Assessment of non‐vitamin K antagonist oral anticoagulants for the management of left ventricular thrombus

Abstract Although several studies have assessed the effect of non‐vitamin K antagonist oral anticoagulants (NOACs) relative to that of vitamin K antagonists (VKAs) in patients with left ventricular thrombus, the results remain controversial. Herein, a meta‐analysis was performed to compare the effectiveness and safety of NOACs versus VKAs for the treatment of left ventricular thrombus. We systematically searched the Cochrane Library, PubMed and Embase databases until November 2020 for studies that compared the effects of NOACs versus VKAs in patients with left ventricular thrombus. The treatment effects were expressed as odds ratios (ORs) with 95% confidence intervals (CIs) and pooled by a random‐effects model. Seven retrospective studies involving 865 patients with left ventricular thrombus (266 NOAC and 599 VKA users) were included. The pooled analysis suggested no difference in the rate of thrombus resolution between the NOAC and VKA groups (OR = 0.83, 95% CI 0.61–1.13). There were also no differences in the rates of stroke or systemic embolism (OR = 0.62, 95% CI 0.20–1.97), bleeding events (OR = 0.73, 95% CI 0.37–1.45), or all‐cause death (OR = 0.92, 95% CI 0.50–1.69) between patients treated with NOACs and those treated with VKAs. In addition, the rates of thrombus resolution, stroke or systemic embolism, bleeding events, and all‐cause death between NOAC‐ and warfarin‐treated patients were also similar. Our current evidence suggested that NOAC and VKA users had similar rates of thrombus resolution and clinical outcomes among patients with left ventricular thrombus. Further large‐scale prospective studies should confirm our results.

systemic embolism and substantial morbidity and mortality. 4 Therefore, patients with left ventricular thrombus often require anticoagulation therapy. Vitamin K antagonists (VKAs), such as warfarin, are recommended by expert consensus and guidelines and are clinically used for anticoagulation therapy in patients with left ventricular thrombus. 5 However, VKAs have several shortcomings, including marked inter-and intra-individual variations in medication dosage, a narrow therapeutic window, frequent international normalized ratio monitoring, and many drug-drug or drug-food interactions. 6 In recent years, non-vitamin K antagonist oral anticoagulants (NOACs) have been introduced for stroke prevention in patients with atrial fibrillation (AF). Novel drugs, including one direct thrombin inhibitor (dabigatran) and three direct Xa inhibitors (rivaroxaban, apixaban, and edoxaban), could improve the disadvantages of VKAs mentioned above. Evidence from randomized clinical trials and observational studies consistently demonstrated that NOACs are at least as effective as VKAs for stroke prevention and sometimes have better improved safety profiles in Asian or non-Asian patients with AF. [7][8][9][10][11] As such, NOACs have currently been recommended as first-line oral anticoagulants in the AF guidelines. [12][13][14][15] However, the effectiveness and safety of off-label use of NOACs to treat left ventricular thrombus are still unclear. 16 Several previous reviews qualitatively described that off-label use of NOACs could be a reasonable and valid option for the treatment of left ventricular thrombus. 5,17 However, these studies did not compare the effectiveness and safety of NOACs and VKAs in treating left ventricular thrombus. Moreover, there are still no direct head-to-head randomized clinical trials for this purpose. In recent years, several observational studies have assessed the effect of NOACs relative to that of VKAs in patients with left ventricular thrombus, but the results remain contradictory. [18][19][20][21][22][23][24] Therefore, we quantitatively performed a meta-analysis of observational studies comparing the effectiveness and safety of NOACs and VKAs on the rates of thrombus resolution and clinical outcomes in patients with left ventricular thrombus.

| METHODS
As described previously, our current meta-analysis was conducted according to Cochrane methodological standards, and the presentations were performed under the preferred reporting items for reporting systematic reviews and meta-analyses. 6 Ethical approval was not necessary because no patients were involved in setting the research question, the outcome measures, the design, or the implementation of this meta-analysis. The data, methods, and materials of this metaanalysis are available to others for purposes of reproducing the results or replicating procedures by contacting the corresponding author.

| Literature search strategy
Two reviewers systematically searched the Cochrane Library, PubMed and Embase databases until November 2020 for studies that compared the effectiveness and/or safety of any NOAC (dabigatran, rivaroxaban, apixaban, or edoxaban) with that of VKAs in patients with ventricular thrombus. The following key words and their similar search terms were combined using the Boolean operator "and": (1) 'ventricular thrombus' OR 'intraventricular thrombus' OR 'ventricular thrombi'; (2) 'non-vitamin K antagonists' OR 'NOAC' OR 'new oral anticoagulants' OR 'novel oral anticoagulants' OR 'direct oral anticoagulants' OR 'DOAC' OR 'oral thrombin inhibitors' OR 'oral factor Xa inhibitors' OR 'dabigatran' OR 'rivaroxaban' OR 'apixaban' In addition, we further searched the reference lists of the included studies to identify additional studies. We did not apply any restriction on the language of publication.

| Inclusion and exclusion criteria
Studies were eligible if they met the following criteria: (1) design of the study: observational prospective or retrospective study; (2) study population: patients with ventricular thrombus, regardless of the etiology, such as heart failure with reduced ejection fraction, acute myocardial infarction, and nonischemic cardiomyopathy; (3) comparisons: any NOAC (dabigatran, rivaroxaban, edoxaban, or apixaban; any dose) versus VKAs (e.g., coumadin, acenocoumarol, phenprocoumon, and warfarin); (4) clinical outcomes: thrombus resolution, stroke or systemic embolism, bleeding events, and all-cause death. We accepted the original definitions of the included studies; and (5) follow-up duration: no restrictions.
Certain publication types (e.g., reviews, case series, case reports, meta-analyses, editorials, and conference abstracts) or studies with insufficient data were excluded. If the study population had a substantial overlap among different studies, we included the study with the longest follow-up or largest sample size.

| Data extraction
All of the retrieved studies were independently screened by two reviewers. The first phase of screening was performed by reading the titles and abstracts of the records. Then, the second phase of screening involved reviewing the full text of the studies to identify potentially eligible studies. Discrepancies were resolved through discussion or dealt with by consultation with a third reviewer. Ultimately, we included the studies that met the eligibility criteria mentioned above.
For each included study, the following basic information was collected: study characteristics (e.g., the first author and publication year, study design, study period), patient characteristics (e.g., age, sex), type of NOACs, type of VKAs, follow-up time, and outcomes of interest. In the NOAC or VKA groups, the number of events, event rates, and sample size were extracted for the reported outcomes (thrombus resolution, stroke or systemic embolism, bleeding events, and all-cause death).

| Quality assessment
The Newcastle-Ottawa Scale (NOS) tool was used to assess the quality of observational studies. This tool had a total score of nine points.
Each included study was awarded a maximum of one point for each numbered item within the selection of cohorts (four points), the comparability of cohorts (two points), and the assessment of the outcome (three points). In this meta-analysis, we defined an NOS score of ≥6 points and <6 points as moderate-to-high quality and low quality, respectively. 6,25,26

| Statistical analysis
The Cochrane Q test and I 2 statistic were used to assess consistency across the included studies. For the Q statistic, a p value of <.1 indicated substantial heterogeneity. For the I 2 statistic, 25% or less, 50%, and 75% or more indicated low, moderate, and high heterogeneity, respectively. For each study, the number of events and sample size in each treatment group were pooled by a random-effects model. The pooled treatment effects were expressed as odds ratios (ORs) with 95% confidence intervals (CIs). Sensitivity analysis was performed to examine the influence of each study on the pooled results. We also reperformed the analysis by using a fixed-effects model. Publication

| Study selection
The literature retrieval process of this meta-analysis is presented in Figure 1. We identified 231 studies through the Cochrane Library, PubMed and Embase electronic databases after we excluded duplicate publications. Based on the title/abstract screenings, 110 studies were excluded according to the predefined criteria. Then, three conference abstracts were excluded based on the full text screenings (Table S1). Finally, a total of seven observational retrospective studies that were published in 2020 [18][19][20][21][22][23][24] involving 865 patients with left ventricular thrombus (n = 266 for NOACs and n = 599 for VKAs) were included in this meta-analysis.

| Study characteristics
The baseline characteristics of these included studies are shown in Table 1. Five studies used warfarin as the reference, while the combined VKAs, including warfarin, acenocoumarol, and fluindione, were regarded as the control group in the study by Daher et al. 20 Specifically, Robinson et al. 18 (Table S2).

| Effect of NOACs versus that of VKAs on thrombus resolution
All seven included studies reported the rate of thrombus resolution. A total of 154 events were found in 266 patients with NOACs, and 386 events were observed in 599 patients with VKAs. As shown in Figure 2, a random-effects model analysis suggested no difference in the rate of thrombus resolution between the NOAC and VKA groups F I G U R E 1 The literature retrieval process of this meta-analysis (OR = 0.83, 95% CI 0.61-1.13; p = .23). No heterogeneity was found across the included studies (Q statistic: p = .79, and I 2 = 0%).

| Effect of NOACs versus that of VKAs on stroke or systemic embolism
All seven included studies reported the event rate of stroke or systemic embolism. As shown in Figure 3, a total of 21 and 42 events were found in 266 NOAC users and 599 VKA users, respectively. In the pooled analysis, we found no significant difference in the event rate of stroke or systemic embolism between the two studied groups treated with NOACs versus VKAs (OR = 0.62, 95% CI 0.20-1.97; p = .42), with moderate heterogeneity (Q statistic: p = .03 and I 2 = 58%).  Figure 3, the pooled data suggested no difference in the rate of bleeding events between NOACs and VKAs (OR = 0.73, 95% CI 0.37-1.45; p = .37). We found no heterogeneity (Q statistic: p = .67, and I 2 = 0%).

| Effect of NOACs versus that of VKAs on allcause death
Two studies assessed the effect of NOACs versus that of VKAs on allcause death. A random-effects model analysis suggested that the event rate of all-cause death between NOACs and VKAs was not significantly different (OR = 0.92, 95% CI 0.50-1.69; p = .79), with no heterogeneity of this part (Q statistic: p = .49, and I 2 = 0%). The results of this part should be interpreted cautiously due to the limited number of included studies.   Figures S1 and S2). In addition, we also reperformed the analysis with a fixed-effects model, which suggested that NOACs versus VKAs yielded nonsignificantly different risks for the outcomes of thrombus resolution, stroke or systemic embolism, bleeding events, and all-cause death ( Figures S3 and S4).