Resolution of left atrial appendage thrombi: No difference between phenprocoumon and non‐vitamin K‐dependent oral antagonists

Abstract Background Atrial fibrillation is the most important risk factor for left atrial appendage (LAA) thrombi, a potentially life‐threatening condition. Thrombus resolution may prevent embolic events and allow rhythm‐control strategies, which have been shown to reduce cardiovascular complications. Hypothesis There is no significant difference between phenprocoumon and non‐Vitamin K‐dependent oral anticoagulants (NOACs) in the resolution of LAA‐thrombi in a real‐world setting. Methods Consecutive patients with LAA‐thrombi from June 2013 to June 2017 were included in an observational single‐center analysis. The primary endpoint was defined as the resolution of the thrombus. The observational period was 1 year. Resolutions rates in patients on phenprocoumon or NOACs were compared and the time to resolution was analyzed. Results We identified 114 patients with LAA‐thrombi. There was no significant difference in the efficacy of resolution between phenprocoumon and NOACs (p = .499) at the time of first control which took place after a mean of 58 ± 42.2 (median 48) days. At first control most thrombi were dissolved (74.6%). The analysis after set‐time intervals revealed a resolution rate of 2/3 of LAA‐thrombi after 8–10 weeks in the phenprocoumon and NOAC groups. After 12 weeks a higher number of thrombi had resolved in the presence of NOAC (89.3%) whereas in the presence of phenprocoumon 68.3% had resolved (p = .046). Conclusion In this large observational study NOACs were found to be potent drugs for the resolution of LAA‐thrombi. In addition, the resolution of LAA‐thrombi was found to be faster in the presence of NOAC as compared to phenprocoumon.


| INTRODUCTION
Over 26 million people worldwide suffer from a stroke every year. In western countries 20% of all strokes and transient ischemic attacks (TIAs) are of cardioembolic origin. 1,2 Cardioembolic strokes are more severe than other ischemic strokes. 3,4 There has been a steady increase of cardioembolic strokes in the last few years. 5 In patients with atrial fibrillation (AF) as the most important risk factor, a consequent anticoagulant therapy avoids 70% of all cardioembolic strokes. 5,6 Therefore, the presence of intracardiac thrombi is a potentially life-threatening condition because of the risk of embolization. To allow rhythm control in patients with AF which may be of prognostic relevance 7 the presence of intracardiac thrombi has to be excluded in advance.
Thus, not only prevention but also adequate therapy of existing thrombi is of utmost importance. Today non-vitamin K-dependent oral anticoagulants (NOACs) are favored for the prevention of intracardiac thrombi in the majority of patients with AF. 8 However, only limited data exist for the use of NOACs in the resolution of already existing left atrial appendage (LAA) thrombi. 9 NOACs benefit from their simpler application compared to the use of vitamin K antagonists (VKAs) and are largely independent on alimentation or co-medication. Additionally, therapeutic levels are consistent for most patients and their use is mostly restricted by renal insufficiency. In contrast, VKAs show many interdependencies especially with the frequently used antiarrhythmic drug amiodarone and therapeutic levels can vary depending on alimentation. Still, the amount of practical experience on thrombus resolution acquired over decades may favor VKA. 10,11 This study aims at comparing the efficacy of phenprocoumon and NOACs in the resolution of LAA-thrombi in a real-world setting.

| METHODS
This analysis included all consecutive patients diagnosed with an intracardiac thrombus from June 2013 to June 2017 in a general cardiology clinic (SFH Münster, Germany). The intent was to compare the resolving potential of NOACs and the VKA phenprocoumon on thrombi of the LAA. Patients with non-LAA thrombi (e.g., left ventricular thrombi) were excluded. The primary endpoint was defined as the resolution of the thrombus when patients presented again for follow-up.
Informed consent was obtained from all individual participants included in the study. The datasets generated and analyzed during the current study are not publicly available, as per internal protocol, but are available from the corresponding author on reasonable request.
Persistence of the intracardiac thrombi after 1 year was defined as the secondary endpoint. Controls were made by TEE and follow-up appointments were scheduled according to hospital capacity between 4 and 13 weeks after diagnosis. All changes in anticoagulant therapy with date and specific substance were analyzed.
This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of "Ärztekammer Westfalen-Lippe" (Nr. 2019-641-f-S).
LAA thrombi (n = 114) were diagnosed by transesophageal echocardiography (TEE; n = 111), computer tomography (CT, n = 2), or magnetic resonance imaging (MRI, n = 1). The vast majority of patients presented with symptomatic AF. Some patients already presented with cardioembolic complications and were diagnosed with AF in the following diagnosis of a thromboembolic event.

| Statistical analysis
After collecting all data, the evaluation and statistical analysis was made using SPSS Version 25 for Mac OS X (SPSS Inc.). All descriptive data were specified by absolute and relative frequency and complemented with median, arithmetic mean, and SD where necessary. The χ 2 test was used to test for independency. The Kaplan-Meier estimator was used in the context of event history analysis to identify time to resolution under therapy with different types of anticoagulation. Log-rank tests were used to test for significance. In case of small size of groups Fisher's exact test was applied.

| Baseline characteristics and comorbidities
One hundred and sixty-three consecutive patients were included, out of those 114 were diagnosed with an intracardiac thrombus located in the left atrium (n = 1) or LAA (n = 113). The remaining patients had LV-thrombi (n = 36) or thrombi of other locations (n = 13) and were not taken into account.

| AF and LAA thrombi
Most patients (91.9%) with an LAA-thrombus had documented AF or atrial flutter at the time of diagnosis. Persistent AF was the most common overall as well as in the subgroups of patients on phenprocoumon or NOAC (68.8% and 78.9%, respectively; Table 1).
Correspondingly, the majority of patients for whom all relevant data points could be acquired (n = 99) showed varying degrees of dilatation of the LA, only 23.2% of patients showed no dilatation according to the American Society of Echocardiography (Table 1). 12 The mean size of LAA thrombi was 128.9 mm² (±142.5) for the 38 patients for whom all relevant data points could be acquired.

| Major adverse events
The majority of patients (88.4%) did not suffer a stroke before being diagnosed with an intracardiac thrombus. During the course of the T A B L E 1 Characteristics of the patient collective with LAA-thrombi (n = 114) in absolute numbers and percentage including SD were needed at the time of inclusion

| Time to resolution
Out of the 67 patients who were controlled at least once, 56 showed the resolution of the thrombus within 1 year. Irrespective of the type of oral anticoagulation the average time to resolution was 77.8 (SD ± 7.4) days, the median was 53.5 days. A resolution rate of twothirds of thrombi was reached after 71 days (Figure 1), correspondingly a control rate of more than 80% was reached after 10 weeks (Table 2). Additionally, one patient was first controlled after 211 days and one after 245 days. Both thrombi were dissolved.  Table 3.

| Resolution of LAA-thrombi in dependence of the type of oral anticoagulation
Comparable percentages of controlled patients were seen, for example, after 8 weeks control rates were 82.1% and 81.3% in the phenprocoumon and NOAC group, respectively.
After 4 and 6 weeks, resolution rates were overall low, but slighter higher in the NOAC group. With regard to the control after   In the NOAC group in our study 73.7% of thrombi were dissolved at the time of first control. This indicates that the first control probably took place later in CLOT-AF compared to our study. 13 Comparing the results of the X-TRA trial with our data it can be noted that after 6 weeks there was a very similar resolution rate of 40% across the NOAC group as a whole. It may overall be possible that the time to first control was too short to fully investigate the dissolving potential of Rivaroxaban 9 resulting in a resolution rate of 41.5% in X-TRA.
Comparing the dissolving potential of NOACs with the potential of VKA reported from other studies the effectiveness of the NOACs seems to be at least as good as that of VKA. Only very limited data exist that investigates the potential of VKA in thrombus resolution. In Factor Xa when already bound to thrombi. 16 The clinical consequence of a potentially faster resolution when using NOACs is first and foremost reflected in the possibility for earlier controls of LAAthrombi making an earlier antiarrhythmic strategy for AF possible.
When deciding on the type of oral anticoagulation it should also be taken into account that the number of thromboembolic complications may be reduced by faster resolution of the thrombus.

| CONCLUSION
In this large observational study the overall efficacy of thrombus resolution did not differ significantly between VKA and NOACs at the Further studies are needed to prove the role of NOACs and to differentiate between the resolving potential of thrombin-and Factor Xa inhibition as well as for each NOAC individually.

ACKNOWLEDGMENT
Open Access funding enabled and organized by Projekt DEAL.

CONFLICTS OF INTEREST
L. Eckardt reports receiving lecture honora from Boehringer Ingelheim, Daiichy Sankyo, BMS Pfizer, and Bayer Medical. H.
Wedekind reports receiving lecture honora from Bayer Medical and AstraZeneca. The remaining authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT
The datasets generated and/or analyzed during the current study are not publicly available, as per internal protocol, but are available from the corresponding author on reasonable request.