Patient Population and Study Design

We prospectively enrolled consecutive eligible and consenting adult patients with NVAF who were referred to undergo TEE to “rule out” LAA thrombus before procedures involving the restoration of sinus rhythm or LA instrumentation or for excluding LAA thrombus as a source of systemic embolization. Patients were enrolled from the echocardiography laboratories of Rush University Medical Center (Chicago, IL) and Advocate Illinois Masonic Medical Center (Chicago, IL) in the period between January 7, 2011, and November 14, 2013. We excluded patients with mitral stenosis, mitral regurgitation >2+ in severity (scale, 0-4), any mitral valve surgical or percutaneous intervention, complex congenital heart disease, and orthotopic heart transplantation. Patients with mitral annular calcification without evidence of mitral stenosis, isolated aortic valvular disease, aortic valve prosthesis, and right-sided valvular heart disease were not excluded. Patients receiving parenteral or oral anticoagulants were not excluded. All subjects provided informed and Health Insurance Portability and Accountability Act consent before enrollment. The study was funded by an internal research grant and approved by the institutional review board of each participating institution.

Clinical Data

Data including patient demographics, comorbidities, anticoagulation status, and antiplatelet use were prospectively collected before TEE. Atrial fibrillation chronicity was classified as follows: (1) paroxysmal, defined as spontaneously reverting to sinus rhythm within 7 days; (2) persistent, defined as lasting >7 days but <12 months, or any atrial fibrillation events terminated by electrical or chemical cardioversion or radiofrequency ablation within 12 months of onset; or (3) permanent, defined as lasting >12 months. The CHADS ₂ score was calculated from the sum of risk predictors of congestive heart failure, hypertension, age ≥ 75 years, diabetes mellitus, and stroke or transient ischemic attack, weighing each risk predictor by 1 except for prior stroke or transient ischemic attack, which was weighed by 2. The CHA ₂ DS ₂ -VASc score was calculated from the sum of the risk factors of congestive heart failure, hypertension, age 65 to 74 or ≥75 years, diabetes mellitus, stroke or transient ischemic attack, vascular disease, and female gender; weighing each by 1 except for stroke or transient ischemic attack and age ≥75 years, which were weighed by 2. Congestive heart failure was defined as a history of clinical heart failure or LVEF < 40%. Vascular disease was defined as peripheral arterial disease, complex aortic plaque, or prior myocardial infarction.

TTE

Immediately before the induction of sedation for TEE, investigational TTE was performed by a trained sonographer. All transthoracic echocardiograms were reviewed offline by a board-certified echocardiographer who was blinded to the transesophageal and clinical data. The mitral inflow early diastolic velocity (E) and the septal and lateral mitral annular spectral tissue Doppler early diastolic velocities (e′) were measured ( Figure 1 ). All Doppler measurements were obtained by averaging data from three to five consecutive beats. E/e′ ratios were calculated from the septal and lateral e′ velocities, and an average E/e′ value was calculated. Unless otherwise specified, all analyses presented hereafter are related to lateral e′ velocity and lateral E/e′ ratio. Findings were verified using septal and averaged e′ velocity and E/e′ values.

Representative examples. (A) Pulsed-wave Doppler recording of early diastolic mitral inflow velocity (E). (B) Spectral tissue Doppler imaging of lateral mitral annular early diastolic velocity (e′). (C) LAA thrombus (arrow). (D) LA SEC.

Reproduced with permission from Doukky et al .

The left ventricular septal and posterior wall thickness and end-diastolic as well as end-systolic internal dimensions were measured from two-dimensional or M-mode transthoracic echocardiographic images. Left ventricular mass was calculated using the Devereux formula. Left ventricular end-systolic and end-diastolic volumes and LVEF were measured using the biplane Simpson method. LA volume was measured using the single-plane Simpson method in the apical four-chamber view. All volume and mass measurements were indexed to body surface area.

TEE

TEE was performed under moderate sedation according to standard protocol. An expert National Board of Echocardiography–certified echocardiographer, who was blinded to the transthoracic and clinical data, reviewed all transesophageal images to determine the presence or absence of LAA thrombus, defined as a circumscribed and uniformly echodense intracavitary mass distinct from the underlying LA or LAA endocardium and the pectinate muscles and present in more than one imaging plane ( Figure 1 ). LAA sludge, defined as a dynamic gelatinous, precipitous echodensity, without a discrete mass, present throughout the cardiac cycle was categorized as LAA thrombus. SEC was defined as dynamic “smokelike” echoes with the characteristic swirling motion with optimal gain setting during the entire cardiac cycle ( Figure 1 ) and was classified from 1 to 4 on the basis of the criteria described by Fatkin et al . Peak LAA emptying velocity at the LAA orifice was measured using pulsed-wave Doppler and averaged over three to five beats.

The primary outcome of the study was TEE-identified LAA thrombus. The secondary outcome was TEE-identified SEC.

Statistical Analysis

The two-tailed Student’s t test was used to compare normally distributed continuous variables, which are expressed as mean ± SD. The Mann-Whitney U test was used to compare skewed data. The χ ² test was used to compare categorical variables, which are expressed as frequencies and percentages. The Spearman method was used to evaluate linear correlations.

Multivariate logistic regression models were fit to determine the predictive value of diastolic function parameters adjusted for known clinical predictors (CHA ₂ DS ₂ -VASc score). Risk was expressed as odds ratios (ORs) with 95% CIs. Goodness of fit for multivariate models was confirmed using the Hosmer-Lemeshow test. To avoid model overfitting, the number of covariates included in multivariate logistic regression models was limited to approximately one covariate for every 10 outcome events.

The receiver operating characteristic (ROC) methodology was used to analyze the discriminatory capacity of clinical and echocardiographic predictors of LAA thrombus. ROC analyses were expressed as curve plots and calculated area under the curve (AUC) with CI and associated P value representing the likelihood of the null hypothesis (indicated by an AUC of 0.5). The χ ² test was used to compare the areas under two ROC curves. All tests were two tailed, and P values < .05 were considered to indicate statistical significance. SPSS 22 (IBM, Armonk, NY) was used for all data analyses.

Patient Population and Study Design

We prospectively enrolled consecutive eligible and consenting adult patients with NVAF who were referred to undergo TEE to “rule out” LAA thrombus before procedures involving the restoration of sinus rhythm or LA instrumentation or for excluding LAA thrombus as a source of systemic embolization. Patients were enrolled from the echocardiography laboratories of Rush University Medical Center (Chicago, IL) and Advocate Illinois Masonic Medical Center (Chicago, IL) in the period between January 7, 2011, and November 14, 2013. We excluded patients with mitral stenosis, mitral regurgitation >2+ in severity (scale, 0-4), any mitral valve surgical or percutaneous intervention, complex congenital heart disease, and orthotopic heart transplantation. Patients with mitral annular calcification without evidence of mitral stenosis, isolated aortic valvular disease, aortic valve prosthesis, and right-sided valvular heart disease were not excluded. Patients receiving parenteral or oral anticoagulants were not excluded. All subjects provided informed and Health Insurance Portability and Accountability Act consent before enrollment. The study was funded by an internal research grant and approved by the institutional review board of each participating institution.

Clinical Data

Data including patient demographics, comorbidities, anticoagulation status, and antiplatelet use were prospectively collected before TEE. Atrial fibrillation chronicity was classified as follows: (1) paroxysmal, defined as spontaneously reverting to sinus rhythm within 7 days; (2) persistent, defined as lasting >7 days but <12 months, or any atrial fibrillation events terminated by electrical or chemical cardioversion or radiofrequency ablation within 12 months of onset; or (3) permanent, defined as lasting >12 months. The CHADS ₂ score was calculated from the sum of risk predictors of congestive heart failure, hypertension, age ≥ 75 years, diabetes mellitus, and stroke or transient ischemic attack, weighing each risk predictor by 1 except for prior stroke or transient ischemic attack, which was weighed by 2. The CHA ₂ DS ₂ -VASc score was calculated from the sum of the risk factors of congestive heart failure, hypertension, age 65 to 74 or ≥75 years, diabetes mellitus, stroke or transient ischemic attack, vascular disease, and female gender; weighing each by 1 except for stroke or transient ischemic attack and age ≥75 years, which were weighed by 2. Congestive heart failure was defined as a history of clinical heart failure or LVEF < 40%. Vascular disease was defined as peripheral arterial disease, complex aortic plaque, or prior myocardial infarction.

TTE

Immediately before the induction of sedation for TEE, investigational TTE was performed by a trained sonographer. All transthoracic echocardiograms were reviewed offline by a board-certified echocardiographer who was blinded to the transesophageal and clinical data. The mitral inflow early diastolic velocity (E) and the septal and lateral mitral annular spectral tissue Doppler early diastolic velocities (e′) were measured ( Figure 1 ). All Doppler measurements were obtained by averaging data from three to five consecutive beats. E/e′ ratios were calculated from the septal and lateral e′ velocities, and an average E/e′ value was calculated. Unless otherwise specified, all analyses presented hereafter are related to lateral e′ velocity and lateral E/e′ ratio. Findings were verified using septal and averaged e′ velocity and E/e′ values.

Representative examples. (A) Pulsed-wave Doppler recording of early diastolic mitral inflow velocity (E). (B) Spectral tissue Doppler imaging of lateral mitral annular early diastolic velocity (e′). (C) LAA thrombus (arrow). (D) LA SEC.

Reproduced with permission from Doukky et al .

The left ventricular septal and posterior wall thickness and end-diastolic as well as end-systolic internal dimensions were measured from two-dimensional or M-mode transthoracic echocardiographic images. Left ventricular mass was calculated using the Devereux formula. Left ventricular end-systolic and end-diastolic volumes and LVEF were measured using the biplane Simpson method. LA volume was measured using the single-plane Simpson method in the apical four-chamber view. All volume and mass measurements were indexed to body surface area.

TEE

TEE was performed under moderate sedation according to standard protocol. An expert National Board of Echocardiography–certified echocardiographer, who was blinded to the transthoracic and clinical data, reviewed all transesophageal images to determine the presence or absence of LAA thrombus, defined as a circumscribed and uniformly echodense intracavitary mass distinct from the underlying LA or LAA endocardium and the pectinate muscles and present in more than one imaging plane ( Figure 1 ). LAA sludge, defined as a dynamic gelatinous, precipitous echodensity, without a discrete mass, present throughout the cardiac cycle was categorized as LAA thrombus. SEC was defined as dynamic “smokelike” echoes with the characteristic swirling motion with optimal gain setting during the entire cardiac cycle ( Figure 1 ) and was classified from 1 to 4 on the basis of the criteria described by Fatkin et al . Peak LAA emptying velocity at the LAA orifice was measured using pulsed-wave Doppler and averaged over three to five beats.

The primary outcome of the study was TEE-identified LAA thrombus. The secondary outcome was TEE-identified SEC.

Statistical Analysis

The two-tailed Student’s t test was used to compare normally distributed continuous variables, which are expressed as mean ± SD. The Mann-Whitney U test was used to compare skewed data. The χ ² test was used to compare categorical variables, which are expressed as frequencies and percentages. The Spearman method was used to evaluate linear correlations.

Multivariate logistic regression models were fit to determine the predictive value of diastolic function parameters adjusted for known clinical predictors (CHA ₂ DS ₂ -VASc score). Risk was expressed as odds ratios (ORs) with 95% CIs. Goodness of fit for multivariate models was confirmed using the Hosmer-Lemeshow test. To avoid model overfitting, the number of covariates included in multivariate logistic regression models was limited to approximately one covariate for every 10 outcome events.

The receiver operating characteristic (ROC) methodology was used to analyze the discriminatory capacity of clinical and echocardiographic predictors of LAA thrombus. ROC analyses were expressed as curve plots and calculated area under the curve (AUC) with CI and associated P value representing the likelihood of the null hypothesis (indicated by an AUC of 0.5). The χ ² test was used to compare the areas under two ROC curves. All tests were two tailed, and P values < .05 were considered to indicate statistical significance. SPSS 22 (IBM, Armonk, NY) was used for all data analyses.