Echocardiography is a widely used and versatile technique that can provide comprehensive information concerning thromboembolic risk in patients with atrial fibrillation. The authors review the potential contributions of echocardiography to thromboembolic risk stratification and to decreasing the thromboembolic risk associated with procedures such as cardioversion and ablation. Unsolved questions and new possibilities that have arisen from the development of strain and strain rate imaging are also discussed.
Echocardiography is a widely used and versatile technique that can provide comprehensive information concerning thromboembolic risk in patients with atrial fibrillation (AF).
Although transesophageal echocardiography (TEE) is a mainstay tool for preventing cardioversion and catheter ablation–related thromboembolism, the potential role of transthoracic echocardiography (TTE) in the risk stratification of patients with AF and in decision making regarding long-term thromboembolism prophylaxis has been overlooked. However, recent investigations have shown that despite the generalized used of clinical risk stratification based solely on the CHADS 2 and CHA 2 DS 2 -VASc scores (measures that have very modest discriminative capability, with C-statistics ranging from 0.54 to 0.65 ), echocardiography may be useful for fine-tuning these classifications.
In the next sections, we review the echocardiographic parameters that play a role in patients with AF in three different situations: (1) evaluation before cardioversion or ablation, (2) predicting the presence of left atrial (LA) appendage thrombus on TEE, and (3) thromboembolic risk stratification to select the appropriate antithrombotic strategy.
Emphasis is placed on different cardiac structures according to the chosen image acquisition technique (transthoracic or transesophageal) and the provided information. For the sake of clarity, we refer to two-dimensional (2D) TTE and TEE unless clearly specified otherwise (e.g., three-dimensional [3D] TTE or 3D TEE).
The promising development of strain and strain rate imaging, alongside the unsolved matters in the field of thromboembolic risk assessment of patients with nonvalvular AF, is also explored.
Transesophageal Echocardiography for Assessment of the Left Atrial Appendage
Two-dimensional TEE provides excellent characterization of the LA appendage and LA environment because of the anatomic relationship of these structures to the esophagus. Despite being a narrow tubular structure with very complex anatomy (variable shape, size, and orientation, with the possibility of several lobes and branches), the LA appendage thrombi can be very accurately identified using 2D TEE, with values of sensitivity and specificity approaching 99%. Figure 1 illustrates the capabilities of 2D TEE for providing very clear and diagnostic images of LA thrombi ( [CR] ; available at www.onlinejase.com ). However, careful examination is required to avoid false-negative findings. Furthermore, the muscular ridges and pectinate muscles must be carefully observed, because they can be misinterpreted as clots.
Three-dimensional TEE allows a more comprehensive assessment of multiple lobes of the LA appendage, which may be located in different planes, and a more accurate estimation of LA appendage geometry and size. This is advantageous, because 2D TEE shows only a “slice” of the LA appendage at a given time (which may result in underestimation or overestimation of orifice size, depth, and number of lobes). Another advantage of 3D TEE is better distinction between the pectinate muscles and thrombi.
The use of contrast has also been proposed as an option for improving LA thrombus detection on 2D TEE, thereby reducing the rate of adverse events in patients undergoing cardioversion.
An epiphenomenon of intracardiac thrombus is the presence of spontaneous echocardiographic contrast (SEC), also known as “swirl” or “smoke,” which is associated with low blood flow velocity. The dynamic smoke like echoes are thought to be composed either of aggregated activated platelets and leukocytes or aggregates or of red blood cells that are interacting with plasma proteins. SEC can be classified into four groups (1 to 4+, depending on the intensity, location, and presence of the swirling movement) according to the method proposed by Fatkin et al. Figure 2 illustrates the presence of dense SEC with swirling movements in the left atrium and LA appendage ( [CR] ; available at www.onlinejase.com ). Increasing grades of SEC have been associated with decreased LA appendage flow velocity and increased LA size.
Sludge is a dynamic and gelatinous, but not solid or well-formed, echodensity present throughout the cardiac cycle ( Figure 3 , [CR] ; available at www.onlinejase.com ). It is often difficult to differentiate sludge from a thrombus. Along the continuum of thrombus formation, sludge is thought to represent a stage beyond SEC (the stages are mild, moderate, and severe SEC; sludge; and thrombus) and may have greater prognostic significance.
The presence of thrombi in the LA appendage, SEC, low LA appendage flow velocities, and complex aortic plaques on TEE has long been associated with stroke, thromboembolism, and adverse prognosis. The presence of at least one of LA appendage thrombus, low LA appendage flow velocities, or SEC is designated as an LA abnormality and is associated with a high risk for stroke.
A high likelihood of cerebral embolism (either clinically assessed or observed on magnetic resonance images) and/or death at medium term despite anticoagulation in patients with AF and dense SEC (3+ or 4+) has been described.
LA appendage thrombus and dense SEC are very powerful predictors of cardiovascular death independent of other known clinical risk factors, including congestive heart failure, diabetes mellitus, hypertension, and vascular disease.
The following transesophageal echocardiographic changes are part of a dynamic process: LA appendage thrombus disappears either because of embolization or adequate anticoagulation; SEC appears or increases over time alongside with dilatation, fibrosis, and progressive atrial dysfunction or stunning; and LA appendage flow decreases with time as AF episode duration progresses or increases some weeks after an effective cardioversion. Therefore, one transesophageal echocardiographic examination is like a single photograph of the LA milieu and may not be fully representative of past and future changes at this level.
The presence of a thrombus in the LA appendage is a contraindication to the performance of cardioversion because of the associated risk for stroke. Therefore, when a rhythm-control strategy is chosen in patients with AF of >24 to 48 hours in duration, guidelines recommend TEE to exclude LA appendage thrombus as an alternative to 3 weeks of effective preprocedural anticoagulation.
This recommendation results from the landmark findings of the Assessment of Cardioversion Using Transesophageal Echocardiography trial and was recently tested using the new anticoagulant dabigatran in the Randomized Evaluation of Long Term Anticoagulant Therapy cardioversion substudy. In a post hoc analysis of trial patients undergoing cardioversion (of whom 13%–25% had undergone previous TEE), no differences in the prevalence of LA appendage thrombus and SEC were found when comparing the three treatment arms (adjusted warfarin vs dabigatran 110 mg twice daily or 150 mg twice daily). Moreover, the analysis of data from before, during, and 30 days after cardioversion revealed comparable incidences of stroke, systemic embolism, and major bleeding.
Recent investigations have assessed the ability of the CHADS 2 and CHA 2 DS 2 -VASc scores to discriminate transesophageal echocardiographic risk factors (LA appendage thrombus, SEC, sludge, and low LA appendage flow velocities). However, these scores showed low to moderate abilities (with areas under the curve ranging from 0.6 to 0.7) to discriminate between the risk factors.
According to the 2012 consensus statement on catheter and surgical ablation of AF, preprocedural TEE is indicated in all patients with AF of >48 hours in duration and without anticoagulation at a therapeutic level in the previous 3 weeks. Still, in many centers, all patients undergoing AF ablation, even those who are in sinus rhythm and are effectively anticoagulated, undergo preprocedural TEE. This treatment plan is derived from the fact that patients with CHADS 2 or CHA 2 DS 2 -VASc scores of zero and a minority of therapeutically anticoagulated patients (1.6%–2.1%) may develop thrombus or “sludge” in the LA appendage. However, some studies suggest that patients under anticoagulation and with CHADS 2 and CHA 2 DS 2 -VASc scores < 2 (negative predictive value approaching 100%) may be spared TEE before catheter ablation of AF.
Alternatives to TEE in these circumstances (e.g., the ratio of left ventricular [LV] ejection fraction to LA volume index, among others) are discussed in the next section.
Other less validated and less frequently used parameters, such as intracardiac intensity variation at the orifice of the LA appendage after contrast infusion and LA appendage ejection fraction evaluated through vector velocity imaging, have been associated with cerebrovascular events and the formation of LA appendage thrombus.
The most relevant evidence concerning transesophageal echocardiographic parameters and their association with thromboembolism and prognosis is presented in Table 1 .
| Study | Study design and setting | Main findings |
|---|---|---|
| Leung et al. (1994) | Cohort 272 patients with nonvalvular AF undergoing TEE Mean follow-up, 17.5 mo | SEC was the only positive predictor (OR, 3.5; P = .03) of stroke or embolic events on multivariate analysis |
| Zabalgoitia et al. (1998) | Cross-sectional 789 participants undergoing TEE at entry to SPAF III | LA appendage thrombi (RR, 2.5; P = .04), dense SEC (RR, 3.7; P < .001), LA appendage peak flow velocities ≤ 20 cm/sec (RR, 1.7; P = .008), and complex aortic plaque (RR, 2.1; P < .001) were independently associated with increased thromboembolic risk (predefined categories for the SPAF III trial) |
| The Stroke Prevention in Atrial Fibrillation Investigators Committee on Echocardiography (1998) | Cohort 382 high-risk patients with nonvalvular AF Adjusted-dose vs low-intensity warfarin plus aspirin 325 mg/daily (combination therapy) Mean follow-up, 1.1 y | The presence of LA abnormality was associated with a high risk for stroke (7.8% per year) in patients under combination therapy In the presence of a complex aortic plaque, the stroke rate was 12.0% per year and reached 20.5% when both LA appendage abnormality and complex aortic plaques were present In patients without any of these changes, the stroke rate was 1.3% per year. |
| Klein et al. (2001) ACUTE trial | Multicenter, randomized, prospective 1,222 patients with AF >48 h in duration Anticoagulant treatment vs TEE-guided strategy 8-wk follow-up | Similar efficacy for the prevention of thromboembolic events: 0.8% TEE-guided strategy vs 0.5% conventional oral anticoagulant therapy ( P = .50) Patients in the TEE-guided strategy group presented less bleeding complications (2.9% vs 5.5%, P = .03) Greater rate of successful restoration of sinus rhythm with the TEE-guided strategy (71.1% vs 65.2%, P = .03), but no differences were found at 8 wk concerning maintenance of sinus rhythm, functional status, and death |
| Bernhardt et al. (2005) | Single-center prospective 128 consecutive patients with dense SEC and AF vs 143 patients with faint SEC and AF, all under continued oral anticoagulation 12-mo follow-up | A very high (22%) likelihood of cerebral embolism (either clinically assessed or by magnetic resonance images) and/or death was observed in patients with dense SEC |
| Dawn et al. (2005) | Cohort 175 patients with AF and no or only mild mitral regurgitation Mean follow-up, 31 mo | LA appendage thrombus (RR, 5.52; P = .024) and LA SEC (RR, 7.96; P = .013) were the only predictors of cardiovascular death, independently of other known clinical risk factors such as congestive heart failure, diabetes mellitus, hypertension, and vascular disease |
| Bernhardt et al. (2006) | Cohort 43 patients with LA appendage thrombi and persistent or permanent AF, all under oral anticoagulation 3-y follow-up | Cerebral embolism and/or death occurred in 22 (51%) patients Thrombi disappeared in 31 subjects (72%) (those with smaller thrombi, lower echogenicity, and lower LA volumes), due either to embolism or to effective anticoagulation |
Transthoracic Echocardiography
The main focus of TTE-derived measures as predictors of stroke or risk stratification in patients with AF has been almost exclusively restricted to depressed LV ejection fraction (which has strong evidentiary support). Nevertheless, other parameters, including LA diameter measured on M-mode echocardiography, LA area, and LA volume, have already shown some evidence of accuracy, suggesting a possible role in risk stratification.
The incorporation of TTE-derived parameters seems to be plausible for the refinement of clinical risk scores. However, despite the recommendations of the National Institute for Health and Clinical Excellence, TTE-derived parameters play only a minor role in current risk classification schemes.
Another possible use of these noninvasively collected measurements may be the prediction of prothrombotic changes in the left atrium and LA appendage to obviate the need for TEE in individuals thought to be at very low thromboembolic risk.
An overview of the different transthoracic echocardiographic parameters and their roles is presented in Tables 2 and 3 .
| Study | Study design and setting | Main findings |
|---|---|---|
| The Stroke Prevention in Atrial Fibrillation Investigators (1992) | Cohort 568 patients with nonrheumatic AF Mean follow-up, 1.3 y | 14 transthoracic echocardiographic variables were assessed for predicting ischemic stroke or systemic embolism Only LA size (measured on M-mode echocardiography) and depressed LVEF were independent predictors of thromboembolism on multivariate analysis and improved risk stratification when combined with three clinical risk factors: history of hypertension, recent congestive heart failure, and previous thromboembolism |
| Osranek et al. (2005) | Cohort 45 patients with lone AF Mean follow-up, 27 y | Individuals with indexed LA volume ≥ 32 mL/m 2 had worse event-free survival (HR, 4.46; P = .005) Cerebral infarction occurred in 7 patients, all with indexed LA volumes ≥ 32 mL/m 2 |
| Lee et al. (2008) | Cross-sectional 330 patients with persistent AF and preserved LVEF | E/E′ ratio was independently associated with ischemic stroke on multivariate analysis |
| Shin et al. (2010) | Cohort 148 patients with AF and heart failure with preserved LVEF Median follow-up, 27 mo | S′ and E′, particularly when combined, were independent predictors of a composite of cardiovascular death, recurrent heart failure, and ischemic stroke |
| Azemi et al. (2013) | Case-control (57 patients each group) Nonvalvular AF CHADS 2 score ≤ 1 before index event | Patients with stroke presented reduced peak negative and peak positive LA strain values, when compared with controls |
| Author, year | Study design and setting, n | Main findings |
|---|---|---|
| Omran et al. (1999) | Cross-sectional 117 patients with stroke or TIA TTE and TEE were performed | Imaging of the LA appendage was possible using TTE in 75% of patients Both methods were concordant for the detection of LA appendage thrombus in 10 patients TTE failed to detect a thrombus in 1 patient that was later diagnosed on TEE |
| Moreira et al. (2005) | Cross-sectional 51 patients with acute ischemic neurologic events undergoing both TTE and TEE | Doppler and LA appendage area evaluation was possible by TTE in 98% of patients A moderate correlation was found between LA appendage peak velocities ( r = 0.63, P < .001) and LA appendage area ( r = 0.73, P < .001) when comparing both techniques |
| Karakus et al. (2008) | Cross-sectional 92 patients referred for exclusion of LA appendage thrombus underwent 2D and 3D TTE and 2D TEE | 3D TTE had a very high correlation with 2D TEE Compared with 3D TTE, 2D TEE seemed to more frequently misdiagnose the pectinate musculature as thrombus |
| Sallach et al. (2009) | Cross-sectional, multicenter 118 patients with AF > 2 days Undergoing both TTE and TEE | Thrombi in the LA appendage were detected using TTE using harmonic imaging and intravenous contrast LA appendage apical peak E′ ≤ 9.7 cm/sec and anterior S′ ≤ 5.2 cm/sec on TTE-DTI could identify patients with sludge or thrombus |
| Puwanant et al. (2009) | Cross-sectional 1,058 patients undergoing TEE before catheter ablation of AF | LVEF < 35% and congestive heart failure were independent predictors of LA thrombus or sludge 50% of patients with CHADS 2 scores of 1 and LA thrombus had LVEFs ≤ 35% |
| Kleemann et al. (2009) | Prospective, single-center 295 patients with nonvalvular AF with CHADS 2 scores of 0 or 1 undergoing TEE before cardioversion | LVEF < 40% and LA dimension ≥ 50 mm were independent predictors of the presence of SEC and thrombus Echocardiography was proposed as a tool to further stratify patients with low CHADS 2 scores |
| Ayirala et al. (2011) | Cross-sectional 334 patients with AF undergoing TEE | LVEF, indexed LA volume, and CHADS 2 score were independent predictors of LA appendage thrombus formation, and their combination was a useful measure for identifying patients at low risk for LA appendage thrombus formation A normal LVEF and normal LA volume index were associated with the absence of LA appendage thrombus |
| Providência et al. (2012) | Cross-sectional 376 patients with nonvalvular AF undergoing TEE | The addition of LA area and LVEF to the CHADS 2 and CHA 2 DS 2 -VASc scores increased their discriminative capability with regard to LA appendage thrombus, dense SEC, and low LAA flow velocities |
| Providência et al. (2012) | Cross-sectional 180 patients with nonvalvular AF undergoing TEE before cardioversion | The incorporated independent predictors in a model to detect the presence of LA appendage thrombus were C-reactive protein, indexed LA volume, troponin I, episode duration, and previous stroke or embolism This model displayed a high discriminative capability (AUC, 0.816) for LA appendage thrombus No patients with LA appendage thrombus were observed among those with scores ranging from 0 to 2, which corresponded to 49.4% of the sample ( n = 89) |
| Doukky et al. (2013) | Cross-sectional 245 patients with nonvalvular AF undergoing TEE before cardioversion or electrophysiology procedures | External validation of the ratio of LVEF to indexed LA volume, which confirms its high discriminative power (AUC, 0.83) for the presence of LA appendage thrombus, which were not detected in patients with a ratio ≥ 1.5 (i.e., 50.7% of the population) |
LV Ejection Fraction
In the early 1990s, the use of TTE for the risk stratification of patients with AF was proposed following data from the Stroke Prevention in Atrial Fibrillation study, in which a depressed LV ejection fraction was identified as an independent predictor of ischemic stroke and systemic embolism. However, the results of further testing and prospective validation of this classification were not published.
In a 2008 comparison of 12 risk stratification schemes used to predict stroke in patients with nonvalvular AF, TTE parameters were present in half the schemes and included depressed LV ejection fraction (2D echocardiography) and fractional shortening (M-mode echocardiography).
In the more recent CHA 2 DS 2 -VASc classification, an LV ejection fraction < 40% was included as a surrogate for congestive heart failure.
Another key element is that approximately 90% of thrombi in patients with AF arise from the LA appendage. Thus, a parameter that could refine the predictive models of LA appendage thrombus would be useful for predicting thromboembolic events during follow-up. A normal LV ejection fraction has been associated with the absence of LA appendage thrombus formation in patients with AF undergoing TEE. Moreover, assessment of the LV ejection fraction has increased the discriminative capability of clinical classifications (CHADS 2 and CHA 2 DS 2 -VASc scores) with regard to LA appendage thrombi and other markers of LA stasis.
LV Hypertrophy and Diastolic Dysfunction
The only evidence concerning LV hypertrophy as a possible risk factor for stroke in patients with AF is from studies conducted in the 1980s and 1990s. However, these also included an important percentage of patients with valvular heart disease, including rheumatic mitral stenosis and aortic stenosis, meaning that the results are difficult to interpret.
A recent investigation suggested that indexed LV mass has a very high discriminative capability for the prediction of LA appendage thrombi: the C-statistic for this parameter was 0.98, which was far higher than the value for SEC (0.73–0.75). However, these data still need to be reproduced and further validated.
Parameters concerning LV diastolic function have been shown to be related to thromboembolic risk in patients with AF: the ratio of transmitral Doppler early filling velocity to tissue Doppler early diastolic mitral annular velocity (E/E′ ratio) was independently associated with ischemic stroke, and systolic (S′) and early diastolic (E′) mitral annular velocities were independent predictors of a composite end point of cardiovascular death, recurrent heart failure, and ischemic stroke.
LA Size
A relationship between increased LA dimensions (measured on M-mode echocardiography), AF, and stroke was proposed in the 1980s in a small, retrospective case-control study of 40 patients with AF, of whom 20 had previous ischemic strokes. Similar evidence was presented by Aronow et al. However, the fact that some of the patients from these two studies had rheumatic heart disease or valvular heart disease renders the data difficult to validate in the context of nonvalvular AF.
In the Stroke Prevention in Atrial Fibrillation study, LA size (measured on M-mode echocardiography) was an independent predictor of thromboembolism and added incremental predictive value to the clinical risk factors.
The majority of investigations performed in the 1980s and 1990s, including LA dimensions, focused only on LA diameter measured on M-mode echocardiography, which is now known to be an inappropriate and outdated method. The current recommendations of the European Association of Echocardiography and the American Society of Echocardiography support measuring LA biplane volume using either the area-length formula or the modified Simpson’s rule as the preferred method for assessing LA size. However, when more recent studies are evaluated, favorable evidence concerning parameters such as indexed LA volume results from small size studies. Some of the studies that failed to confirm an association between LA volume and thromboembolic events were not appropriately powered for AF. Nonetheless, a normal LA volume index has been associated with the absence of LA appendage thrombi formation in patients with AF.
On the basis of the fact that TEE is a semi-invasive procedure that carries risks over transthoracic imaging, finding TTE-derived parameters that can accurately separate patients with AF into groups at very low risk (who can be spared TEE) and moderate or high risk (who must undergo TEE) for having a LA appendage thrombus has been the aim of recent investigations.
A combination of LA size and LV ejection fraction has been shown to have good discriminative capability for detecting patients with LA appendage thrombus or other transesophageal echocardiographic findings. A ratio of LV ejection fraction to indexed LA volume > 1.5 has recently been shown to be highly accurate at excluding the presence of LA appendage thrombus in patients with AF who are candidates for AF catheter ablation or cardioversion. This ratio has been successfully validated, confirming its clinical applicability.
Another possibility is to combine transthoracic echocardiographic parameters with clinical data to further refine the prediction rules. LA area and LV ejection fraction were found to increase the discriminative capability of the CHADS 2 and CHA 2 DS 2 -VASc scores with regard to LA appendage thrombus, dense SEC, and low LA appendage flow velocities (≤20 cm/sec). Ayirala et al. have also shown that higher CHADS 2 scores, increased LA volume indexes, and lower LV ejection fractions were predictors of LA appendage thrombus.
A risk score that combines clinical data with biomarkers and echocardiographic parameters has been recently developed to accurately detect patients with AF with very low risk for thromboembolism who can be spared TEE. Indexed LA volume was included in this model, which may prove very useful in the precardioversion or preablation assessment. However, validation of these results is still needed.
Imaging of the LA Appendage
Some attempts have been made to examine and acquire indexes of LA appendage function using 2D TTE. This can be accomplished using parasternal short-axis or apical two-chamber views. In a small study that compared 2D TTE and 2D TEE, imaging of the LA appendage was possible in 75% of patients using 2D TTE, at expense of lower sensitivity for the detection of thrombus.
A very good correlation between the peak velocities in the LA appendage measured by 2D TTE and 2D TEE was also described. In another small study, LA appendage emptying velocities were reduced on 2D TTE in patients with acute ischemic stroke with LA appendage thrombus or SEC.
Doppler tissue imaging (DTI) of the LA appendage has been shown using 2D TTE and presented a strong correlation with the values acquired using 2D TEE. This parameter seems to be compromised in patients with AF and high thromboembolic risk. In the Comprehensive Left Atrial Appendage Optimization of Thrombus study, LA appendage thrombi were accurately detected on 2D TTE with intravenous contrast. Cutoff values and the best segments of the LA appendage for DTI assessment and measurement of peak S′ and E′ to predict LA appendage thrombi, sludge, or severe SEC were also defined.
However, despite these promising results, imaging of the LA appendage using 2D TTE is operator dependent and has a learning curve.
Two studies have shown that 3D TTE may be as accurate as 2D TEE for assessing the LA appendage if the acoustic window is accurate. However, strong evidence concerning this technique as a substitute for 2D TEE is still lacking, as the sample size of the studies was small, and >50% of the participants evaluated had conditions other than AF.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
