Background
When stratifying thromboembolic risk to patients with atrial fibrillation (AF), left atrial appendage (LAA) thrombus is currently the only echocardiographic index that absolutely contraindicates cardioversion. The aim of this study was to identify the predictors of LAA “sludge” and its impact on subsequent thromboembolism and survival in patients with AF.
Methods
A total of 340 patients (mean age, 66 ± 12 years; 75% men) who underwent transesophageal echocardiography to exclude LAA thrombus before electrical cardioversion or radiofrequency pulmonary vein isolation) for AF were retrospectively studied. LAA sludge was defined as a dynamic, viscid, layered echodensity without a discrete mass, visualized throughout the cardiac cycle. Follow-up was obtained after a mean of 6.7 ± 3.7 years, and patients were analyzed according to LAA thrombus ( n = 62 [18%]), sludge ( n = 47 [14%]), or spontaneous echocardiographic contrast ( n = 84 [25%]). Patients without these transesophageal echocardiographic characteristics served as controls ( n = 147 [43%]).
Results
LAA sludge was independently predicted by enlarged left atrial area (odds ratio, 4.54; 95% confidence interval [CI], 2.38–8.67; P < .001), reduced LAA emptying velocity (odds ratio, 12.7; 95% CI, 6.11–26.44; P < .001), and reduced left ventricular ejection fraction (odds ratio, 2.11; 95% CI, 1.03–4.32; P < .001). Thromboembolic event and all-cause mortality rates in patients with sludge were 23% and 57%, respectively. Multiple logistic regression analyses identified the presence of LAA sludge to be independently associated with thromboembolic complications (adjusted hazard ratio, 3.43; 95% CI, 1.42–8.28; P = .006) and all-cause mortality (adjusted hazard ratio, 2.02; 95% CI, 1.22–3.36; P = .007).
Conclusion
Sludge within the LAA is independently associated with subsequent thromboembolic events and all-cause mortality in patients with AF.
Among the clinical and echocardiographic indices used to stratify thromboembolic risk to patients with atrial fibrillation (AF), left atrial (LA) appendage (LAA) thrombus is a powerful predictor of increased risk and stands as the only absolute contraindication to cardioversion.
Spontaneous echocardiographic contrast (SEC) is associated with a high incidence of thrombus formation and thromboembolic events, but there is no consensus that cardioversion is contraindicated in patients without coexistent LA or LAA thrombus identified. Current AF management guidelines do not recognize “sludge” within the LAA as an entity identified by transesophageal echocardiography (TEE). Although sludge may be regarded as a precursor to thrombus, the long-term fate of these patients has not been reported. In this study, we investigated the predictors of LAA sludge without thrombus and its impact on subsequent thromboembolism and survival in patients with AF.
Methods
Patients
The study was approved by the institutional review board. The source population for this retrospective case-control series was identified from the Cleveland Clinic Adult Echocardiography Laboratory Database between 1996 and 2005. During the study period, we identified 491 consecutive patients with electrocardiographically documented permanent or paroxysmal AF who were referred for TEE to exclude thrombus within the LAA. Patients with valvular AF ( n = 17) or prosthetic valves ( n = 42) were excluded because of the established association with AF and thromboembolism. We also excluded patients who did not undergo transthoracic echocardiography within 1 month of index TEE ( n = 76) and those without Social Security numbers ( n = 16). A total of 340 patients were included for the final analysis. The baseline clinical and echocardiographic characteristics of the 340 study subjects are described in Table 1 and Table 2 , respectively.
| Variable | Control ( n = 147) | SEC ( n = 84) | Sludge ( n = 47) | Thrombus ( n = 62) | P |
|---|---|---|---|---|---|
| Age (y) | 64 ± 12 | 65 ± 11 | 68 ± 9 | 69 ± 14 | .040 |
| Gender (male) | 71% | 76% | 81% | 81% | .246 |
| Undergoing PVI | 34% | 39% | 23% | 18% | .217 |
| Previous embolic events, TIA, or ischemic stroke | 14% | 12% | 28% | 18% | .047 |
| Hypertension | 61% | 62% | 72% | 56% | .234 |
| Congestive heart failure | 31% | 37% | 70% | 74% | <.001 |
| Diabetes mellitus | 19% | 18% | 23% | 40% | .698 |
| Ischemic heart disease | 36% | 38% | 55% | 47% | .033 |
| Current cigarette smoking | 7% | 5% | 15% | 10% | .168 |
| Dyslipidemia | 48% | 33% | 45% | 34% | .849 |
| Aspirin | 27% | 30% | 30% | 16% | .832 |
| Warfarin at echocardiography | 56% | 54% | 41% | 47% | .079 |
| INR at echocardiography | 2.0 ± 1.2 | 1.9 ± 1.0 | 1.7 ± 0.4 | 1.8 ± 0.9 | .185 |
| Warfarin at follow-up | 84% | 87% | 96% | 94% | .061 |
| INR at follow-up | 1.9 ± 0.6 | 2.0 ± 0.7 | 1.9 ± 0.3 | 1.9 ± 0.7 | .970 |
| Heparin | 42% | 43% | 51% | 47% | .361 |
| BNP (pg/mL) | 133 (37–228) | 210 (75–258) | 510 (278–869) | 847 (131–1801) | .016 |
| CRP (mmol/L) | 2.40 (0.7–10.7) | 1.70 (0.9–8.5) | 13.2 (4.2–22.2) | 8.5 (5.8–11.3) | .025 |
| CHADS 2 score | 1.6 ± 1.1 | 1.4 ± 1.3 | 2.5 ± 1.2 | 2.4 ± 1.5 | <.001 |
| LA area (cm 2 ) | 24 ± 6 | 28 ± 7 | 32 ± 8 | 33 ± 7 | <.001 |
| Moderate or severe LA enlargement (>30 cm 2 ) | 14% | 24% | 64% | 66% | <.001 |
| LAA emptying velocity (cm/sec) | 47 ± 18 | 28 ± 19 | 16 ± 6 | 17 ± 9 | <.001 |
| Reduced LAA emptying velocity (<20 cm/sec) | 2% | 17% | 76% | 69% | <.001 |
| LVEF (%) | 47 ± 11 | 47 ± 12 | 31 ± 14 | 29 ± 14 | <.001 |
| Impaired LVEF (<55%) | 52% | 52% | 77% | 92% | .004 |
| LV hypertrophy | 66% | 62% | 65% | 62% | .997 |
| RV dysfunction | 26% | 27% | 72% | 74% | <.001 |
| Aortic stenosis | 2% | 4% | 6% | 3% | .361 |
| Aortic regurgitation | 6% | 11% | 12% | 11% | .296 |
| Mitral regurgitation | 26% | 30% | 34% | 24% | .380 |
| Tricuspid regurgitation | 15% | 20% | 45% | 33% | <.001 |
| Atrial fibrillation or flutter at the time of TEE | 68% | 79% | 69% | 72% | .932 |
| Ascending and arch aortic atheroma >4 mm or mobile | 12% | 14% | 17% | 12% | .526 |
| Outcome | Control | SEC | P | Sludge | P | Thrombus | P |
|---|---|---|---|---|---|---|---|
| Death | 40/147 (27%) | 15/84 (18%) | .11 | 27/47 (57%) | <.001 | 39/62 (63%) | <.001 |
| Thromboembolic event | 10/137 (7%) | 9/78 (12%) | .29 | 11/44 (25%) | .014 | 11/55 (20%) | .034 |
Echocardiographic Studies
An Acuson Sequoia (Siemens Medical Solutions USA, Inc, Mountain View, CA) or ATL 5000 (Philips Medical Systems, Andover, MA) echocardiographic system and a multiplane 5- to 7.5-MHz transducer were used for dedicated TEE to visualize the left atrium and LAA from multiple planes. Digital cine loops and/or standard VHS tapes of all echocardiographic studies were evaluated offline (ProSolv Cardiovascular Analyzer 3.5; Problem Solving Concepts, Indianapolis, IN) in consensus by two experienced readers (B.S.L., A.L.K.) blinded to the clinical data.
SEC was recognized as dynamic, swirling, smokelike echoes within the LAA cavity. LAA sludge was defined as an intracavitary echodensity with viscid gelatinous qualities giving the impression of impending precipitation, but without a discrete organized mass, continuously seen throughout the cardiac cycle. Sludge appears more dense and layered than severe SEC ( Figure 1 , Videos 1–4 ; available at www.onlinejase.com ). LAA thrombus was defined from multiple imaging planes as a well-circumscribed, solid echodensity acoustically distinct from the underlying endocardium and pectinate muscles that was less heterogeneous and dynamic than sludge.
Patients were analyzed according to the appearance within the left atrium and LAA on the index TEE: thrombus ( n = 62 [18%]), sludge ( n = 47 [14%]), and SEC ( n = 84 [25%]). The control group consisted of 147 patients (43%) without LAA thrombus, sludge, or SEC. Patients with coexistent appearances were assigned to the higher thrombogenic spectrum group for analysis (thrombus > sludge > SEC > none).
LAA emptying velocities were evaluated during TEE using pulsed-wave Doppler with a sample volume placed 1 cm into the LAA at transducer phased-array angles of 45° to 120°. Low LAA velocities were defined as <20 cm/sec at end-diastole.
Transthoracic studies were performed using 2.5- and 3.5-MHz phased-array transducers with an Acuson Sequoia system using standard echocardiographic views. Left ventricular ejection fraction (LVEF), LA area (planimetry from the apical four-chamber view), left ventricular wall thickness by M-mode echocardiography, presence of a mechanical valve prosthesis, and assessment of valvular disease were also determined during transthoracic echocardiography according to current guidelines. Left ventricular mass by two-dimensionally guided M-mode echocardiography was derived using the American Society of Echocardiography’s guideline formula. Normal LVEF was defined as ≥55%, normal LA area <20 cm 2 , and normal wall thickness as <1 cm in women and <1.1 cm in men. Mild LA dilation was defined as LA area 20 to 30 cm 2 , moderate as LA area 30 to 40 cm 2 and severe as LA area > 40 cm 2 . Right ventricular dysfunction was defined as right ventricular systolic dysfunction (right ventricular fractional area change < 40%) from the two-dimensional transthoracic images. There were no patients with valvular disease in the final analysis, because of the established association with AF and thromboembolism. No patients were excluded from analysis because of suboptimal echocardiographic image quality.
Anticoagulation Profile
Patients were referred for TEE before cardioversion or pulmonary vein isolation (PVI) to assess for LAA thrombus if no international normalized ratio (INR) testing was performed or there was a documented subtherapeutic INR (<2) within the preceding 3 weeks of their procedural hospital admission or reevaluation of a preexisting LAA thrombus. Patients with subtherapeutic INRs and no LAA thrombus before cardioversion or PVI were given either intravenous or low–molecular weight heparin periprocedurally. The cardioversion or PVI of patients with LAA thrombus were deferred and treated with warfarin anticoagulation for ≥4 weeks before re-referral.
Reproducibility
To assess the agreement of grading transesophageal echocardiographic findings of LAA SEC, sludge, or thrombus, the digital cine loops of 30 randomly selected patients were reevaluated by a second observer blinded to the initial assignment. Kappa statistics were used to assess the concordance of findings between two investigators.
Outcomes
The primary outcome was thromboembolic events, including cerebrovascular accidents, transient ischemic attacks, or other systemic emboli identified by medical chart review and/or a validated patient questionnaire. The secondary end point was all-cause mortality. Death was identified by the 2012 Social Security Death Index database, which has been shown to be highly specific (>99.5%) and unbiased. The results of the echocardiographic studies were not available to the investigator assessing end points.
Statistical Analysis
Continuous variables are summarized as mean ± SD if normally distributed and as medians and interquartile ranges if not normally distributed. Normality was assessed by the Shapiro-Wilk W test. Categorical variables are summarized as proportions and frequencies. Differences in continuous clinical variables across patient groups were assessed by Student’s t tests or one-way analysis of variance if normally distributed and by Wilcoxon or Kruskal-Wallis (rank sum) tests if not normally distributed. Differences in clinical proportions across patient groups were assessed by contingency table analysis. Odds ratios were calculated through logistic regression analysis and evaluated according to the likelihood ratio test. The Cox proportional-hazards model was used to assess the clinical risk for all-cause mortality or thromboembolic event (transient ischemic attacks, cerebrovascular accidents, and systemic embolism) associated with LAA sludge. Covariate selection for model entry was based on clinical experience and the identification of known correlates of thromboembolic events. Because of the relatively small numbers of events, the numbers of covariate entered into the models were restricted. The effects of LAA sludge on mortality and thrombus were analyzed and compared using Kaplan-Meier plots stratified by transesophageal echocardiographic appearance (SEC, sludge, and thrombus). The CHADS 2 score was used to stratify stroke risk using clinical data, with two points assigned for a history of stroke or transient ischemic attack and one point each for age >75 years and history of hypertension, diabetes mellitus, or heart failure. Interreader agreement is expressed as κ coefficients. All P values reported were from two-sided tests and considered statistically significant when <.05. Statistical analyses were performed using JMP version 8.0.2 and SAS version 9.0 (SAS Institute Inc, Cary, North Carolina).
Results
Characteristics of the Population
Among the 340 study subjects, sludge within the LAA was identified by TEE in 47 (14%). Compared with the control group, patients with LAA sludge were older, and previous thromboembolic events, heart failure, and ischemic heart disease were more prevalent ( Table 1 ). The frequency of warfarin or heparin use and the intensity of oral anticoagulation at index TEE were similar between the study groups, including a comparison of the sludge and thrombus groups (mean INR, 1.7 ± 0.4 vs 1.8 ± 0.9; P = .46).
Sixty-two patients (18%) were found to have thrombus within the LAA by TEE. Patient age and the frequency of heart failure, ischemic heart disease, and current tobacco smoking were similar among the LAA sludge and thrombus study groups at baseline. The mean CHADS 2 score of patients with LAA sludge was not significantly different compared with patients with LAA thrombus (2.5 ± 1.2 vs 2.4 ± 1.5, P = .81). Increasing brain natriuretic peptide levels were associated with increasing grades of thrombogenesis, lowest with SEC and highest with thrombus.
Interreader Variability
There was excellent interreader agreement for the assignment of SEC, sludge, and thrombus on TEE (κ = 0.86).
Baseline Echocardiographic Characteristics
Moderate LA enlargement, reduced LAA emptying velocity, and moderate impairment of global left ventricular systolic function were common in patients with LAA sludge ( Table 1 ). LA size, LAA emptying velocity, LVEF, and right ventricular dysfunction in this study group bore a remarkable similarity to these values in patients with LAA thrombus. The occurrence of more than mild or mobile ascending or transverse aortic atheroma defined by TEE did not differ between the study groups.
Echocardiographic Predictors of LAA Sludge
In multivariate logistic regression analysis, enlarged LA area (>20 cm 2 ; odds ratio, 4.54; 95% confidence interval [CI], 2.38–8.67; P < .001), reduced LAA emptying velocity (<20 cm/sec; odds ratio, 12.7; 95% CI, 6.11–26.44; P < .001), and reduced LVEF (to <55%; odds ratio, 2.11; 95% CI, 1.03–4.32; P < .001) were independent predictors of LAA sludge.
Outcomes
Mortality and thromboembolic event follow-up was complete for all patients in the study. Anticoagulation status follow-up was available in 315 of 340 patients (93%). Of the patients with LAA sludge, anticoagulation status follow-up was available in 96% of patients (45 of 47). This compares with 92% of patients with SEC (77 of 84) and 94% of patients with thrombus (58 of 62). Two hundred fifty-three patients were receiving oral anticoagulation therapy with warfarin alone at follow-up, nine patients were receiving aspirin only, 40 patients receiving combined warfarin and aspirin, 10 patients were receiving aspirin and clopidogrel, and three patients were receiving no anticoagulation therapy. Antithrombotic therapy data were extracted from hospital clinic note reviews, electronic laboratory results for INR testing, family physician phone interviews, and patient questionnaires. INR values closest to the date preceding a thromboembolic event or death were analyzed. In patients without event outcomes, the most recent INR was recorded. At the time of follow-up, INR values were measured in all patients with warfarin therapy. There was no significant difference in the available INR values between groups (control vs SEC vs sludge vs thrombus, 1.8 ± 0.6 vs 2.0 ± 0.7 vs 1.9 ± 0.3 vs 1.9 ± 0.7), and the follow-up INR value was not associated with thromboembolism in univariate logistic regression analysis. In addition, we compared the rate of thromboembolic events in patients with INR follow-up with that in patients without INR follow-up.
The mean follow-up duration was 6.4 ± 3.6 years from index TEE, 122 patients (36%) died, and 41 patients (12%) had systemic embolic events. The total rates for the LAA sludge group were 27 deaths (57%) and 11 thromboembolic events (23%) during follow-up ( Table 2 ). The identification of sludge within the LAA was associated with a higher risk for all-cause mortality (hazard ratio [HR], 2.11; 95% CI, 1.47–3.03; P < .001) and a higher risk for thromboembolism (HR, 3.43; 95% CI, 1.56–7.52; P = .002) compared with the control group ( Table 3 ).
