74, and ≥75 years. Models with interactions were also evaluated, but no significant interactions were found. Results of the logistic regression are reported as odds ratios (ORs) and corresponding 95% confidence intervals (CIs). For both schemes, the relation of risk scores and categories to TEE risk factors was examined using χ 2 tests for trend. Receiver operating characteristic curve analyses were performed to compare the ability of the two risk schemes to predict any LAA abnormality, smoke, sludge, thrombus, and decreased LAA emptying velocity. C-statistics (area under the curve) along with their 95% CIs are reported for both schemes. The differences between the areas under the curves of the two schemes and the corresponding 95% CIs are also reported for each LAA abnormality. Because the European Society of Cardiology recommends anticoagulation for patients with CHADS 2 and CHA 2 DS 2 -VASc scores ≥ 2 (i.e., for those who are assigned to the high-risk category), we were most interested in measuring the performance of this risk category for predicting thrombogenic milieu in our multiethnic population. We therefore report the sensitivity, specificity, and positive and negative predictive values of CHA 2 DS 2 -VASc and CHADS 2 scores ≥ 2 for detecting LAA abnormalities.
To assess the reproducibility of the measurements of LAA emptying velocity, LAA velocities were remeasured in 20 randomly selected transesophageal echocardiograms by the original observer several weeks after his initial assessment and also by a second echocardiographer. Intraobserver and interobserver variability are reported as mean percentage error (absolute difference between the two measurements for each subject divided by their mean) and intraclass correlation coefficient with 95% CI. All statistical tests with two-sided P values < .05 were considered statistically significant. All analyses were performed using SPSS version 19 for Windows (SPSS, Inc., Chicago, IL) and SAS version 9.2 (SAS Institute Inc., Cary, NC).
The study was approved by the institutional review boards of both institutions. Patients gave permission for use of their clinical data as part of the consent for TEE imaging.
Results
Clinical Characteristics and TEE Findings
The mean age of the patients was 66.3 ± 11.6 years, and 146 were men (87.4%). Our patients included 100 non-Hispanic whites (59.9%), 40 Hispanics (24.0%), and 27 blacks (16.2%). Table 1 shows demographic characteristics, type of atrial fibrillation, indications for TEE studies, individual clinical components of risk scores, mean CHADS 2 and CHA 2 DS 2 -VASc scores, and LAA abnormalities for the entire study group and for each ethnic group. Whites tended to be older and included a higher proportion of men. The large majority of TEE studies were performed to evaluate the LAA for thrombus before cardioversion, radiofrequency ablation, or closure. The overall mean CHADS 2 and CHA 2 DS 2 -VASc scores were 2.2 ± 1.4 and 3.2 ± 1.8, respectively. Comorbidities were common in this group of patients with NVAF referred for TEE imaging, with nearly half having histories of heart failure or moderate left ventricular dysfunction. Also, 75 patients (44.9%) had at least one LAA abnormality. Smoke, sludge, thrombus, and abnormal LAA emptying velocity were present in 64 (38.3%), 22 (13.2%), five (3.0%), and 36 (21.6%) patients, respectively. Blacks tended to more commonly have smoke, and a larger proportion of them had sludge compared with the other two ethnic groups.
Variable | Overall | Whites | Hispanics | Blacks | P |
---|---|---|---|---|---|
Ethnicity | 167 (100.0%) | 100 (59.9%) | 40 (24.0%) | 27 (16.2%) | |
Age (y) | 66.3 ± 11.6 | 68.1 ± 9.5 | 64.5 ± 14.0 | 62.3 ± 13.5 | .063 ∗ |
Men | 146 (87.4%) | 93 (93.0%) | 32 (80.0%) | 21 (77.8%) | .028 |
AF classification | .836 | ||||
Paroxysmal | 54 (32.7%) | 32 (32.3%) | 12 (30.0%) | 10 (38.5%) | |
Persistent | 67 (40.6%) | 39 (39.4%) | 19 (47.5%) | 9 (34.6%) | |
Permanent | 44 (26.7%) | 28 (28.3%) | 9 (22.5%) | 7 (26.9%) | |
Indications for TEE imaging | .287 † | ||||
Cardioversion | 114 (68.3%) | 66 (66.0%) | 26 (65.0%) | 22 (81.5%) | |
Ablation/closure | 15 (9.0%) | 6 (6.0%) | 6 (15.0%) | 3 (11.1%) | |
Mitral valve | 12 (7.2%) | 11 (11.0%) | 1 (2.5%) | 0 (0%) | |
Endocarditis | 11 (6.6%) | 6 (6.0%) | 4 (10.0%) | 1 (3.7%) | |
Cardiac source of embolus | 8 (4.8%) | 4 (4.0%) | 3 (7.5%) | 1 (3.7%) | |
Aortic valve | 4 (2.4%) | 4 (4.0%) | 0 (0%) | 0 (0%) | |
Heart failure | 2 (1.2%) | 2 (2.0%) | 0 (0%) | 0 (0%) | |
Tricuspid valve | 1 (0.6%) | 1 (1.1%) | 0 (0%) | 0 (0%) | |
Hypertension | 126 (75.4%) | 74 (74.0%) | 28 (70.0%) | 24 (88.9%) | .184 |
Diabetes | 60 (35.9%) | 33 (33%) | 15 (37.5%) | 12 (44.4%) | .531 |
Stroke/TIA/peripeheral embolism | 30 (18.0%) | 16 (16.0%) | 8 (20.0%) | 6 (22.2%) | .702 |
Vascular disease | 55 (32.9%) | 32 (35.0%) | 14 (35.0%) | 6 (22.2%) | .433 |
Heart failure/low EF | 74 (44.3%) | 44 (44.0%) | 15 (37.5%) | 15 (55.6%) | .343 |
CHADS 2 score | 2.2 ± 1.4 | 2.1 ± 1.4 | 2.2 ± 1.5 | 2.5 ± 1.3 | .261 ∗ |
CHA 2 DS 2 -VASc score | 3.2 ± 1.8 | 3.1 ± 1.8 | 3.1 ± 1.9 | 3.4 ± 1.7 | .711 ∗ |
Any LAA abnormality | 75 (47.5%) | 43 (45.3%) | 16 (43.2%) | 16 (61.5%) | .285 |
Smoke | 64 (38.3%) | 34 (34.0%) | 14 (35.0%) | 16 (59.3%) | .050 |
Sludge | 22 (13.2%) | 8 (8.0%) | 4 (10.0%) | 10 (37.0%) | <.001 |
Thrombus | 5 (3.0%) | 2 (2.0%) | 1 (3.5%) | 2 (7.4%) | .336 |
Mean LAA emptying velocity < 20 cm/sec | 36 (23.2%) | 20 (21.3%) | 8 (21.6%) | 8 (33.3%) | .443 |