By Michael H. Crawford, MD, Editor

SYNOPSIS: A study of patients with atrial fibrillation or flutter undergoing transesophageal echocardiography (TEE) showed a left atrial appendage thrombus in 8% overall and 4% in those appropriately anticoagulated. Repeat TEE after anticoagulation for a mean of 96 days showed resolution of thrombus in only 59% of subjects.

SOURCE: Niku AD, Shiota T, Siegel RJ, Rader F. Prevalence and resolution of left atrial thrombus in patients with nonvalvular atrial fibrillation and flutter with oral anticoagulation. Am J Cardiol 2019;123:63-68.

Embolic stroke after electric cardioversion of atrial fibrillation (AF) or flutter (AFI) is a catastrophic complication that potentially can be prevented by warfarin or direct oral anticoagulant (DOAC) therapy. Often, transesophageal echocardiography (TEE) is used to assess for left atrial appendage (LAA) thrombus before cardioversion, even if patients have been on warfarin or DOAC therapy. However, there are few data to support this practice.

Investigators from Cedars-Sinai Medical Center in Los Angeles performed a retrospective observational study of more than 2,000 TEE studies performed on 1,485 patients with AF or AFI between 2013 and 2017 to gain insight into the utility of TEE prior to cardioversion. LAA thrombus included discrete masses and viscid echo densities distinct from spontaneous echo contrast. Patients with LAA occluder or past LAA ligation were excluded. The initial TEE in the 1,485 patients exhibited LAA thrombus in 117. Among these 117 patients, 61 were on antiplatelet therapy, and all but one were on an anticoagulant. Fifty-six patients were considered adequately anticoagulated by taking warfarin with an international normalized ratio (INR) > 2.0 or on a DOAC. The thrombus rate in this group was 4%. In 63 of 117 patients with LAA thrombus, a repeat TEE was performed within one year (mean, 96 days) of being on continuous anticoagulant therapy. Thrombus resolution was observed in 37 of those 63 patients. The only clinical variable associated with LAA thrombus persistence on anticoagulation was the presence of diabetes. Resolution rates were not influenced by the configuration of the LAA. Resolution did not vary by anticoagulation type. Among the 117 patients with LAA thrombus, eight suffered a thromboembolic event, and only two were noncompliant with oral anticoagulants. The authors concluded that LAA thrombus persistence on anticoagulant therapy is common and not clinically predictable. Thus, they recommended TEE prior to cardioversion in all patients with a history of LAA thrombus, regardless of anticoagulation therapy.


Increasingly, my institution’s electrophysiology section is requesting TEEs on all patients who are undergoing electrical or chemical cardioversions or who might cardiovert in association with therapy for AF or AFI, despite taking (presumably) adequate anticoagulant therapy. However, very few of the TEEs performed for this purpose reveal LAA thrombus. Thus, this study caught my attention. This analysis of the authors’ five-year experience with TEEs conducted on patients with AF or AFI reached several conclusions: Eight percent of the patients had LAA thrombus, yet almost all were on oral anticoagulants. In those thought to be appropriately anticoagulated, it still was 4%. Forty percent of those with LAA thrombus who underwent a repeat TEE after about three months of appropriate anticoagulant therapy still showed LAA thrombus. There was no difference in thrombus incidence or resolution between warfarin or DOAC therapy. The only clinical predictor of persistent LAA thrombus was diabetes.

It is common practice to perform cardioversion of AF/AFI without a TEE if the patient has been on adequate oral anticoagulation for more than three weeks. This study suggests this practice runs the risk of cardioverting four out of 100 patients with a thrombus. Whether these thrombi would embolize and hit a vital organ is unknown. Other studies of TEE prior to cardioversion have shown embolism rates in those with negative TEEs of 1-2%. The risk of TEE is minimal in well-selected patients, but there are patients in whom the risk of TEE is competitive with these embolism rates. Thus, if almost complete avoidance of thromboembolism with cardioversion is the goal, then a routine TEE regardless of anticoagulant status is reasonable. The authors noted that the ascertainment of who is compliant with DOAC therapy is challenging since there is no INR-type test for these agents. In particular, I worry about patients on DOACs that have to be taken twice a day (dabigatran and apixaban).

There are limitations to this study. There could be a selection bias toward performing TEEs on higher-risk patients. Most patients recorded CHA2DS2-VASc scores of ≥ 3. Three-dimensional echo was not performed routinely, so the sensitivity for detecting thrombi may be less if 3D echo was not conducted. No robust analysis of medication noncompliance beyond INR was performed. Also, there were no data on the persistence or recurrence of AF/AFI and how this may have affected the results. Finally, the number of patients in this one center study was relatively small.

Clearly, patients with known prior LAA thrombus should undergo TEE-guided cardioversion based on the poor resolution rates reported in this study. Whether everyone with AF/AFI undergoing elective cardioversion or any procedure or medication that might lead to cardioversion needs TEE is uncertain. Still, the data in this study support the trend in this direction that I have observed.