By Michael H. Crawford, MD, Editor

SYNOPSIS: A large, multicenter, observational study of the relative efficacy of warfarin vs. direct oral anticoagulants (DOACs) for left ventricular thrombi has shown that DOAC use is associated with a higher risk of embolic events than warfarin. Investigators advised caution with off-label use of DOACs.

SOURCE: Robinson AA, Trankle CR, Eubanks G, et al. Off-label use of direct oral anticoagulants compared with warfarin for left ventricular thrombi. JAMA Cardiol 2020; Apr 22. doi: 10.1001/jamacardio.2020.0652. [Epub ahead of print].

A high degree of patient acceptance has driven off-label use of direct oral anticoagulants (DOACs) for the treatment of left ventricular thrombi (LVT), despite a paucity of supportive data.

In three tertiary care, academic medical centers in Virginia and North Carolina, Robinson et al organized the Retrospective Evaluation of DOACs and Vascular Endpoints of Left Ventricular Thrombi (RED VELVT) observational study. From 2013 to 2019, patients with echocardiographically diagnosed LVT were identified and followed for a median of 351 days (range, 51-866 days). The resulting population consisted of 514 patients (74% men, mean age 58 years), of whom 300 were treated with warfarin and 185 with a direct oral anticoagulant (DOAC). These groups included a mixed group of 64 patients who switched treatment such that there were 236 patients exclusively treated with warfarin and 121 with a DOAC. Most switches were from warfarin to DOAC. Ninety-three patients received no oral anticoagulants, including 43 who received no anticoagulant, oral or parenteral.

There were 54 embolic events: 36 strokes and 18 systemic emboli. After a multivariate analysis, anticoagulant type was significantly associated with embolic events (hazard ratio [HR] for DOAC vs. warfarin, 2.64; 95% confidence interval [CI], 1.28-5.43; P = 0.01), as was a history of prior emboli (HR, 2.07; 95% CI, 1.17-3.66; P = 0.01). Interestingly, neither thrombus size nor mobility were associated with embolic events. Censoring the follow-up data at three, six, and 12 months did not alter the results. The authors concluded that for the treatment of LVT, DOAC use was associated with a higher risk of embolic events than warfarin, even after adjusting for other clinical factors.


This was an important study. In three large, academic medical centers over a six-year period, 44% of patients with LVT were treated with DOACs. At the time, these were perhaps reasonable clinical decisions. Patients do not like warfarin, and DOACs have been highly efficacious for preventing embolic events in patients with atrial fibrillation and deep venous thrombosis. Surely, LVT must be a similar blood stasis-related condition. Also, there were several small, observational studies that suggested DOACs were efficacious for LVT. However, those were single-center investigations, with small patient cohorts, shorter follow-up periods, and few embolic events. The Robinson et al study was multicenter, with more than 500 patients and 54 embolic events. Also, patients with LVT actually were contacted; the authors did not rely on chart review alone.

Why these remarkably different results? There are several possible explanations. First, DOACs may be good at preventing thrombus formation, such as atrial fibrillation, but they may not be as efficacious for resolving thrombi. Second, LVT probably involves more than just stasis. There are certainly endothelial injury factors, too, especially in acute myocardial infarction. Third, the authors of prior studies only considered visible thrombus resolution on imaging. Robinson et al focused on clinically reported embolic events. Finally, it is possible that weaknesses in the Robinson et al study biased the study in favor of warfarin.

Regarding weaknesses, this was a retrospective, observational study, which could mean there were unmeasured confounders that biased the results. To the investigators’ credit, they conducted extensive sensitivity analyses, and none changed the results.

Still, there was no central review of the echoes. Other imaging results (if any other imaging occurred), were not considered. Transthoracic echo has limited sensitivity for thrombi. Thrombus escape from the LV could not be ascertained. Another reason Robinson et al focused on clinical events is that one-third of their patients did not undergo a follow-up echo. In addition, there was no information on the dosing of DOACs or adherence to this therapy. Almost all the DOACs used were factor Xa inhibitors. Finally, there was no analysis of bleeding events or other safety information.

Despite these potential limitations, this study was carefully performed. It is the largest to address this issue to date. Also, given the low incidence of LVT, it is doubtful a randomized, controlled trial will be conducted. For now, the off-label use of DOACs for LVT should be undertaken with caution until more data are acquired. Perhaps LVT should be added to the list of conditions for which DOACs are not useful, along with mechanical prosthetic valves.