Efficacy and safety of antithrombotic regimens in patients with durable left ventricular assist devices: a systematic review and bayesian network meta-analysis

Abstract Background Advances in durability and hemocompatibility have improved outcomes for advanced heart failure (HF) patients with left ventricular assist devices (LVADs). However, the anticoagulation management in this population is challenging, and the optimal strategy remains unclear, especially concerning the need for antiplatelet therapy and the potential role of direct oral anticoagulants (DOACs) over vitamin K antagonists (VKAs). Purpose We aimed to compare the efficacy and safety of three antithrombotic regimens in HF patients with durable LVADs: (1) VKA monotherapy, (2) VKAs plus aspirin (ASA), and (3) DOACs plus ASA. Methods A systematic review and network meta-analysis (NMA) were conducted in accordance with PRISMA guidelines. PubMed, Embase, and Cochrane databases were searched for randomized controlled trials (RCTs) and non-RCTs comparing thrombotic and bleeding outcomes among antithrombotic regimens in HF patients with LVADs. A random-effects Bayesian NMA was performed using the Markov Chain Monte Carlo algorithm to estimate pooled odds ratios (ORs) with 95% credible intervals (CrIs). Treatment ranking was assessed using the Surface Under the Cumulative Ranking (SUCRA) score. Data synthesis was conducted using the gemtc package in R software. Results Ten studies (four RCTs and six non-RCTs) involving 972 patients met the inclusion criteria. Among them, 406 patients received VKA monotherapy, 502 received VKAs plus ASA, and 64 were treated with DOACs plus ASA. The majority were male (78.9%), with a mean age and ejection fraction of 58.7 ± 11.3 years and 15.1 ± 0.6%, respectively. VKA monotherapy was associated with a significantly lower risk of non-surgical bleeding events (NSBE) (OR 0.21; 95% CrI 0.06–0.78; Figure 1A) and gastrointestinal bleeding (GIB) (OR 0.33; 95% CrI 0.12–0.86; Figure 1B) compared to VKAs plus ASA. No significant differences were found when comparing DOACs plus ASA to either VKAs plus ASA or VKA monotherapy regarding NSBE or GIB (Figure 1). Based on SUCRA probabilities, VKA monotherapy ranked highest for reducing NSBE (0.7648) and GIB (0.7708). No significant differences were found among anticoagulation strategies regarding ischemic stroke/transient ischemic attack (IS/TIA), pump thrombosis (PT), or all-cause mortality (Figure 2). SUCRA rankings indicated that VKA monotherapy had the highest probability of reducing IS/TIA (0.8670) and all-cause mortality (0.6272), while DOACs plus ASA ranked highest for reducing PT (0.7320). Conclusion This NMA suggests that VKA monotherapy is significantly associated with a reduced risk of bleeding events compared to VKAs plus ASA, without a significant increase in thrombotic complications. No statistically significant differences between DOACs and VKAs were observed in this population, although further large-scale, high-quality RCTs are needed to confirm these findings, guide clinical practice, and refine anticoagulation strategies for HF patients with durable LVADs.