Left atrial thrombus formation and resolution during dabigatran therapy: A Japanese Heart Rhythm Society report

3.1 LA thrombus prevalence

A total of 198 patients underwent TEE to rule out the presence of LA thrombus before ECV. The average age was 61.6 years with an 84% male predominance. The average CHADS₂ score was 0.99 with 3.5% of patients having a history of stroke or transient ischemic attack.

Dabi 150 mg b.i.d and 110 mg b.i.d were given to 98 and 100 patients, respectively. LA thrombus was found in one patient receiving the higher dosage while seven patients received the lower dosage (p=0.076) (Table 1). The decision to administer the lower dosage in 58% of patients was based on Japan's recommendations, which requires the fulfillment of one of the following criteria: creatinine clearance rate, 30–50 mL/min; co-administration of p-glycoprotein inhibitors such as amiodarone or verapamil; age ≥70 years; or a history of gastrointestinal bleeding. In the remaining 42% of patients, physicians arbitrarily selected the lower dosage.

Dabi was administered for <3 weeks in 21%, for 3–6 weeks in 24%, and for ≥6 weeks in 55% of patients prior to TEE.

3.2 Characteristics of LA thrombus patients

LA thrombus was found in eight (4%) out of 198 patients (Fig. 1). The eight patients with LA thrombus tended to be older (67.3 vs. 61.3 years, p=0.175), had higher CHADS₂ scores (1.88 vs. 0.95, p=0.058), and had a higher prevalence of prior stroke or transient ischemic attack (22.2% vs. 2.6%, p=0.034) than those without LA thrombus (Table 1).

Of the eight patients with LA thrombus, one was receiving a Dabi dosage of 150 mg b.i.d, whereas the remaining seven were receiving 110 mg b.i.d for ≥3 weeks. Notably, in three of these seven patients, the dosage of 110 mg b.i.d was arbitrarily selected and did not reflect the Japanese recommendations for dosage reduction as described in Section 3.1.

3.3 LA thrombus fate

A second TEE was performed in six of the eight patients with LA thrombus, which revealed complete resolution of the thrombus in five patients with the earliest resolution within 23 days after the first TEE (Fig. 1). Of these five patients, one was receiving a prolonged Dabi dosage of 150 mg b.i.d, two had an increase in dosage from 110 mg to 150 mg b.i.d, and the remaining two were switched to warfarin (Table 2, Figs. 2 and 3). One patient undergoing the second TEE 9 days after warfarin therapy only showed LA thrombus shrinkage.

3.4 ECV complications

ECV was carried out in both 190 patients without LA thrombus and in five patients after LA thrombus resolution. ECV was not attempted in the three remaining patients with no LA thrombus resolution confirmation by TEE. Two patients (1%) had a stroke at days 3 and 15 after ECV while on Dabi 110 mg b.i.d, although LA thrombus was not detected before ECV. The former patient had CHADS₂ score zero, but received a lower dosage of 110 mg b.i.d arbitrarily selected by a physician, and the latter was an 83-year-old patient with hypertension. A 68-year-old male diabetic patient receiving a Dabi dosage of 110 mg b.i.d had a major large intestine bleeding episode secondary to colon carcinoma and required hospitalization.

4.1 Main findings

LA thrombus developed in 4% of patients with AF despite Dabi administration before elective ECV, as shown on TEE. Older patients with higher CHADS₂ score, an inappropriately Dabi lower dosage, or a shorter treatment duration appeared to be more susceptible to LA thrombus formation. LA thrombus resolution was achieved typically after an increase in Dabi dosage or switching to warfarin. Finally, although peri-ECV thromboembolic complications were rare, they occurred despite the previous absence of LA thrombus.

LA thrombus formation in AF patients receiving NOAC treatment has not been studied extensively. Although modalities to examine LA thrombus such as computed tomography or intracardiac echocardiography exist, TEE is by far the most popular and reliable method being used in clinical practice [3], [4]. However, because of its invasive nature, TEE is indicated only in patients undergoing ECV or AF ablation to exclude the presence of LA thrombus before the procedure. This survey has thus provided an opportunity to ascertain LA thrombus prevalence among patients with persistent AF selected to undergo elective ECV.

4.2 LA thrombus development during warfarin therapy

This survey demonstrated the presence of LA thrombus in 4% of patients with AF despite all receiving anticoagulant therapy with Dabi for at least 3 weeks. Although it may not be appropriate to compare this result with those of previous studies using warfarin, a single-center study in Germany including 799 patients with persistent AF receiving warfarin for at least 3 weeks and with a prothrombin time–international normalized ratio (PT–INR) between 2 and 3 detected LA thrombus using TEE in 7.7% [5]. In that study, thromboembolic events developed within 4 weeks of ECV at a rate of 0.8% in both groups of patients receiving the TEE guided and non-guided conventional approach, indicating that LA thrombus, even if present, may not necessarily be dislodged to cause thromboembolic events occurring after ECV. Although the prevalence of LA thrombus has been variable depending on the subjects studied, it was 4% if TEE was performed within 48 h of AF without prior anticoagulation [6]. ECV without oral anticoagulant therapy was previously accepted in patients with <48 h of AF [7]. However, a recent Finnish study demonstrated that even in this set of patients, the risk of thromboembolic events can reach up to 9.8% among those with both heart failure and diabetes [8]. On the other hand, LA thrombus prevalence was reported to be 3% in the above-mentioned German study when TEE was performed in 295 low-risk patients with AF with a CHADS₂ score of 0 or 1 [9]. LA thrombus prevalence or sludge increases with increasing scores of CHADS₂ and CHA₂DS₂-VASc [10], [11]. Therefore, it was not surprising that in our survey, patients with LA thrombus tended to be older and had a higher CHADS₂ score and prevalence of previous stroke or transient ischemic attack than those without demonstrable LA thrombus.

4.3 LA thrombus development during NOAC use

Although Dabi appears to be an effective alternative to warfarin in preventing stroke or systemic embolism without requiring PT–INR monitoring, the fact that 4% of patients receiving Dabi still had LA thrombus in our survey suggests that TEE is useful when performed before elective ECV of persistent AF. Thrombus may develop even under Dabi treatment in the presence of mechanical heart valves or rheumatic mitral stenosis [12], [13], but a few cases of failure to prevent LA thrombus formation with the use of this agent have also been reported in patients with AF without such backgrounds [14]-[16]. Although it is hard to prove, we speculate that not only higher CHADS₂ scores, but also the insufficient Dabi dosage given to some of our patients might have led to inadequate anticoagulation and subsequent clot formation. Three of the seven patients taking this lower dosage did not fulfill any of the recommendation criteria for using a 110 mg b.i.d dosage, but whether a 150 mg b.i.d dosage could have prevented LA thrombus formation in these patients remains uncertain.

In the RE-LY study, where TEE was performed according to physicians' preference, LA thrombus was detected in 1.8% of 165 patients and in 1.2% of 162 patients receiving 110 mg b.i.d and 150 mg b.i.d dosage, respectively [17]. Although the results of the RE-LY study were in line with ours, we observed a higher incidence of LA thrombus at the lower Dabi dosage, which could be a reflection of shorter anticoagulation duration as well as a potentially lower patient adherence to Dabi treatment, in a real-world scenario. Dabi treatment adherence was not carefully evaluated in this survey, but poor adherence in some patients could have been responsible for inadequate anticoagulation. As LA thrombus can develop in a substantial number of patients receiving Dabi, we recommend TEE for every ECV candidate, but whether this also applies to other NOACs remains uncertain. Peri-ECV complication sub-analysis in prospective studies yielded variable stroke or systemic embolism rates of 0.8% and 0.3% at Dabi dosages of 110 mg and 150 mg b.i.d in the RE-LY study, respectively, 1.8% with rivaroxaban in the ROCKET AF study, and 0% with apixaban in the ARISTOTLE study 30 days after ECV [18], [19]. A recent prospective randomized trial comparing rivaroxaban and vitamin K antagonists for AF cardioversion failed to show any significant difference in the primary efficacy and safety outcomes [20]. However, in a small patient subgroup where TEE was performed optionally, LA thrombus was found in six out of 33 patients versus zero out of 31 patients receiving rivaroxaban or vitamin K antagonists, respectively, for 3–8 weeks prior to ECV.

4.4 Stroke after negative TEE

In our study, two patients (1%) had a stroke after ECV. Although not every preexisting LA thrombus might cause thromboembolism after ECV, stroke incidence could have been higher if ECV had been performed without prior TEE. However, it is noteworthy that TEE was not able to detect LA thrombus in any of these two patients before ECV. Thromboembolic events after a negative TEE are common and have significant clinical importance. One possibility is that TEE has limited power to detect a preexisting small LA thrombus or sludge. Although dense spontaneous echo contrast was not systematically evaluated in this survey, sludge is not only considered to be a precursor, but may also mask the detection of a well-formed LA thrombus. Baseline myocardial structure such as the pectinate muscles of the LA appendage may also be sometimes erroneously diagnosed as LA thrombus, while the contrary can lead to its underdiagnosis. However, as the CHADS₂ score was zero in one of these two patients, a more plausible explanation would be de novo LA thrombus formation as a consequence of ECV itself.

Atrial function is known to remain reduced for ≥3 weeks after ECV, and this may provide a milieu for de novo thrombus formation [21]. Based on repeated TEE in five patients with stroke after negative TEE, Black et al. found new or increased LA spontaneous echo contrast in four patients including one patient with a new LA appendage thrombus [22]. When TEE was carried out within 5–15 min of ECV, Grimm et al. found spontaneous echo contrast developing de novo in 4 of the 20 patients [23]. Furthermore, Fatkin et al. carried out ECV in 15 patients while TEE was in place and found that within 10 s of successful ECV, LA spontaneous echo contrast appeared in five and increased in one, which was associated with a transient atrial dysfunction (stunning) [24]. LA thrombus formation immediately after ECV has also been demonstrated on TEE or intracardiac echo during AF ablation [25], [26]. Therefore, the possibility of ECV causing or unmasking atrial stunning and potentially inactivating thrombogenic factors in the atrial endocardium or blood constituents should not be underestimated. More importantly, the limited power of TEE in predicting later LA thrombus formation should be kept in mind.

4.5 LA thrombus resolution with NOAC use

LA thrombus resolution is another important observation in this survey. Of the six patients in whom LA thrombus was detected in the first TEE and later underwent a second TEE, five experienced thrombus disappearance ≥3 weeks later, and one showed thrombus shrinkage after 9 days of anticoagulant therapy. Although prolonged treatment with Dabi 150 mg b.i.d and increasing the dosage up to 150 mg b.i.d or switching to warfarin appeared to be effective in thrombus resolution in three and two patients, respectively, the number of patients was too small and the protocol was not standardized or randomized to be able to reach a definite cause-effect relationship regarding this observation. Stoddard et al. found that of 21 patients undergoing serial TEE, LA thrombus resolved only in one patient at 3–5 weeks of warfarin therapy, and in the 8 additional patients at 5–17 weeks [27]. An ACUTE study sub-analysis revealed complete thrombus resolution in 26 (76.5%) of 34 patients undergoing a repeat TEE after an average of 31.7 days of warfarin anticoagulation therapy [28]. Therefore, LA thrombus resolution is a common phenomenon that can result from either active or passive effects of any anticoagulant.

The potential thrombolytic effect of Dabi has been previously described [29]-[31]. Warfarin anticoagulation effects may not only be difficult to control at times, but could also exert a transient thrombogenic action in its loading phase [32]. By contrast, Dabi may create a consistent anticoagulation milieu easier and faster while inhibiting thrombin binding to fibrin and fibrin degradation products [33], [34]. LA thrombus resolution has also been reported with rivaroxaban and apixaban [35]-[38]. However, additional prospective and detailed studies are necessary before Dabi or any NOAC is recommended for clot resolution.

4.6 Study limitations

There are several limitations in this study. This retrospective study was carried out without following a consistent protocol for Dabi dosage and therapy duration prior to or after ECV. The possibility of selection bias cannot be excluded as a TEE was performed at the discretion of individual doctors. No protocol was provided for dosage change or type of oral anticoagulants, or for the timing of the second TEE after LA thrombus detection, nor was an assessment available to ascertain Dabi adherence before and after ECV. Individual doctors with potentially variable skills made LA thrombus diagnosis by using TEE.