Atrial fibrillation (AF) is the most prevalent cardiac arrhythmia in adults, confers a 3–5 fold increased risk of acute ischemic stroke (AIS) compared to non-AF populations which is accompanied by significantly higher mortality rates and worse functional outcomes (1–3). Therefore, long-term oral anticoagulation (OAC) was recommended for secondary prevention (4–8). However, the optimal time of OAC initiation post-AIS remains contentious, particularly balancing thromboembolic recurrence risk against hemorrhagic risk (9–12).

Endovascular therapy (EVT) demonstrates substantial clinical benefit in AF-related AIS patients with large vessel occlusion, particularly within 24 h of symptom onset (13), and enhanced therapeutic efficacy observed in patients with salvage brain tissue during extended time windows of 6–24 h (14). EVT can restore blood flow to occluded arteries, but reperfusion is more likely to lead to hemorrhagic transformation (HT) (15). The suboptimal utilization of OAC both during hospitalization and post-discharge periods may be attributed to heightened post-EVT hemorrhagic transformation risk and stroke severity. Whether timely anticoagulation management after EVT can improve outcomes in all patients with AF and AIS still requires further corroboration from relevant studies. Therefore, this study will utilize propensity score matching (PSM) to analyze the safety and efficacy of early postoperative anticoagulation in these patients, specifically investigating its association with functional outcomes, recurrent ischemic stroke/systemic embolism (IS/SE), safety, and all-cause mortality.

In AF-associated AIS, during hospitalization and post-discharge periods, timing of initiation of anticoagulation remains suboptimal (18), potentially attributable to heightened post-EVT hemorrhagic transformation risk and stroke severity. Prior research has demonstrated the incidence of HT in AIS is approximately 18–42 percent (19). During the acute phase of stroke, neuroinflammatory responses exacerbate blood-brain barrier injury, thereby increasing vascular permeability. Concurrently, rapid post-EVT cerebral reperfusion may facilitate the transmural passage of water and osmotic solutes across the compromised barrier, elevating hemorrhagic transformation risk (15). A meta-analysis of 25 cohort studies showed that age and a higher NIHSS score at admission were predictors of symptomatic intracranial hemorrhage (sICH) in patients with AIS after EVT (20). Tian et al. included 633 patients who received EVT treatment in their study and found that the multivariate model indicated that the baseline NIHSS score was an independent predictor of sICH after EVT (21). In this study, we also found that patients who did not receive OACs had higher HAS-BLED scores, higher NIHSS scores, higher proportion of postoperative HT in the hospital compared with those who received OAC at discharge, which may affect the use of OAC.

Existing research indicates that optimal anticoagulation strategies enhance long-term functional recovery while mitigating mortality and disability risks in stroke patients with AF (22, 23). Yan et al. conducted a study of 400 AF-associated AIS patients, demonstrating that in-hospital OAC initiation independently predicted superior functional recovery (mRS 0–1) and favorable neurological status (mRS 0–2) at 1-year assessment (24). Weller et al. analyzed 82 AF patients with acute large vessel occlusive stroke undergoing EVT and carotid stenting, demonstrating significantly elevated 90-day mortality rates in non-anticoagulated patients compared to discharged OAC recipients (25). In this study, there was a selection bias in the administration of anticoagulant therapy to stroke patients, as those who received it tended to have milder stroke severity, smaller infarct volumes, and a lower risk of hemorrhage. We employed PSM analysis to perform 1:1 matching between the two groups on some baseline characteristics potentially affecting prognostic outcomes, incorporating a total of nine matching factors. Among these, the admission NIHSS score, HAS-BLED score, LVEF, post-operative HT, infarct size, discharge NIHSS score, and the discharge mRS score were variables that showed statistically significant differences in baseline characteristics between the two groups (P < 0.05). Furthermore, the study was specifically restricted to patients with AF complicated by stroke who underwent EVT. It was necessary to control for two factors that could potentially influence the study results—AF subtype and surgical technique. Therefore, these were included as matching factors to minimize the disparity in confounding variables. Our results demonstrate that in a majority group of stroke patients without severe bleeding risk who were clinically selected for anticoagulation therapy, regular long-term anticoagulation was significantly associated with improved functional and mortality outcomes. This association remained significant after PSM adjusted for key between-group differences.

Patients in the anticoagulation group exhibited a significantly lower 1-year all-cause mortality in this study. Kaplan–Meier analysis revealed that the survival benefit associated with anticoagulant therapy primarily emerged within the first month after discharge. Acute-phase inflammatory and prothrombotic responses peak within 48–72 h post-stroke and can persist for several weeks (26). Initiating anticoagulation early in the post-EVT period might provide protection during this critical time window, thereby contributing to the observed reduction in all-cause mortality. In addition, the lack of observed difference in recurrent IS/SE events between the anticoagulation and non-anticoagulation groups in this study may be attributed to residual confounding by unmeasured factors, such as postoperative rehabilitation, frailty, and postoperative infection. In the non-anticoagulation group, a higher burden of these unmeasured negative prognostic factors likely contributed to more severe illness and higher short-term mortality. For some critically ill patients, the imminent risk of death may have precluded them from reaching the point where a stroke event could be observed, thereby obscuring a difference in effectiveness outcomes.

HT demonstrates higher incidence in AF patients with AIS undergoing EVT, predominantly manifesting as asymptomatic radiologic findings within 36 h post-procedure (15). Anticoagulation therapy needs to be restarted after individualized evaluation, and there is no specific time for initiation of anticoagulation. ELAN trial subanalyses demonstrated early direct oral anticoagulant initiation in HT patients did not significantly increase symptomatic intracranial hemorrhage incidence. However, functional outcomes may be compromised in those with parenchymal hematoma conversion (27). Escudero et al. conducted a multicenter analysis of 1,489 nonvalvular atrial fibrillation (NVAF) patients from the SITS registry cohort receiving acute stroke interventions [intravenous thrombolysis (IVT), EVT, or combined treatment]. Their findings demonstrated the safety profile of initiating dabigatran within 1–3 days post-stroke, with minimal thromboembolic or hemorrhagic events observed during 3-month follow-up across all treatment modalities (2.2%) (28). In this study, 72 patients developed HT post-EVT, but no clear classification study was carried out. Of these, 14 patients resumed OAC at discharge following multidisciplinary assessment. Comparative analysis demonstrated no significant between-group disparity in 1-year hemorrhagic risk despite comparable HT incidence. Notably, one non-anticoagulated patient receiving dual antiplatelet therapy (DAPT) experienced symptomatic intracranial hemorrhage during follow-up. Consequently, patients developing post-EVT HT necessitate meticulous evaluation of antithrombotic regimen-related bleeding risks, while the net clinical benefit of early anticoagulation initiation remains uncertain, the large-scale randomized trials remain essential to establish evidence-based OAC resumption timelines in this high-risk population.

AF can be divided into KAF and AFDAS depending on the time of diagnosis. In stroke involving the insula, cerebral ischemia and cell death may affect the autonomic nervous system and cause AF, so it is uncertain whether AFDAS is the result of brain injury or the cause of stroke associated with multiple neurogenic mechanisms, it remains unsubstantiated by relevant evidence (29). Patients with AFDAS demonstrate distinct vascular risk profiles, including significantly lower rates of both baseline comorbidities and subsequent stroke events than KAF (30). The patients with AFDAS may have a lower atrial burden, a smaller left atrium, and a higher LVEF than those with KAF, with no differences in age, sex, stroke severity, or mortality (31–33). Ryu et al. investigated the relationship between detection time and functional outcomes of AF after mechanical thrombectomy, and the results showed that AFDAS was positively correlated with functional independence (34). Another study demonstrated that KAF demonstrates an elevated HT risk vs. AFDAS, but without concomitant mortality risk elevation (35). OAC is recommended across all AF subtypes to mitigate recurrent stroke risk. However, the net clinical benefit of OAC therapy in patients with AFDAS remains uncertain, because patients with AFDAS tend to exhibit more benign features. In this study, whether AF was AFDAS showed no significant interaction with functional or efficacy outcomes. Instead, anticoagulation was consistently linked to better functional outcomes across all AF subtypes. This suggests that early anticoagulation may confer a prognostic benefit regardless of the specific atrial fibrillation subtype. Therefore, more data from randomized controlled trials are warranted to determine the risk-benefit ratio of anticoagulation therapy in AFDAS populations.

This study had several limitations. First, it was a single-center design with a limited sample size, which may introduce selection bias. Second, this study did not perform a detailed analysis of OAC regimens, including specific drug types and dosage adjustments. Third, while we assessed the impact of timely OAC initiation at discharge, we did not evaluate longitudinal changes in anticoagulation therapy during the 1-year follow-up period. Notably, patients with irregular medication adherence or delayed post-discharge OAC initiation were excluded from analysis. Fourth, we acknowledged that PSM may not have eliminated all selection bias and confounding, including differences in frailty, fall risk, baseline cognitive function, and independence. Future randomized controlled trials are needed to confirm these observations. Consequently, our findings primarily reflect the prognostic benefits of early, guideline-adherent anticoagulation in compliant patients. Thus, further randomized controlled trials are warranted to validate these observations.

This study offers novel real-world evidence supporting the role of post-EVT anticoagulation therapy in AF-associated AIS. Analysis indicates that among stroke patients clinically selected for anticoagulation therapy and without high risk of severe bleeding, OAC therapy is associated with improved functional and mortality outcomes compared with those not receiving OAC. The benefit remained statistically significant following PSM to adjust for intergroup disparities.