Author + information
- Received June 16, 2009
- Accepted July 8, 2009
- Published online November 24, 2009.
- Sarinya Puwanant, MD⁎,†,
- Brandon C. Varr, BS⁎,
- Kevin Shrestha, AB⁎,
- Sarah K. Hussain, MD⁎,
- W.H. Wilson Tang, MD⁎,
- Ruvin S. Gabriel, MD⁎,
- Oussama M. Wazni, MD⁎,
- Mandeep Bhargava, MD⁎,
- Walid I. Saliba, MD⁎,
- James D. Thomas, MD⁎,
- Bruce D. Lindsay, MD⁎ and
- Allan L. Klein, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Allan L. Klein, Heart & Vascular Institute, Cleveland Clinic, 9500 Euclid Avenue, Desk F15, Cleveland, Ohio 44195
Objectives The goals of this study were to determine: 1) if low-risk patients assessed by a CHADS2score, a clinical scoring system quantifying a risk of stroke in patients with atrial fibrillation (AF), require a routine screening transesophageal echocardiogram (TEE) before pulmonary vein isolation (PVI); and 2) the relationship of a CHADS2score with left atrial (LA)/left atrial appendage (LAA) spontaneous echo contrast, sludge, and thrombus.
Background There is no clear consensus of whether a screening TEE before catheter ablation of AF should be performed in every patient.
Methods Initial TEEs for pre-PVI of 1,058 AF patients (age 57 ± 11 years, 80% men) were reviewed and compared with a CHADS2score.
Results CHADS2scores of 0, 1, 2, 3, 4, 5, and 6 were present in 47%, 33%, 14%, 5%, 1%, 0.3%, and 0% of patients, respectively. The prevalence of LA/LAA thrombus, sludge, and spontaneous echo contrast were present in 0.6%, 1.5%, and 35%. The prevalence of LA/LAA thrombus/sludge increased with ascending CHADS2score (scores 0 [0%], 1 [2%], 2 [5%], 3 [9%], and 4 to 6 [11%], p < 0.01). No patient with a CHADS2score of 0 had LA/LAA sludge/thrombus. In a multivariate model, history of congestive heart failure and left ventricular ejection fraction <35% were significantly associated with sludge/thrombus.
Conclusions The prevalence of LA/LAA sludge/thrombus in patients with AF undergoing a pre-PVI screening TEE is very low (<2%) and increases significantly with higher CHADS2scores. This suggests that a screening TEE before PVI should be performed in patients with a CHADS2score of ≥1, and in patients with a CHADS2score of 0 when the AF is persistent and therapeutic anticoagulation has not been maintained for 4 weeks before the procedure.
- atrial fibrillation
- left atrial appendage
- pulmonary vein isolation
- transesophageal echocardiography
Catheter ablation has emerged as an established strategy for a restoration of sinus rhythm in patients with atrial fibrillation (AF) (1–4). Recent studies have shown that catheter ablation of AF was associated with thromboembolic events, with most events occurring within 2 weeks after the procedure (5). The ablation procedure leaves patients with scar and endothelial injury, which creates a thrombogenic milieu leading to thrombus formation. The 2007 Heart Rhythm Society/European Heart Rhythm Association/European Cardiac Arrhythmia Society guidelines (6) recommend performing pre-procedural transesophageal echocardiography (TEE) screening for the presence of thrombus in patients with persistent AF who are in AF at the time of ablation (6). However, there is no clear consensus of whether screening TEE before catheter ablation of AF should be performed in every patient.
The aims of this study were: 1) to determine if low-risk patients with AF assessed by a CHADS2score, a clinical scoring system quantifying a risk of stroke in patients with AF, necessarily require a routine screening TEE before pulmonary vein antral isolation (PVI); 2) to determine the relationship of a CHADS2score with pre-procedural left atrial (LA)/left atrial appendage (LAA) characteristics of thrombogenic milieu identified by a TEE; and 3) to determine the prevalence of LA/LAA thrombus and/or sludge in these patients.
From January 1999 to December 2006, 1,208 consecutive patients with AF who underwent TEE for pre-PVI screening at Cleveland Clinic were identified from Cleveland Clinic Echocardiography and Electrophysiology databases. We excluded 125 subsequent TEEs on the same patients (only the initial TEE was used for analysis when a patient underwent TEE more than once during study period) and 25 patients who did not have available CHADS2score data for analysis. A total of 1,058 patients were included in the study. The study was approved by the Cleveland Clinic Institutional Review Board.
A retrospective medical record review was performed by an observer blinded to echocardiographic findings. The clinical data were reviewed beginning 3 months before TEE. A CHADS2score (7,8) ranging from 0 to 6 was calculated for each patient at the time of TEE as: congestive heart failure (CHF) (1 point); hypertension (1 point); age >75 years (1 point); and history of stroke, transient ischemic attack (TIA), or systemic embolism (2 points). A low thromboembolic risk was defined as a CHADS2score of 0 to 1. A history of coronary artery disease, thyrotoxicosis, obstructive sleep apnea, chronic obstructive pulmonary disease, cardiomyopathy, prior ablative procedure or direct current cardioversion (DCC), the use of systemic anticoagulation, the cardiac rhythm at the time of PVI and the findings on TEE in the case that the PVI was cancelled or deferred, and the classification of AF were reviewed. Paroxysmal atrial fibrillation (PAF) was defined as recurrent AF (≥2 episodes) that terminates spontaneously within 7 days (6,9). Persistent AF was defined as AF that is sustained beyond 7 days, or lasting <7 days but necessitating pharmacologic or electrical cardioversion (6,9). Persistent AF also included longstanding persistent AF (continuous AF of >1-year duration). Permanent AF was defined as AF in which cardioversion has either failed or not been attempted (6,9). In patients who had AF episodes that fell into 1 or more of these categories, their most frequent pattern of AF was used.
TEE was performed using commercially available equipment equipped with a multiplane phase array transducer. All TEEs were performed within 24 h before the PVI procedure, including procedures cancelled due to LA/LAA thrombus or sludge. All patients gave written consent before TEE. Lidocaine was used for local anesthesia of hypopharynx. Conscious sedation with intravenous midazolam and/or fentanyl was given in all patients. Standard esophageal intubation and TEE acquisition were conducted (10). Cine loop of LA and LAA from 0° to 180° was stored for offline analysis.
Echocardiographic data analysis
TEEs were reviewed for the presence of LA/LAA, spontaneous echo contrast (SEC), sludge, and thrombus by 2 observers who were blinded to clinical history. Interobserver differences were resolved by a third observer.
SEC was defined as dynamic “smoke-like” echoes with the characteristic swirling motion with optimal gain setting during cardiac cycle (11,12). Sludge was defined as a dynamic gelatinous, precipitous echodensity, without a discrete mass, present throughout the cardiac cycle (11). Thrombus was defined as circumscribed and uniformly echodense intracavitary mass distinct from the underlying LA or LAA endocardium and the pectinate muscles, and present in more than 1 imaging plane (11,13).
Anticoagulation and catheter ablative procedure
In our institution, we routinely perform this procedure while patients are fully anticoagulated, with a therapeutic international normalized ratio (INR) between 2 and 3.5 (14–16). INRs are usually checked every 3 weeks for at least 3 months before the PVI. In the 3 weeks just before the PVI, the INRs are checked on a weekly basis. At our institution, patients are almost always anticoagulated for at least 3 months before the PVI; however, the anticoagulant regimen has changed over the course of the 7-year study period. From January 1999 to December 2005, anticoagulation was usually discontinued for 3 to 4 days before the PVI. Since 2006, anticoagulation was continued until the time of the PVI. The TEEs are done during the day of the PVI. Patients with chronic AF must have therapeutic INRs for at least 2 to 3 months before the procedure. Heparin bolus (100 to 150 U/kg) is always given before transseptal punctures. The patient is maintained on a heparin infusion rate of 15 to 20 U/kg/h (14,15) and the infusion rate is adjusted to keep the activated clotting time in the range of 350 to 450 s. The PVI procedure was performed according to a standard measure as previously described elsewhere (3,4,6,17,18). There is no consensus in the management of LA/LAA sludge. The decision of whether the PVI procedure should be further continued, deferred, or cancelled after a screening TEE was individually based on the judgment of the electrophysiologist.
The frequency and percentage of TEE characteristics of the thrombogenic milieu (SEC, sludge, thrombus) and mean ± SD of LAA emptying velocities and left ventricular ejection fraction (LVEF) were reported in each CHADS2score and in each classification of AF (PAF with sinus rhythm at the time of PVI, PAF with AF at the time of PVI, and persistent/permanent AF). TEE characteristics and continuous variables were compared across level of CHADS2score and 3 AF classification groups using the chi-square or Fisher exact test as appropriate and analysis of variance test, respectively. Multivariable logistic regression was performed to identify independent clinical predictors of LA/LAA thrombus and sludge. The variables tested in the logistic model include CHADS2score ≥2, persistent/permanent AF, hypertension, prior stroke/TIA/systemic emboli, LVEF <35%, CHF, diabetes mellitus, age >75 years, AF at the time of PVI, previous cardioversion, previous PVI, coronary artery disease, and INR <2. Statistical significance was defined as p < 0.05.
To determine the intra- and interobserver variation, echocardiographic data (control data [n = 5], SEC [n = 5], sludge [n = 5], and thrombus [n = 5]) from a nested case group of 20 randomly selected patients were analyzed by the same observer at 2 points in time at least 4 weeks apart in a blinded fashion. The same nested sample was reviewed in a similar manner on 2 occasions by a second observer. A kappa statistic analysis was used to describe agreement.
Patient characteristics are shown in Table 1.The mean age of participants was 57 ± 11 years; 80% were men. Most (81%) had PAF, and 73% were anticoagulated at the time of PVI/TEE. Impaired left ventricular function (LVEF <35%) was present in 9% of patients. The majority of patients (80%) were in a low-risk group as CHADS2scores of 0 and 1 were present in 47% and 33%, respectively. Of studied patients, 1% (n = 10) had mechanical prosthesis (scores 0 [15%], 1 [23%], 2 [23%], 3 [31%], and 4 to 6 [8%]), and 0.3% (n = 3) had rheumatic heart diseases (scores 0 [0%], 1 [33%], 2 [33%], 3 [0%], and 4 to 6 [33%]).
Prevalence of LA/LAA thrombus or sludge
Among 1,058 patients who underwent screening TEE before PVI, the prevalence of LA/LAA thrombus and sludge were only 0.6% (n = 6) and 1.5% (n = 16), respectively. The PVI procedures were cancelled in 6 and 7 patients due to LA/LAA thrombus and sludge, respectively.
Relationship of CHADS2score and TEE features of a thrombogenic milieu
As demonstrated in Table 2,the prevalence of thrombus/sludge increased with ascending CHADS2scores (scores 0 [0%], 1 [2%], 2 [5%], 3 [9%], and 4 to 6 [11%], p < 0.01). Peak LAA emptying velocities were also significantly reduced with ascending CHADS2score. No patients with a CHADS2score of 0, irrespective of LVEF and classification of AF, had LA/LAA thrombus or sludge. LA/LAA thrombus or sludge was present in 8 patients (2%) with a CHADS2score of 1 who were presumed to be at a relatively low risk of stroke (Fig. 1).Table 3details the individual clinical and echocardiographic characteristics in those low-risk patients with a CHADS2score of 1 who had LA/LAA sludge or thrombus identified on TEE. Four of these patients (50%) had impaired LVEF, and 7 of these patients (88%) were in AF at the time of PVI. None of these patients had clinical risk factors of diabetes or prior stroke/TIA. Figures 2 and 3⇓illustrate LA/LAA findings of sludge and thrombus in patients with CHADS2scores of 1 (Patients #4 and #7) (Table 3).
Relationship of CHADS2score and classification of AF with TEE features of a thrombogenic milieu
As indicated in Table 4,LA/LAA thrombus and sludge were present in 1% of patients with PAF with normal sinus rhythm at the time of PVI.
Independent clinical predictors of LA/LAA thrombus or sludge
In the multivariable model, a history of CHF and LVEF <35% was only significantly associated with LA/LAA thrombus/sludge (Table 5).
Follow-up thromboembolic events
In our study, the patients were anticoagulated post-PVI for at least 3 months. Follow-up data were available in 1,039 of 1,058 patients (98%) with a mean follow-up duration of 60.9 ± 53.7 months. Of 1,039 patients, 19 (1.8%) and 2 (0.02%) patients had ischemic stroke and TIA, respectively. One of 22 (5%) patients with LA/LAA thrombus/sludge at the time of PVI/TEE developed post-PVI thromboembolic stroke whereas 21 of 1,017 (2%) patients without LA/LAA thrombus/sludge developed post-PVI thromboembolic stroke; however, this was not statistically significant (p = 0.48).
To determine the intraobserver and interobserver variation, echocardiographic data (control data [n = 5], SEC [n = 5], sludge [n = 5], and thrombus [n = 5]) from a nested case group of 20 randomly selected patients were analyzed by the same observer at 2 points in time at least 4 weeks apart in a blinded fashion. The same nested sample was reviewed in a similar manner on 2 occasions by a second observer. The concordance between 2 observers in identification of LA/LAA characteristics of thrombogenic milieu was high (overall kappa = 0.933). No discordance between thrombus and sludge was demonstrated. The intraobserver concordance was very high as well (kappa = 1.00).
The major findings of our study are: 1) the yield of LA/LAA thrombus/sludge identified on TEE in patients with AF referred for pre-PVI TEE screening was very low (0.6% for thrombus and 1.5% for sludge); 2) LA/LAA thrombus or sludge was present in only patients with at least 1 clinical risk factor(s) (CHADS2score of ≥1); 3) the prevalence of LA/LAA thrombus/sludge significantly increased with increasing CHADS2score; and 4) in addition to CHADS2score, a history of CHF and LVEF <35% were independent predictors of the presence of LA/LAA thrombus/sludge in these patients. The implication is that TEEs are not needed to assess stroke risk in patients with a CHADS2score of 0 when the therapeutic anticoagulation has been maintained, but they may demonstrate echocardiographic findings associated with increased risk of stroke in patients with a CHADS2score ≥1.
Prevalence of LA/LAA thrombus or sludge
The presence of LA/LAA thrombus is associated with an increased risk of thromboembolic events (11,19,20). Our findings of the low prevalence of pre-procedural LA/LAA thrombus (0.6%) agreed with the previous report by Khan et al. (21) who studied 601 patients with AF at our institution and showed that the prevalence of LA/LAA thrombus was <1%. However, our study population was larger. Khan et al. (21) did not describe whether they excluded patients with subsequent TEEs on the same patients. In contrast to our study, they included all patients with TEE regardless of available CHADS2score data for analysis. Furthermore, we have reassessed all TEE studies blinded to CHADS2score whereas their study used TEE reports from the medical records. Additionally, our database includes the TEE patients regardless of computed tomography (CT) scan, but their database includes the CT patients andthe patients who had both TEE/CT. Champagne et al. (22) reported that LA thrombus was present in only 1 patient among 85 (1%) that underwent TEE-guided AF ablation. However, compared with patients with AF undergoing TEE before DCC, the prevalence of LA/LAA thrombus in our study population was far less frequent (12% vs. 0.6%) (20,23). A potential explanation for the low prevalence of LA/LAA thrombus in our study was the fact that the majority of studied patients (80%) were in a low-risk (CHADS2score ≤1) group and were anticoagulated (73%).
The clinical significance of LA/LAA “sludge” (11,24) has not been well described. Our investigators have previously reported that the mortality and thromboembolic risk in AF patients with sludge were comparable to those with thrombus and could be considered as a contraindication to DCC and PVI (25). Similar to the prevalence of LA/LAA thrombus, the prevalence of sludge was low (1.5%). Our findings suggest that a more selective TEE approach to target high-risk patients could be implemented.
Selective TEE approach stratified by a CHADS2score
The CHADS2score, a clinical scoring system quantifying a risk of stroke in patients with AF, was recently devised and validated in 1,733 Medicare beneficiaries with nonvalvular AF (7) and in 51,807 general practice research database patients with chronic AF in the United Kingdom (8), respectively. The CHADS2scoring scheme was found to be a good predictor of stroke risk, with higher scores associated with an increased stroke rate (7,8). In our cohort, it was interesting that no patients with a CHADS2score of 0 had LA/LAA thrombus or sludge identified on TEE. However, our findings were in contrast to those of Rader et al. (26) and Zabalgoitia et al. (27). Rader et al. (26) previously reported that LA/LAA thrombus was identified in 2 patients with nonvalvular AF and a CHADS2score of 0 (70 and 74 years old; only 1 of them received anticoagulation). It was noted that overall, fewer patients in their study were anticoagulated (69% vs. 73%) and the prevalence of clinical risk factors (heart failure, hypertension, diabetes, stroke, and impaired left ventricular function) was higher than those in our study. Additionally, they excluded patients in sinus rhythm at the time of TEE, and their study did not target patients undergoing PVI. These factors may have affected the characteristics of the thrombogenic milieu resulting in different findings. Similar to the study of Rader et al. (26), Zabalgoitia et al. (27) found that the prevalence of LA/LAA thrombus in patients with AF was 3% in a low-risk group defined as an absence of all the following clinical risk factors: women over age 75 years, hypertension, heart failure or fractional shortening <25%, and prior thromboembolism. In their study, 38% of patients in a low-risk group were anticoagulated and 28% had a history of PAF, whereas over 50% of patients with a CHADS2score of 0 in our study were anticoagulated and had a history of PAF. Although there is conflicting evidence, it was clearly shown that, in a study population of AF patients undergoing TEE before PVI, no patients with a CHADS2score of 0 had LA/LAA thrombus or sludge. Furthermore, our study population was in a lower-risk group than those undergoing DCC. Our study suggests that PVI may be performed safely without TEE in patients with a CHADS2score of 0 who received periprocedural warfarin therapy. Patients with PAF who are in sinus rhythm at the time of PVI appear to be at lower risk, but TEE should still be considered with a CHADS2score ≥1.
The recent Heart Rhythm Society Task Force document found no consensus regarding the need to perform a TEE screening before PVI in patients with PAF who are in sinus rhythm at the time of ablation (6). In our study, 1 patient (0.5%) with a history of PAF who was in sinus rhythm at the time of PVI had LA/LAA sludge identified on a screening TEE before PVI. However, this patient had warfarin discontinued before the procedure. Furthermore, 1 patient with PAF who was in sinus rhythm at the time of TEE was found to have LA/LAA thrombus. The PVI procedure was cancelled due to LAA thrombus. The INR on the day of TEE was 1.8. This illustrates that using a clinical history of PAF and rhythm at the time of PVI/TEE alone may not completely exclude the thrombogenic state of LAA sludge and thrombus. Rather, combining a CHADS2score as well as adequacy of anticoagulation with a history of PAF and sinus rhythm at time of PVI is more useful in mitigating a need of pre-PVI TEE screening.
Relationship of CHADS2score and TEE features of the thrombogenic milieu
Previous studies have shown that the prevalence of LA/LAA thrombus increased with an increasing prevalence of clinical risk factors (11,27–31) and a higher CHADS2score (26). Our findings agreed with those studies. The most powerful predictors of LA/LAA sludge/thrombus in our study included impaired left ventricular function and a history of CHF, which were in agreement with other studies (11,26,27,29–31) despite different study populations. In our study, patients with LA/LAA sludge or thrombus had at least 1 clinical risk factor of a CHADS2score. Thus, those patients with at least 1 clinical risk factor of CHADS2score as well as a history of CHF and LVEF <35% should receive a screening TEE to exclude LA/LAA sludge and thrombus before PVI.
Relationship of TEE features of the thrombogenic milieu and subsequent thromboembolic events
There have been conflicting data pertaining to the association between LA thrombus and thromboembolic stroke/events. In our study, we found that patients with LA thrombus/sludge on TEE had insignificantly higher subsequent thromboembolic events; however, our study was underpowered for stroke. Our findings were concordant with the study by Archer et al. (32). They also did not find the significant association between LA thrombus and embolic events (stroke/TIA and peripheral embolism) in patients with chronic nonrheumatic valvular AF. They reported stroke in 1 of 5 (20%) subjects with LA thrombus compared with an embolic event in 6 of 50 (12%) patients without LA thrombus during a mean follow-up duration of 34 months. The study was underpowered to reach statistical significance, and LA thrombus was not assessed by multivariate analysis as an independent predictor of thromboembolic events. In a TEE substudy of the Stroke Prevention in Atrial Fibrillation-III trial, a multivariate analysis examining TEE parameters failed to demonstrate an independent predictor of LA thrombus for embolic events in 192 patients with nonvalvular AF (33). The prevalence of LA thrombus in this study was 14%. However, in this trial, univariate analysis showed an association between embolism and LA thrombus. Furthermore, Stollberger et al. (34) showed that LA thrombus insignificantly (p = 0.09) predicts stroke or peripheral embolism in patients with AF. In this study, the incidence of LA thrombus was 2.5% (34). However, Stoddard et al. (35) recently demonstrated that LA thrombus significantly predicts TIA in 261 patients with AF during a mean follow-up duration of 30 months. The incidence of LA thrombus in their study was 18%.
This was a retrospective study. The definition of sludge was relatively subjective; nevertheless, interobserver variability in our study was very low among those TEE characteristics. We did not grade SEC as mild or severe. TEEs were performed by various echocardiographers; however, they were all specialized and experienced in TEE. A CHADS2score has not been validated in patients with AF undergoing pre-PVI TEE, although it was useful in excluding LA/LAA thrombus and sludge in those with a CHADS2score of 0. The data for the adequacy of anticoagulation were limited in this study and may have affected the features of the thrombogenic milieu. There was heterogeneity of aggressiveness of anticoagulation and/or target-achieved INR before ablation. These were primarily based on the judgment of each electrophysiologist. Clinical usefulness of the CHADS2score in patients with AF undergoing PVI in this study was tested in patients treated as they were at our site. This needs to be further tested in a similar patient population treated with a different anticoagulation protocol.
Our study shows that the prevalence of LA/LAA sludge/thrombus in patients with AF undergoing a pre-PVI screening TEE is very low (0.6% for thrombus and 1.5% for sludge) and significantly increases in patients with higher CHADS2scores. This suggests that a more selective TEE approach to target high-risk patients may be beneficial. The results of our study demonstrate that a screening TEE before PVI should be performed in patients with a CHADS2score ≥1 regardless of the rhythm at the time of the procedure, and in patients with a CHADS2score of 0 when the therapeutic anticoagulation has not been maintained for 4 weeks before the procedure.
Dr. Wazni serves as a consultant for Biosense Webster.
- Abbreviations and Acronyms
- atrial fibrillation
- congestive heart failure
- computed tomography
- direct current cardioversion
- international normalized ratio
- left atrial
- left atrial appendage
- left ventricular ejection fraction
- paroxysmal atrial fibrillation
- pulmonary vein isolation
- spontaneous echo contrast
- transesophageal echocardiogram
- transient ischemic attack
- Received June 16, 2009.
- Accepted July 8, 2009.
- American College of Cardiology Foundation
- Oral H.,
- Knight B.P.,
- Tada H.,
- et al.
- Josephson M.E.
- Oral H.,
- Chugh A.,
- Ozaydin M.,
- et al.
- Calkins H.,
- Brugada J.,
- Packer D.L.,
- et al.
- Rietbrock S.,
- Heeley E.,
- Plumb J.,
- van Staa T.
- Fuster V.,
- Ryden L.E.,
- Asinger R.W.,
- et al.
- Shanewise J.S.,
- Cheung A.T.,
- Aronson S.,
- et al.
- Wazni O.M.,
- Tsao H.M.,
- Chen S.A.,
- et al.
- Fatkin D.,
- Kelly R.P.,
- Feneley M.P.
- Aschenberg W.,
- Schluter M.,
- Kremer P.,
- Schroder E.,
- Siglow V.,
- Bleifeld W.
- Wazni O.M.,
- Beheiry S.,
- Fahmy T.,
- et al.
- Gill J.S.
- Klein A.L.,
- Gabriel R.S.
- Lowe B.S.,
- Varr B.,
- Shrestha K.,
- Whitman C.,
- Klein A.L.
- Zabalgoitia M.,
- Halperin J.L.,
- Pearce L.A.,
- Blackshear J.L.,
- Asinger R.W.,
- Hart R.G.