Atrial fibrillation (AF), the most commonly diagnosed sustained cardiac arrhythmia, affects approximately 743 and 664 individuals per 100,000 population worldwide and in Asia (Linz et al., 2024), and is associated with an increased and heterogeneous risk of stroke depending on the presence of individual risk factors (Chao et al., 2024). Chronic obstructive pulmonary disease (COPD) affects cardiac rhythm through multiple interacting pathways involving functional and structural remodeling (Simons et al., 2021), and is related to a higher risk of incident AF (Liao and Chen, 2017; Grymonprez et al., 2019). Beyond its potential role in the development of AF, COPD may also adversely affect patients who already have AF. Compared to those with AF alone, patients with concomitant COPD have been observed to exhibit higher risks of ischemic stroke (Nadeem et al., 2015; Tsai et al., 2025) and bleeding events (Durheim et al., 2018; Rodríguez-Mañero et al., 2019). In addition, COPD exacerbations may contribute to subsequent cardiovascular diseases, including ischemic stroke (Simons et al., 2024).

Since 2009, the pivotal randomized controlled trials have established the efficacy and safety of non–vitamin K antagonist oral anticoagulants (NOACs) for stroke prevention in AF, leading to their widespread adoption as the cornerstone of stroke prevention (Joglar et al., 2024; Van Gelder et al., 2024). Despite the established efficacy of NOACs, data from the Global Anticoagulant Registry in the Field–Atrial Fibrillation (GARFIELD-AF), encompassing patients from multiple countries worldwide, indicated that around 30% of individuals with AF received antiplatelet agents alone or no stroke-prevention therapy (Steinberg et al., 2017; Fox et al., 2021). This finding highlights a substantial treatment gap between clinical evidence and real-world practice in stroke prevention among patients with AF.

In this context, although COPD may influence the pathophysiology, symptom burden, and therapeutic decision-making of AF, the impact of COPD on stroke-prevention medication use among patients with AF remains unclear. To date, no study has comprehensively characterized the prescription patterns of stroke-prevention therapies in AF patients with concomitant COPD. Prior investigations concluded in 2015 (Liao et al., 2023), underrepresenting contemporary practice amid the rapid uptake of NOACs, evolving guideline recommendations, and changing perceptions of ischemic stroke and bleeding risk in patients with COPD. A current, population-level assessment was therefore conducted among patients with AF concomitant with COPD to characterize temporal trends in the use of oral anticoagulants (OACs) and oral antiplatelet agents (OAPT), to investigate concurrent temporal trends in ischemic stroke and major bleeding, and to further examine whether prescribing trends differed according to COPD exacerbation history in this specific, high-risk population.

A total of 632,635 patients were enrolled between 1 January 2013, and 31 December 2022. During follow-up, 27,356 individuals within the defined COPD cohort (n = 339,795) developed AF based on the predefined criteria. After excluding patients whose AF diagnosis occurred prior to their COPD diagnosis and patients with less than 1 year of follow-up, a total of 13,072 patients remained eligible for the final prescription pattern analysis. The cohort selection process is illustrated in Figure 1.

In the study population, the mean (SD) age was 75.0 (10.2) years, and the majority of patients were male (n = 9,066; 69.4%). The most prevalent comorbidities included hypertension (n = 8,898; 68.1%), diabetes (n = 4,007; 30.7%), hyperlipidemia (n = 3,220; 24.6%), heart failure (n = 4,727; 36.2%), and ischemic heart disease (n = 4,881; 37.3%). Over 87% of population accompanied with a CHA₂DS₂-VASc score ≥2, reflecting a moderate-to-high baseline ischemic stroke risk in this population. Overall, most baseline characteristics were well balanced between patients with and without a history of COPD exacerbation. However, notable imbalances were observed in sex distribution (male: 75.4% vs. 67.5%; ASMD = 0.18) and the prevalence of heart failure (41.1% vs. 34.7%; ASMD = 0.13). Detailed baseline characteristics for both groups are presented in Table 1.

The temporal trends in the prescription patterns of stroke-prevention medications among patients with COPD and AF from 2013 to 2022 is shown in Figure 2. In 2013, OAPTs were the predominant option for stroke prevention (74.6%), followed by warfarin (24.0%), whereas the use of NOACs was relatively limited (15.4%). Over time, NOAC prescriptions increased substantially, rising from 15.4% in 2013 to 65.4% in 2022, and rapidly overtook warfarin, which declined steadily to 6.1% by 2022. The overall proportion of patients receiving any OAPT decreased modestly, from 74.6% in 2013 to 51.7% in 2022. The temporal trends in all drug categories were statistically significant (p-value for trend <0.0001).

As shown in Figure 3, patient-level prescription patterns of stroke-prevention therapies demonstrated substantial changes from 2013 to 2022. In 2013, half of the patients with COPD and AF (50.0%) received OAPT only within 1 year after their AF diagnosis, making it the most common regimen. Receiving warfarin plus OAPT was the second most frequent pattern (12.8%), followed by NOAC plus OAPT (7.6%) and warfarin only (6.0%). Non-users accounted for 15.7% of the cohort in 2013. Marked shifts occurred over the study period. The proportion of patients receiving NOAC only increased more than tenfold, rising from 2.6% in 2013 to 33.1% in 2022. Similarly, NOAC plus OAPT demonstrated a steady and substantial increase from 7.6% to 29.6%, and had become the second most common regimen after 2018. In contrast, use of warfarin plus OAPT exhibited a progressive decline from 12.8% in 2013 to 1.7% in 2022, while warfarin only also decreased from 6.0% to 1.7% over the period. The proportion of patients receiving OAPT only dropped sharply from 50.0% to 19.0%, reflecting a shift away from antiplatelet-based strategies. The prevalence of NOAC-and-warfarin pattern remained consistently low (<5% annually), and receiving three kinds of stoke-prevention medications (warfarin + NOAC + OAPT) remained a rare practice throughout the study window. The proportion of non-users decreased modestly from 15.7% to 12.2%, suggesting improved adoption of stroke-prevention therapies over time. The Cochran–Mantel–Haenszel trend test confirmed significant temporal changes across all treatment categories (p < 0.0001), highlighting a major and sustained shift from warfarin- and OAPT-centered regimens toward NOAC-dominant therapeutic strategies.

From 2013 to 2022, the annual distribution of stroke-prevention medications shifted substantially toward NOAC-based therapy. Over the same period, population-level incidence rates of ischemic stroke and major bleeding declined significantly (Figure 4). The 1-year ischemic stroke incidence rate (95% CI) decreased from 19.5 (16.8–22.6) per 100 person-years in 2013 to 14.8 (12.7–17.2) in 2022 (IRR 0.975, 95% CI 0.959–0.991; p = 0.0025). Major bleeding (95% CI) similarly fell from 21.9 (19.0–25.1) to 16.0 (13.9–18.5) per 100 person-years (IRR 0.965, 95% CI 0.951–0.979; p < 0.0001).

During the follow-up period, no statistically significant differences were observed between patients with and without a history of COPD exacerbation in the temporal trends of drug prescriptions. The trend comparison between the two groups showed no significant divergence in the trajectories of warfarin (p = 0.3819) or NOAC use (p = 0.1119). Similarly, the temporal patterns of OAPT use (p = 0.8153) and non-user proportions (p = 0.8385) did not differ significantly between the two subgroups (Figure 5). These findings suggested that although the overall directions of change differed across medication classes, the temporal trends were similar between patients with and without prior COPD exacerbations.

Using a nationwide claims database in Taiwan, this cohort study investigated the prescription patterns of stroke-prevention medications and their temporal changes in recent years among COPD patients with newly diagnosed AF in the real-world clinical settings.

During the 10-year enrollment period, we found that approximately 8.1% of patients with general COPD had comorbid AF, which is consistent with the finding of previous study of 4.7%–15.0% (Papaporfyriou et al., 2023). Xiao et al. also observed prevalence of AF was up to 18.1% among patients with end-stage COPD on home oxygen who were hospitalized for COPD exacerbation (Xiao et al., 2019). On the other hand, we reported 5.2% incident cases of AF patients diagnosed after COPD, whereas a 9.8% newly diagnosed AF was reported in the COPD cohort observed by Liao et al. (Liao and Chen, 2017). The lower incidence in our study may reflect stricter eligibility criteria—including medication and procedure codes related to arrhythmia—that likely improved AF case ascertainment while restricting the eligible population.

Although several studies have examined stroke-prevention prescribing patterns in patients with AF and comorbid COPD, their findings have been inconsistent and have not provided a comprehensive characterization of the overall prescribing landscape in this population. In the Global Registry on Long-Term Antithrombotic Treatment in Patients with Atrial Fibrillation (GLORIA-AF) registry, AF patients with comorbid COPD had higher odds of receiving OACs than those without COPD (adjusted odds ratio 1.29, 95% CI 1.13–1.47) (Romiti et al., 2023). By contrast, the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) reported similar OACs use between AF patients with and without COPD (76.1% vs. 76.4%, p = 0.763) (Durheim et al., 2018). These divergent findings may reflect differences in geographic coverage (GLORIA-AF is multinational, whereas ORBIT-AF is American-based), study design as well as eligibility criteria, and local practices (Huisman et al., 2014; Steinberg et al., 2014).

Liao et al. conducted the study using Taiwan’s NHIRD and reported that 31.1% of patients were prescribed low-dose aspirin within the first 30 days after AF diagnosis, whereas only a limited proportion received dabigatran (1.7%) or rivaroxaban (1.6%) (Liao et al., 2023). However, the enrollment period in Liao et al.‘s study was 2002–2015. To our knowledge, the reimbursement for AF-related stroke prevention under Taiwan’s National Health Insurance was introduced in the following years: dabigatran (2012), rivaroxaban (2013), apixaban (2014), and edoxaban (2016). Accordingly, our study can be considered an essential extension and refinement of their work. By incorporating data from more recent years, our study captured the contemporary landscape of stroke-prevention medications—reflecting reimbursement-driven uptake of NOACs and evolving practice patterns after 2016—thereby providing a more complete picture of prescribing in COPD patients with AF. However, because we did not include a contemporaneous AF cohort without COPD, we could not directly quantify whether COPD status modified the pace of NOACs adoption beyond temporal trends in AF management during 2013–2022. Accordingly, our findings should be interpreted as descriptive temporal trends among patients with AF and COPD. Nevertheless, quantifying real-world prescription pattern in this high-risk and clinically complex population remains meaningful, as COPD is frequently accompanied by greater comorbidity burden and treatment complexity.

With respect to the overall AF population—regardless of COPD status—several registry cohorts have reported the use of stroke-prevention medications; however, gaps remain in the literature, particularly for the most recent years. GARFIELD-AF, which recruited over 50,000 AF patients from 2010 to 2016 across 35 countries, showed that NOAC use, with or without OAPT, increased substantially during follow-up (from 3% in 2010 to 43% in 2016) (Steinberg et al., 2017; Kakkar et al., 2012). In our study, a similar trend was observed, revealing that 10.2% of patients using NOAC with or without OAPT in 2013, increasing to 35.3% in 2016, and 62.7% in 2022. Compared with the ORBIT-AF launched in the United States, the penetration of NOAC in Taiwan was relatively lower (71% vs. 35.3% in 2016) (Steinberg et al., 2017; Steinberg et al., 2014), which may be attributed to the higher bleeding risk reported in Asian or Chinese populations (Kim et al., 2021). As a result, OAPT was more frequently adopted as the preferred treatment option (Chiang et al., 2015).

Incorporating ecological outcome trends provides clinical context for the pronounced transition from warfarin- or OAPT-based strategies to NOACs in this high-risk COPD-AF population. At the population level, incidence rates of ischemic stroke and major bleeding declined over time, in parallel with the period of increasing NOAC uptake. However, these population-level associations should be interpreted cautiously. Because patient risk profiles, diagnostic coding practices, and other aspects of care changed concurrently over time, this ecological analysis cannot directly attribute outcome changes to NOAC adoption or infer causal treatment effects.

An exacerbation of COPD can precipitate both AF and ischemic stroke through converging pathways. Ventilation–perfusion mismatch with hypoxemia creates an oxygen supply–demand imbalance and, together with oxidative stress, dynamic hyperinflation, and sympathetic activation, promotes atrial remodeling and autonomic instability, heightening susceptibility to AF. In parallel, intercurrent infection and systemic inflammation drive endothelial dysfunction, platelet activation, and coagulation cascade upregulation, fostering a hypercoagulable status that predisposes to ischemic stroke (Simons et al., 2024). Beyond mechanistic plausibility, both moderate and severe COPD exacerbations have been observed to be associated with subsequent cardiovascular disease (CVD), with incidence peaking up to 20-fold hazard in the early post-exacerbation period (e.g., within weeks) and then gradually declining, yet remaining significantly elevated for up to 1 year after the exacerbation (Dransfield et al., 2022; Swart et al., 2023; Hawkins et al., 2024). However, in our analysis stratified by COPD exacerbation history, the post-AF prescribing trends for stroke-prevention medications did not differ significantly; even the prescribing proportions were similar, as suggested by the interweaving curves (Figure 5). Several factors may underlie these observations. First, clinical awareness of the implications of COPD exacerbations may remain heterogeneous, and current understanding may be insufficient to influence downstream prescribing decisions. Second, in our baseline assessment, patients with and without prior exacerbations exhibited similar CHA₂DS₂-VASc score distributions, suggesting generally guideline-concordant practice in Taiwan, which might overshadow the specific impact of COPD exacerbations on stroke-prevention medications choice. Third, according to clinical guidelines, COPD exacerbations are characterized by increased respiratory symptoms and require specific clinical assessments (GOLD Executive Committee, 2025). However, such detailed evaluations cannot be fully captured in claims-based data. Our findings may reflect the lack of granular clinical data on COPD exacerbation characteristics (e.g., laboratory assessments such as blood eosinophil counts and C-reactive protein, or specific triggers). In this study, we used COPD primary diagnosis in combination with prescriptions of oral or parenteral antibiotics or steroids as a proxy indicator to classify whether patients had a history of exacerbations (Larsson et al., 2013; Wang et al., 2021), which may underestimate the occurrence of COPD exacerbations. Nevertheless, given prior evidence that COPD exacerbations substantially increase subsequent CVD risk, this population still warrants particular attention to ensure comprehensive care.

The complete treatment patterns and temporal trends of stroke-prevention medications in patients with COPD and AF were examined using data up to the most recent available year. By incorporating contemporary data spanning the past decade, we observed a clear shift from OAPT- or warfarin-based regimens—which previously constituted the backbone of stroke prevention—to a predominant reliance on NOACs in recent years. This transition mirrors the prescribing patterns reported in the general AF population, suggesting that the evolution of clinical practice toward NOAC-centered therapy has similarly extended to patients with concomitant COPD. Concurrent trends in ischemic stroke and major bleeding incidence rates were examined alongside prescribing trends; both clinical outcomes declined significantly over time. Although a history of COPD exacerbation did not appear to be associated with the use of stroke-prevention medications, this subgroup warrants close attention given the markedly increased risks of CVD and stroke following exacerbations.