Highlights
In a large patient-level meta-analysis of the four pivotal randomized trials of direct oral anticoagulants (DOACs) versus warfarin in atrial fibrillation, Asian patients had worse baseline risk profiles than non-Asians, including younger age, lower body weight, poorer renal function, and more prior stroke or transient ischemic attack.
Among patients receiving warfarin, Asians experienced higher adjusted risks of stroke/systemic embolic events, major bleeding, intracranial hemorrhage, gastrointestinal bleeding, and the composite net clinical outcome, in the setting of lower median time in therapeutic range.
Standard-dose DOACs outperformed warfarin more strongly in Asians than non-Asians for stroke prevention, major bleeding, and net clinical benefit, while lower-dose DOACs were associated with more thromboembolic events than standard-dose therapy in Asians.
Background
Atrial fibrillation is a major driver of ischemic stroke, systemic embolism, heart failure, and premature mortality worldwide. Oral anticoagulation remains the cornerstone of stroke prevention, but the optimal agent and dose can vary across populations. For more than a decade, direct oral anticoagulants have been favored over warfarin in many patients because they reduce intracranial hemorrhage and avoid the narrow therapeutic window and frequent monitoring required with vitamin K antagonists. Even so, many clinicians have remained cautious about anticoagulant selection and dosing in Asian patients, partly because of concerns about bleeding risk, lower body size, different patterns of warfarin management, and regional prescribing habits.
These concerns matter because Asia carries a large and growing atrial fibrillation burden, and stroke prevention remains uneven across health systems. Some earlier analyses suggested that Asian patients may have a different balance of thrombotic and bleeding risk, but high-quality randomized evidence comparing treatment effects by race has been limited. The COMBINE AF collaboration addresses this gap by pooling patient-level data from the four pivotal DOAC trials and examining whether the relative benefits and harms of DOACs versus warfarin differ between Asian and non-Asian patients.
Study Design
Data source and population
This patient-level meta-analysis used data from COMBINE AF, which integrated the randomized evidence from the four landmark DOAC trials in atrial fibrillation. The analysis included 10,212 Asian patients and 61,471 non-Asian patients. Because patient-level data were available, the investigators could compare baseline features, event rates, and treatment effects within and across racial groups with greater precision than would be possible with aggregate trial summaries.
Comparisons and endpoints
The investigators first compared clinical characteristics between Asian and non-Asian participants. They then evaluated outcomes in the warfarin arms and examined the relative efficacy and safety of standard-dose DOACs versus warfarin, testing for interaction between race and treatment effect. They also explored whether body weight and creatinine clearance modified outcomes among Asian patients. Key outcomes included stroke or systemic embolic events (SEE), major bleeding, intracranial hemorrhage, gastrointestinal bleeding, and a primary net clinical outcome defined as stroke/SEE, major bleeding, or death. A secondary net clinical outcome also incorporated intracranial hemorrhage.
Key Findings
Baseline differences between Asian and non-Asian patients
Compared with non-Asians, Asian patients were on average 3.2 years younger and approximately 20 kg lighter. They also had worse renal function, with mean creatinine clearance of 64.9 mL/min versus 77.3 mL/min in non-Asians, and a substantially higher prevalence of prior stroke or transient ischemic attack, 37.2% versus 26.6% for non-Asians. These differences were highly significant, with P < .001 for each comparison.
These baseline patterns are clinically important because age, weight, and kidney function all influence anticoagulant exposure, drug selection, and bleeding risk. They also reinforce why a one-size-fits-all approach to anticoagulation is unlikely to be optimal.
Warfarin performance was less favorable in Asians
In the warfarin arm, the median time in therapeutic range was lower in Asians than in non-Asians, 57.7% versus 66.2%, again with P < .001. Time in therapeutic range reflects how often the international normalized ratio remains in the desired window; lower values generally mean less stable anticoagulation and more risk for both thrombosis and bleeding.
Against this background, Asian patients assigned to warfarin had a higher adjusted risk of stroke/systemic embolic events, major bleeding, intracranial hemorrhage, gastrointestinal bleeding, and the primary net clinical outcome than non-Asians. In practical terms, warfarin appeared to perform worse in Asian patients, likely reflecting both pharmacologic sensitivity and less favorable anticoagulation control in this subgroup.
Standard-dose DOACs showed stronger benefit in Asians
When standard-dose DOACs were compared with warfarin, the treatment effect was more favorable in Asians than in non-Asians for several clinically meaningful outcomes. For stroke/systemic embolic events, the hazard ratio was 0.65 in Asians versus 0.86 in non-Asians, with a significant interaction by race (P for interaction < .02). For major bleeding, the hazard ratio was 0.62 in Asians versus 0.91 in non-Asians, again favoring DOACs more strongly in Asians. For the primary net clinical outcome, the hazard ratio was 0.76 in Asians versus 0.94 in non-Asians, also with significant interaction.
These are not just statistically significant differences; they are clinically relevant. A hazard ratio of 0.65 for stroke/SEE means that standard-dose DOACs were associated with a 35% relative reduction versus warfarin in Asians, whereas the reduction in non-Asians was more modest. Similarly, the observed benefit for major bleeding suggests that, in this dataset, standard-dose DOACs were both safer and more effective than warfarin in Asian patients.
Gastrointestinal bleeding did not increase in Asians
One of the most practically important observations was the contrast in gastrointestinal bleeding. Standard-dose DOACs increased gastrointestinal bleeding only in non-Asians, where the hazard ratio was 1.41, compared with no increase in Asians, where the hazard ratio was 0.92. The interaction was significant (P for interaction = .009).
This finding is noteworthy because gastrointestinal bleeding is often cited as a reason to prefer warfarin in patients perceived to be at bleeding risk. In this analysis, that concern was not supported for Asian patients receiving standard-dose DOACs. The data instead suggest that, at least in the trial populations studied, standard-dose DOACs did not carry an excess gastrointestinal bleeding penalty in Asians.
Benefit across body weight and renal function in Asians
The analysis also explored whether standard-dose DOACs remained beneficial across the range of body weight and creatinine clearance seen in Asian participants. The authors reported that standard-dose DOACs reduced the risk of clinical events across a wide range of both variables. This is important because many Asian patients are smaller and have different renal function profiles than the average non-Asian trial participant, and clinicians often hesitate to use full-dose therapy in these settings.
While dose adjustment remains essential when it is indicated by labeling, these subgroup findings argue against reducing DOAC doses simply because a patient is Asian or has a relatively low body weight within the studied range.
Lower-dose DOACs were associated with worse ischemic outcomes than standard-dose therapy
Perhaps the most practice-changing result was the comparison of lower-dose versus standard-dose DOACs in Asian patients. Lower-dose DOACs significantly increased the risk of stroke/systemic embolic events, with a hazard ratio of 1.57, and also increased the secondary net clinical outcome, with a hazard ratio of 1.23. In other words, lowering the dose was not a benign strategy; it was associated with more thromboembolic harm without an obvious offsetting safety advantage in the reported outcomes.
This result should be interpreted carefully. It does not mean that approved dose-reduction criteria should be ignored. Rather, it suggests that empirical underdosing based on ethnicity alone is not supported. Dose reduction should continue to follow drug-specific indications such as renal impairment, age, concomitant drug interactions, or other label-based criteria.
Expert Commentary
This study strengthens the evidence that DOACs should generally be preferred over warfarin in atrial fibrillation, including in Asian patients. The patient-level design is a major strength because it allows more rigorous subgroup analyses than a conventional study-level meta-analysis. The large sample size also improves confidence that the observed interaction patterns are not merely random noise.
At the same time, several limitations should temper overinterpretation. First, race is an imperfect proxy for biological ancestry, diet, health-system factors, and prescribing behavior. Second, most Asian trial participants in pivotal DOAC studies have historically come from East Asian populations, so the findings may not fully generalize to South Asian, Southeast Asian, or multi-ethnic populations. Third, the analysis is based on randomized trial populations, which tend to be healthier and more closely monitored than patients seen in routine practice. Fourth, the lower-dose DOAC analysis should not be conflated with approved dose-reduction strategies; real-world underdosing remains a separate issue from labeled dose adjustment.
From a mechanistic perspective, the findings are plausible. Warfarin management can be more challenging in smaller patients and in settings with variable time in therapeutic range, which likely contributed to the worse outcomes seen in Asians on warfarin. DOACs have more predictable pharmacokinetics, fewer food and drug interactions, and do not depend on maintaining a narrow INR window. These advantages may translate into a larger absolute and relative benefit in populations where warfarin control is less stable.
For practice, the message is straightforward: do not assume that Asian patients should receive lower-dose DOACs by default. Instead, select the agent and dose according to approved labeling, renal function, age, weight, and drug interactions. In many Asian patients with nonvalvular atrial fibrillation, standard-dose DOACs appear to offer the best balance of stroke prevention and bleeding risk.
Conclusion
In this patient-level meta-analysis from COMBINE AF, Asian patients with atrial fibrillation had higher baseline thromboembolic risk and poorer warfarin control than non-Asians. Standard-dose DOACs provided at least as much benefit as in non-Asians, and in several outcomes the benefit was greater. Importantly, standard-dose DOACs did not increase gastrointestinal bleeding in Asians, while lower-dose DOACs were associated with more ischemic events than standard-dose therapy. The overall evidence supports standard-dose DOACs as the preferred anticoagulant strategy for most Asian patients with atrial fibrillation, provided there is no label-based reason to reduce the dose or avoid the drug class.
Funding and ClinicalTrials.gov
The PubMed abstract does not specify a dedicated funding statement for this analysis. COMBINE AF is a pooled patient-level meta-analysis of previously completed randomized trials, so no new interventional ClinicalTrials.gov registration applies to the present analysis.
References
1. Chao TF, Braunwald E, Palazzolo MG, et al. Direct oral anticoagulants vs warfarin in Asian vs non-Asian patients with atrial fibrillation: a patient-level meta-analysis from COMBINE AF. European Heart Journal. 2026. PMID: 42059513.
2. January CT, Wann LS, Calkins H, et al. 2023 ACC/AHA/ACCP/HRS Guideline for the Diagnosis and Management of Atrial Fibrillation. Circulation. 2024;149:e1-e156.
3. Hindricks G, Potpara T, Dagres N, et al. 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the EACTS. European Heart Journal. 2021;42:373-498.