J Clin Med. 2026 Nov 9;15(21):8736. doi: 10.3390/jcm15218736.
ABSTRACT
Objectives: This study aims to describe real-world oral anticoagulation management in patients with NVAF followed up in Spanish Primary Care, including anticoagulant use, VKA control, treatment switching, patient-reported satisfaction, and clinical outcomes during documented anticoagulant exposure. Methods: RACOVIR was a multicenter observational registry with cross-sectional enrolment and retrospective collection of electronic medical record data from AF diagnosis to inclusion. Adults with NVAF receiving oral anticoagulation for ≥6 months were included. The Barthel Index and Anti-Clot Treatment Scale assessed autonomy and treatment satisfaction at inclusion. Clinical events were evaluated using exposure-based incidence rates, supplemented by weighted marginal structural models and propensity-score-adjusted Cox sensitivity analyses restricted to patients receiving a single anticoagulant strategy. Results: Among 1901 patients, 428 received VKAs and 1473 DOACs at enrolment. Mean exposure duration was 7.6 years for VKAs and 3.8 years for DOACs. Adequate VKA control was achieved by 56.2% and 59.0% using the direct and Rosendaal methods, respectively. Reduced DOAC doses were recorded in 24.8% of apixaban, 54.7% of dabigatran, 26.9% of edoxaban, and 21.9% of rivaroxaban users; appropriateness could not be assessed for every patient. Overall, 395 patients (20.8%) received both strategies. DOAC users reported greater treatment satisfaction. Unadjusted exposure-based incidence rate ratios (DOAC/VKA) were 1.75 (95% CI 1.05-2.92) for stroke, 1.78 (95% CI 1.06-3.00) for ischemic stroke, 1.64 (95% CI 0.97-2.79) for myocardial infarction, and 3.68 (95% CI 1.81-7.46) for major bleeding. In weighted marginal structural models, adjusted incidence rate ratios (DOAC/VKA) were 1.627 (95% CI 0.785-3.375) for stroke and 6.725 (95% CI 2.106-21.473) for major bleeding. A propensity-score-adjusted Cox sensitivity analysis excluding treatment switchers (n = 1492) yielded corresponding hazard ratios of 2.577 (95% CI 0.930-7.144; p = 0.069) and 8.298 (95% CI 3.180-21.649; p < 0.001), respectively. Conclusions: RACOVIR shows substantial heterogeneity in anticoagulation management in Spanish Primary Care. DOACs predominated and were associated with greater patient-reported satisfaction, while suboptimal INR control remained common with VKAs. Although adjusted analyses suggested higher major bleeding rates during DOAC exposure, these findings should be interpreted cautiously because of the observational design, potential selection bias, treatment switching, and residual confounding. No causal conclusions regarding comparative anticoagulant safety or effectiveness can be established.
PMID:43030878 | PMC:PMC13763620 | DOI:10.3390/jcm15218736