Comparison of different direct oral anticoagulants versus vitamin K antagonists in patients with left ventricular thrombus. Bayesian Network Meta-analysis with reconstructed individual patient data.

Background: Left ventricular thrombus (LVT) is associated with a high rate of embolic complications. Despite major advances in revascularization strategies and heart failure treatment, the incidence of LVT in patients with impaired regional or global left ventricular systolic function remains relevant. Vitamin K antagonists (VKAs) are the standard of care, but recent data suggested that direct oral anticoagulants (DOACs) may be equally effective. Whether different DOACs are equally effective remains debatable. Therefore, the aim of this network meta-analysis (NMA) was to compare the effectiveness of different oral anticoagulants in the treatment of LVT.

Methods: We performed a systematic literature search in the databases of MEDLINE, EMBASE and Web of Science for randomized (RCTs) and non-randomized controlled studies (NRS) comparing the effectiveness of different DOACs and/or DOACs and VKAs for LVT treatment. Individual patient data was reconstructed from Kaplan-Meier curves and included in the Bayesian NMA based on a random-effects model with informative priors. This systemic review and network meta-analysis is registered with PROSPERO (CRD42023459096).

Results: Of the 452 articles identified by the literature search 20 were ultimately included in the NMA (4 RCTs, 16 NRS). Rivaroxaban showed significantly higher cumulative LVT resolution in the meta-analysis of reconstructed time-to-event data compared to VKAs (HR 0.64, 95% CI [0.47; 0.87], log-rank p = 0.004, Fig. 1). No time-to-event data was available for other DOACs. Data for both Dabigatran and Edoxaban were limited to only few patients (data not shown). The NMA showed significantly higher rates of LVT resolution for Rivaroxaban (OR 2.09, 95% CrI [1.09; 4.71], Fig. 2) but not for Apixaban (OR 0.51, 95% CrI [0.13, 1.98], Fig. 2) compared to VKAs at 1 month. At 3 months LVT resolution rates were similar for both Apixaban (OR 1.7, 95% CrI [0.57; 6.4], Fig. 2) and Rivaroxaban (OR 1.8, 95% CrI [0.77; 3.5], Fig. 2) compared to VKAs. LVT resolution rates were also similar independent of follow-up duration thereafter for Rivaroxaban (OR 1.1, 95% CrI [0.65; 1.74], Fig. 2) and Apixaban (OR 2.44, 95% CrI [0.78; 8.44], Fig. 2) compared to VKAs. Rivaroxaban showed lower incidence of ischemic stroke (OR 0.22, 95% CrI [0.07; 0.56], Fig. 2) and systemic embolism (OR 0.25, 95% CrI [0.07; 0.70], Fig. 2) compared to VKAs. This was not the case for Apixaban (OR 0.69, 95% CrI [0.25; 1.59] for ischemic stroke, OR 0.26, 95% CrI [0.02; 1.89] for systemic embolism, Fig. 2). The incidence of major bleeding events was lower for Apixaban (OR 0.24, 95% CrI [0.04; 0.93], Fig. 2) but not for Rivaroxaban (OR 0.48, 95% CrI [0.20; 1.03], Fig. 2) compared to VKAs.

Conclusion: In this NMA Rivaroxaban showed faster thrombus resolution compared to VKAs. Overall LVT resolution independent of follow-up duration was similar for Rivaroxaban and Apixaban compared to VKAs. Rivaroxaban showed lower odds of ischemic stroke and systemic embolism compared to VKAs without increase in bleeding events. The results of this NMA emphasize the need for adequately powered RCTs comparing Rivaroxaban to the current standard-of-care treatment of LVT with VKAs.

: Kaplan-Meier curve of cumulative incidence of left ventricular thrombus resolution. Blue line = Rivaroxaban; Red line = Vitamin K antagonists