Identifying the cause of ischemic stroke and prescribing appropriate secondary prevention therapies are critical in reducing the risk of recurrence. Many patients with stroke present with well-defined etiologies, such as large artery atherosclerosis (LAA) or cardioembolism, enabling clinicians to tailor preventive strategies accordingly [1]. However, a significant proportion of patients still present with undetermined stroke mechanisms, complicating the delivery of optimal management strategies [2,3]. Among them, a notable and increasingly recognized subgroup consists of individuals with both atrial fibrillation (AF) and LAA. This group continues to grow with the aging population and wider use of prolonged cardiac monitoring [4-7].

This clinical overlap has prompted ongoing investigations into whether a combination of oral anticoagulants (OACs) with antiplatelet agents, typically indicated for AF and LAA, respectively, may improve outcomes in patients presenting with both cardioembolic and atherosclerotic stroke risk factors. Several observational studies investigating the role of combination therapy (OAC plus antiplatelets) in stroke patients with AF and LAA have reported contradictory results regarding both efficacy and safety [8-15]. A related clinical question has been addressed in patients with AF and coexisting coronary artery occlusive disease (CAOD), for which recent randomized trials and a meta-analysis have shown that OAC monotherapy may offer comparable protection against thrombotic events, while significantly lowering bleeding risk [16-19]. However, there is a lack of both randomized controlled trials and comprehensive systematic reviews or meta-analyses directly comparing OAC monotherapy with combination therapy involving antiplatelet agents for stroke patients with AF and LAA.

Therefore, we conducted a systematic review and meta-analysis to compare the efficacy and safety of combination therapy with OAC and antiplatelet agents with those of OAC alone in ischemic stroke patients with both AF and LAA as causes of stroke. We further examined whether the impact of combination therapy differed according to treatment duration, with the goal of informing individualized therapeutic strategies for this patient population.

The coexistence of AF and cerebral atherosclerosis constitutes a uniquely high-risk stroke population, reflecting the combined impact of two major stroke etiologies: cardioembolism and LAA. Patients with both conditions are increasingly encountered in clinical practice, particularly in aging populations, where both AF and the atherosclerotic burden are increasing [4,5,22]. However, there is a substantial lack of high-quality data to guide antithrombotic management in this group. In contrast to CAOD, in which dozens of randomized trials have defined the optimal timing and duration of combination antithrombotic regimens [23,24], evidence in stroke patients with overlapping etiologies remains sparse [25]. Current clinical guidelines recommend administering OAC monotherapy in patients with AF, as adding antiplatelet therapy provides little benefit, but increases bleeding risk [26]. However, in patients with potentially symptomatic ICAS or ECAS, evidence to guide treatment decisions is insufficient. The present meta-analysis addresses this gap by evaluating the efficacy and safety of combination therapy with OAC and antiplatelet agents (vs. OAC monotherapy) in stroke patients with AF and LAA.

The pooled results in this meta-analysis demonstrated that long-term combination therapy with OAC and antiplatelet agents failed to significantly reduce the rate of recurrent ischemic stroke compared to OAC monotherapy. However, this strategy is associated with a substantially increased risk of major bleeding, including intracranial hemorrhage. These findings are consistent with large-scale data in cardiology, where prolonged combination therapy has been found to increase bleeding complications without proportional ischemic benefit [24,26]. However, direct extrapolation from AF populations with coronary artery disease should be approached with caution because the therapeutic context of cerebrovascular atherosclerosis differs fundamentally from that of coronary artery disease. Coronary lesions are frequently treated with percutaneous revascularization, and antithrombotic therapy in that setting is primarily directed toward preventing stent-related thrombosis during a defined healing period. By contrast, cerebrovascular plaques remain persistently stenotic without structural correction, generating different hemodynamic environments, patterns of plaque instability, and embolic mechanisms. These mechanistic differences underscore the need for cerebrovascular-specific evidence rather than reliance on data derived from AF with coronary artery disease. In particular, the risk of intracranial hemorrhage is notably higher in Asian populations, which are more vulnerable to bleeding complications under antithrombotic therapy regimens [13,27,28]. These findings reinforce current guideline recommendations favoring OAC monotherapy as the standard of care for secondary prevention in stroke patients with concomitant AF and cerebral atherosclerosis, and caution against the routine use of combination therapy beyond specific, time-limited indications.

While long-term combination therapy appears detrimental, our analysis indicates that the short-term use of OAC plus antiplatelet therapy, typically within the first 3 months (i.e., 30- 90 days) after the index stroke, may reduce the risk of early recurrent ischemic events, particularly in patients in whom LAA is a contributing mechanism. This early vulnerability period is characterized by vascular inflammation, active platelet aggregation, and plaque instability [29,30]. These pathophysiologic changes may transiently increase the risk of artery-to-artery embolism, particularly among patients with high-grade symptomatic stenosis or ulcerated plaques. Evidence from stroke trials, such as the Clopidogrel in High-Risk Patients With Acute Non-Disabling Cerebrovascular Events (CHANCE) and Platelet-Oriented Inhibition in New TIA and Minor Ischemic Stroke (POINT) trials strongly supports the use of short-term dual antiplatelet therapy in non-cardioembolic stroke, where dual antiplatelet therapy for a short period (<21 days) reduced early recurrence with an acceptable safety profile [31]. Although these trials did not enroll patients with AF, the biological rationale may extend to AF patients with mixed etiologies, particularly in the acute phase.

The potential benefits of short-term combination therapy may be particularly relevant among patients with stroke, which is more plausibly attributable to atherosclerotic mechanisms than to cardioembolism. Such considerations highlight that antithrombotic strategies should not be uniform across all AF patients with stroke, but should instead be determined by the individualized underlying stroke mechanism and adapted through time-dependent therapy adjustment. Patients with symptomatic atherosclerotic lesions or unstable plaques may derive short-term benefits from the addition of an antiplatelet agent, provided that this is followed by timely de-escalation to OAC monotherapy [31]. This is analogous to the transient use of dual antiplatelet therapy following coronary or carotid stenting [26], which may help balance the efficacy and safety of the use of OAC with antiplatelets.

Taken together, our findings provide a practical treatment framework for this patient group, showing that chronic OAC monotherapy is the safest and most effective long-term treatment option for patients with AF and coexisting LAA. However, short-term combination therapy may be justified in selected patients with high-risk atherothrombotic features. Such decisions should be guided by careful assessment of the bleeding risk, individualized patient characteristics, and informed shared decision-making. Importantly, the duration of combination therapy should be limited to the early post-stroke period, with de-escalation to monotherapy once the high-risk window has passed. Ethnic and regional differences in susceptibility to bleeding must also be considered when applying these strategies to diverse patient populations.

Further, this meta-analysis highlights the urgent need for prospective, well-powered, randomized controlled trials to guide antithrombotic strategies in patients with AF and significant cerebrovascular atherosclerosis. Future trials should clearly define treatment duration, drug classes, and endpoints and incorporate centralized adjudication of ischemic and hemorrhagic events. Mechanism-driven subgroup stratification and modern imaging are critical for patient selection. Several important studies in this field are currently at different stages of investigation, including the completed ADD-ON trial (adding antiplatelets during edoxaban treatment in stroke patients with nonvalvular atrial fibrillation [NCT04010955]) and ATIS-NVAF trial (optimal antithrombotic therapy in ischemic stroke patients with nonvalvular atrial fibrillation and atherothrombosis [NCT03062319]), the ongoing ALLY trial (combination of antiplatelet and anticoagulation for AIS patients with concomitant NVAF and extracranial/intracranial artery stenosis [NCT06058130]), and the planned BEACON-AA trial (Benefits of Apixaban and Clopidogrel On stroke prevention in patients with Atrial fibrillation and cerebral Atherosclerosis). These studies vary in their designs, particularly regarding treatment duration and primary endpoints, thereby addressing different therapeutic windows. The BEACONAA and ALLY trials focus on the high-risk acute-to-subacute phase by administering combination therapy for 1 month and 3 months, respectively, in patients with symptomatic intracranial or extracranial stenosis. In contrast, the ADD-ON study and ATIS-NVAF trials were designed to evaluate the safety and efficacy of long-term combination therapy ranging from 1.5 to 2 years. Notably, the recently published ATIS-NVAF randomized trial reported that long-term combination therapy did not reduce ischemic cardiovascular events but significantly increased major and clinically relevant non-major bleeding, underscoring the safety concerns associated with prolonged dual therapy [28]. Furthermore, regarding endpoints, BEACON-AA utilizes recurrent ischemic lesions on magnetic resonance imaging as a surrogate marker to detect early efficacy, whereas the other trials assess composite clinical events. Highlighting these differences clarifies how short-term intensification may complement long-term maintenance strategies in future guidelines (Supplementary Table 5). These trials will provide essential data to support evidencebased guidelines. However, until such results are available, clinicians must rely on the best available evidence and clinical judgment, applying time-limited combination therapy only in strictly defined high-risk situations with an emphasis on de-escalation and safety monitoring.

Despite the consistent patterns observed, our analysis had important limitations. Most of the included studies were retrospective cohort analyses or registry-based studies, with inherent risks of selection bias and confounding factors. Given that the meta-analysis incorporated crude event counts from observational cohorts, unmeasured confounding and treatment-selection bias are unavoidable. These methodological constraints limit causal inference, and the conclusions should therefore be viewed as exploratory pending confirmation in randomized studies. Significant heterogeneity was identified in the antithrombotic regimens, including differences in the OAC type (vitamin K antagonists [VKAs] vs. DOACs), choice of antiplatelet agent, treatment duration, and outcome definition. Notably, the definition of composite outcomes varied substantially across the included studies. While some studies used predefined composite endpoints combining ischemic stroke, myocardial infarction, and all-cause mortality, others incorporated major bleeding or systemic embolism into their composite measures. This heterogeneity limits the clinical interpretability of the pooled composite outcome and underscores that these results should be considered exploratory rather than definitive. However, the ability to perform detailed stratified analyses was limited by the underlying data. Only one study provided DOAC-exclusive outcomes, and most cohorts included mixed VKA+DOAC populations, preventing a separate DOAC-only meta-analysis. In addition, P2Y12 inhibitor-specific results were not available in the included studies, precluding antiplatelet class-specific analyses that would reflect contemporary practice. Accordingly, the findings of this meta-analysis should not be generalized to specific contemporary regimens such as DOAC plus P2Y12 inhibitor combinations, which were not specifically evaluated in the included studies. These constraints arise from limitations in the existing evidence base rather than the analytic approach and highlight the need for standardized reporting in future studies. Additionally, much of the data was sourced from East Asian populations, where ICAS is more prevalent and bleeding risk under dual antiplatelet therapy is known to be higher [27]. Thus, while this analysis provides clinically valuable insights for these patients, this limits generalizability to Western populations with different risk profiles and genetic backgrounds. In addition, two included studies defined large-artery stenosis using ICD-9/10 diagnostic codes rather than imaging-based confirmation, which risks introducing misclassification of LAA status due to the lower specificity of administrative coding. However, in exploratory subgroup analyses stratified by stenosis ascertainment modality, the overall direction of the treatment effect remained generally consistent between imaging-based and ICD-based cohorts. Although these findings should be interpreted with caution, this suggests that differences in tools to define LAA did not appear to have substantially influenced the main conclusions of our analysis. Detailed information on antiplatelet agent subtype, lesion location, including distinctions between intracranial and extracranial atherosclerosis, and stenosis severity was also not consistently available, which limited further stratified analyses. These variables likely represent important sources of heterogeneity. Moreover, the real-world attribution of stroke mechanisms is often imprecise, particularly when patients harbor both AF and atherosclerotic lesions. Future studies should incorporate high-resolution vessel wall imaging, advanced cardiac monitoring, and standardized endpoint adjudication to improve the precision of the mechanism-specific patient classification.

Taken together, these sources of clinical and methodological heterogeneity indicate that the findings of this meta-analysis should be interpreted as hypothesis-generating rather than prescriptive for clinical practice. Although consistent patterns emerged across observational cohorts, the absence of randomized evidence, coupled with limitations in treatment definitions and drug-specific reporting, constrains the strength of any causal inference. Notably, since most included studies enrolled mixed populations treated with VKA and DOAC, these findings should not be directly extrapolated to current DOAC-based antithrombotic strategies. Accordingly, practice-changing recommendations should await rigorously designed randomized trials that apply standardized criteria, contemporary DOAC-based regimens, and clearly defined antiplatelet subclasses.

This analysis showed that while long-term OAC plus antiplatelet therapy increases bleeding risk without preventing recurrent stroke, short-term combination therapy (<3 months) may offer benefits. Randomized controlled trials are required to confirm the efficacy and safety of these therapeutic strategies.