Comparison of direct oral anticoagulants and low-molecular-weight heparins in cancer patients with nonvalvular atrial fibrillation: insights from a single-center experience

Abstract

Background Introduction: Atrial fibrillation (AF) is frequently encountered in patients with cancer and poses challenges for anticoagulant management. Cancer itself promotes a prothrombotic state, while anticancer therapies further elevate thromboembolic risk and paradoxically predispose to bleeding. Patients with both cancer and AF face a four- to seven-fold higher risk of venous thromboembolism (VTE) and a two-fold higher bleeding risk compared to those with AF alone. While direct oral anticoagulants (DOACs) have been shown to effectively reduce stroke risk in non-valvular AF (NVAF), cancer patients were largely excluded from randomized trials. In practice, concerns about drug interactions and bleeding often lead physicians to prefer low molecular weight heparin (LMWH), yet data comparing these agents in NVAF are scarce. Purpose The present study was conducted to compare clinical outcomes associated with DOACs and LMWH in individuals with active malignancy and NVAF. We hypothesized that this study would provide valuable insights to guide clinical practice, support clinical decision-making, and promote timely initiation of anticoagulant therapy. Methods Data from patients with active cancer who received LMWAH or DOAC for NVAF were retrospectively screened. Efficacy, safety, and survival outcomes were analyzed. Results The study enrolled 222 patients, of whom 25.7% received LMWH and 74.3% received DOACs. Cancer stage, type, and treatment were comparable between groups. The DOAC group had higher CHA2DS2-VA (3.13 ± 1.21 vs. 2.42 ± 1.33; p < 0.001) and HAS-BLED scores (2.13 ± 0.83 vs. 1.84 ± 0.79; p = 0.022). The primary composite endpoint occurred more frequently in the LMWH group (17.5% vs. 12.1%; log-rank p = 0.036). Myocardial infarction was significantly higher in the LMWH group (8.8% vs. 1.8%; log-rank p=0.003), while rates of ischemic stroke (3.5% vs. 6.7%; log-rank p=0.927) and VTE (5.3% vs. 4.8%; p = 0.537) were similar. Any bleeding (17.5% vs. 13.3%; log-rank p=0.039) and major bleeding (7.0% vs. 1.8%; log-rank p=0.010) were more frequent with LMWH, whereas clinically relevant non-major bleeding was comparable (10.5% vs. 11.5%; log-rank p=0.359). All-cause mortality was significantly higher in the LMWH group (75.4% vs. 49.1%; log-rank p < 0.001), and LMWH use independently predicted mortality (HR = 2.14; 95% CI 1.45–3.17; p < 0.001). Conclusions Although unmeasured confounders such as drug adherence and selection bias—since LMWH may have been preferentially prescribed to frail patients—cannot be excluded due to the retrospective design, DOACs appear to be more effective and safer than LMWH in cancer patients with AF.Figure 1 Figure 2