Racial Differences in Bleeding Risks among Patients with Atrial Fibrillation: An Ecological Epidemiological Study Comparing Korea and UK Population

Abstract

Background

Racial differences in the propensity to bleeding may be evident, with a higher risk of bleeding in Asian populations. This study aimed to assess racial differences in bleeding risk among patients with atrial fibrillation (AF) using an ecological epidemiological approach.

Methods

We included patients with AF from the Korean National Health Insurance Service-Health Screening and UK Biobank who underwent health check-ups between 2006 and 2010. The analysis involved 1928 East Asians (62.1% male, median age 60.0 years) and 5917 White Europeans (71.4% male, median age 64.0 years) were analyzed. Primary outcome was composed of intracranial hemorrhage and bleeding from the gastrointestinal, respiratory, and genitourinary systems.

Results

During follow-up, the primary outcome occurred in 126 East Asians and 587 White Europeans. East Asians had a 42% lower 5-year incidence rate compared with White Europeans (weighted incidence rate 1.31 vs. 2.24 per 100 person-years; incidence rate ratio 0.58, 95% confidence interval: 0.41–0.83). Contrary to the primary outcome, the incidence of intracranial hemorrhage tended to be higher among East Asians (weighted incidence rate 0.34 vs. 0.14 per 100 person-years; incidence rate ratio 2.36, 95% confidence interval: 0.88–6.37). These results persisted even in patients naïve to antithrombotic drugs. East Asians who were already taking antithrombotic drugs at baseline showed no significant difference in the incidence of the primary outcome compared with White Europeans.

Conclusion

This ecological study highlights racial differences in the incidence of bleeding influenced by anatomical site and antithrombotic drug use and underscores the necessity for race-based tailored approaches.

 

Scroll to top

Keywords

racial difference - atrial fibrillation - intracranial hemorrhage - gastrointestinal bleeding

Introduction

Atrial fibrillation (AF) is one of the most common arrhythmias requiring treatment, and stroke prevention with oral anticoagulation therapy (OAC) is one of the main pillars of AF management, as recommended in guidelines.[1] [2] While OAC is used to prevent stroke and systemic embolism associated bleeding events can also increase the risk of morbidity and mortality, and in cases of serious bleeding, creates uncertainty in treatment options.[3] Various strategies have been proposed to reduce bleeding risk, including the mitigation of modifiable risk factors, reducing anticoagulant dosage, or considering alternative procedures, such as left atrial appendage occlusion.[4]

However, East Asians have a different thromboembolic/bleeding profile from Western population, sometimes referred to as the East Asian Paradox, suggesting a need for medical planning and resource allocation that reflects race-specific characteristics for long-term bleeding risk management.[5] [6] Yet, the 1.5 billion East Asians have been underrepresented in the pivotal cardiovascular trials, complicating the generalization of research findings; furthermore, there remains a lack of studies based on reliable statistical methods to clarify interracial differences.[7]

We designed an ecological epidemiological study based on a cohort of approximately 1 million individuals extracted from South Korea and the United Kingdom to investigate the overall incidence of bleeding in patients with AF and validate the differences between the two races (East Asian vs. White European) using various analyses, and assessed bleeding risk by anatomical site. This ecological study features the characteristics of observational epidemiological studies that associate exposure and disease at the population level, rather than at the individual level.[5] [8]

 

Scroll to top

Methods

This retrospective study included patient-level data sourced from the Korean National Health Insurance Service-Health Screening (K-NHIS-HealS) and the UK Biobank, detailed information of which is consistent with those described in previous studies ([Supplementary Methods], available in the online version).[9] [10] This study received approval from the Institutional Review Board of Yonsei University Health System (4-2022-1241) and was conducted in accordance with the Declaration of Helsinki. For the K-NHIS-HealS, identifiers were removed after cohort generation, thus waiving the need for informed consent. In the case of the UK Biobank, the use of its resources was approved by the North West Multicenter Research Ethics Committee (REC approval 21/NM/0157, application 77793), with the consent of all participants in the cohort.

Selection of Participants

The detailed process was demonstrated in [Fig. 1]. This study standardized the enrollment periods for two cohorts by excluding 215,331 K-NHIS-HealS participants who underwent health check-ups in 2005, 2011, and 2012, all aged 18 years and older. Additionally, 59,898 UK Biobank participants who did not identify as “British Whites” were excluded. From both K-NHIS-HealS and the UK Biobank, participants either younger than 40 or older than 70 at enrollment (31,633 and 9422, respectively), those who died within 180 days of enrollment (204 and 331, respectively), and those never diagnosed with AF (208,414 and 426,853, respectively) were excluded from the analysis.

Fig. 1 Flow chart of enrollment and analysis of the participants. This study standardized the enrollment periods for two cohorts by excluding 215,331 K-NHIS-HealS participants who underwent health check-ups in 2005, 2011, and 2012, all aged 18 years and older. Additionally, 59,898 UK Biobank participants who did not identify as “British Whites” were excluded. From both K-NHIS-HealS and the UK Biobank, participants either younger than 40 or older than 70 at enrollment (31,633 and 9,422, respectively), those who died within 180 days of enrollment (204 and 331, respectively), and those never diagnosed with AF (208,414 and 426,853, respectively) were excluded from the analysis. After applying the exclusion criteria, 1,928 East Asians and 5,917 White Europeans with AF at baseline were ultimately included in the study. AF, atrial fibrillation; K-NHIS-HealS, Korean National Health Insurance Service-Health Screening.

 

Scroll to top

Covariates and Outcomes

The definitions of baseline characteristics for both cohorts are consistent with those used in previous studies ([Supplementary Table S1], available in the online version).[5] [8] Sex and race were determined based on self-reported data. Diagnostic and procedural codes from hospital visits were employed to identify underlying comorbidities, and information on medication use for cardiovascular diseases was collected from linked general practitioner electronic health records, incorporating both self-report and prescription data.[11]

The primary outcome encompassed a composite of intracranial hemorrhage (ICH) and bleeding from the gastrointestinal, respiratory, and genitourinary systems. As secondary outcomes, we not only analyzed each component of the primary outcome but also considered major bleeding, defined as a composite of ICH, gastrointestinal bleeding, anemia due to bleeding, and bleeding-related death. All outcomes were defined using established ICD-10 codes from the K-NHIS-HealS and UK Biobank claims data, as well as related death records ([Supplementary Table S2], available in the online version).[8] [11] These codes had to be in the primary position; for ICH, codes for accompanying brain imaging procedures were also required. Our analysis focused on the risk of initial bleeding events, rather than recurrent bleeding events, among the patients.[12]

 

Scroll to top

Statistical Analyses

The baseline characteristics of the patients were described using descriptive statistics. To evaluate differences in baseline characteristics between the two racial groups and achieve balance, overlap weighting based on propensity scores was applied. The propensity score was calculated using logistic regression that included age, sex, AF duration, systolic/diastolic blood pressure, underlying comorbidities, and concurrent medication use as covariates. The distribution of propensity scores for the two racial groups, before and after overlap weighting, was depicted in [Supplementary Fig. S1] (available in the online version). A standardized mean difference of less than 0.1 was considered indicative of balance between the two racial groups for that covariate. To account for differences in follow-up duration between the two national cohorts, the 5-year incidence rates (IRs) were calculated by dividing the number of first incident events by the total person-years over five years from enrollment, and reported per 100 person-years. To compare incidence between the two racial groups, weighted IRs, incidence rate ratios (IRRs), and 95% confidence intervals (CIs) were estimated for the overlap-weighted sample, using White Europeans in the UK as the reference group.

To assess the impact of antithrombotic agents (i.e., warfarin, aspirin, P₂Y₁₂ inhibitors) prescribed for stroke prevention on bleeding events, the additional analysis was conducted by dividing patients into those who were not taking these medications at baseline and those who were. For the former group, data were censored from the date they were prescribed antithrombotic agents during follow-up. A chi-square test was performed to determine whether the composition patterns of antithrombotic agents prescribed during follow-up differed between the two races. For the latter group, to explore the effect of antithrombotic agent types on bleeding risk, patients at baseline were further categorized into those taking only warfarin and those taking only antiplatelet agents. A Cox proportional hazards model with an interaction term for race and type of antithrombotic agent (warfarin or antiplatelet agents) was then fitted to the weighted population.

In the sensitivity analysis, to address the imbalance in baseline characteristics between the two racial groups, one-to-one propensity score matching was employed instead of overlap weighting. Second, similar to the approach for K-NHIS-HealS, only ICD-10 codes were used to define underlying comorbidities in the UK Biobank. Third, instead of using the 5-year IRs, the IRs were compared based on the number of events that occurred throughout the entire follow-up period and the total person-years for both cohorts. Lastly, subgroup analyses were conducted considering age, sex, Charlson Comorbidity Index, and obesity at baseline. When defining obesity, the body mass index (BMI) criteria suggested by Korean clinical guidelines were used for East Asians, whereas the BMI criteria suggested by World Health Organization guidelines were used for White Europeans.[13] [14] All tests were two-tailed, and statistical significance was set at p < 0.05. All analyses were performed using R version 4.2.1 (The R Foundation, www.r-project.org).

 

Scroll to top

 

Scroll to top

Results

Baseline Characteristics

After applying the exclusion criteria, 1928 East Asians and 5917 White Europeans with AF at baseline were ultimately included in the study. The median follow-up period (interquartile range, IQR) for East Asians was 80.6 months (61.6–88.2 months), and for White Europeans, it was 138.7 months (129.0–148.5 months). The median (IQR) age at baseline was 60.0 (54.0–65.0) years for East Asians, and 64.0 (60.0–67.0) years for White Europeans, as shown in [Table 1]. The proportion of males was relatively lower in the East Asian in Korea compared with the White Europeans (62.1 vs. 71.4%, p < 0.001). East Asians had a higher burden of comorbidities such as hypertension, diabetes mellitus, ischemic stroke, myocardial infarction, and heart failure, and were more likely to use medications for cardiovascular diseases. Although the proportion of patients using warfarin was lower among East Asians compared with White Europeans (13.6 vs. 16.2%, p = 0.01), the proportion of patients using aspirin (37.1 vs. 7.0%, p < 0.001) and P₂Y₁₂ inhibitors (6.2 vs. 1.3%, p < 0.001) were significantly higher. As presented in [Supplementary Table S3] (available in the online version), after applying overlap weighting, balance was achieved between the two racial groups for all baseline characteristics, except for BMI.

Table 1 Baseline characteristics of study participants by race

Characteristics	East Asians	White Europeans	p-Value
	(N = 1,928)	(N = 5,917)
Follow-up duration (mo)	80.6 (61.6–88.2)	138.7 (129.0–148.5)	<0.001
AF duration (mo)	32.4 (17.1–50.2)	48.3 (22.0–86.5)	<0.001
Age (y)	60.0 (54.0–65.0)	64.0 (60.0–67.0)	<0.001
Male sex	1,197 (62.1)	4,223 (71.4)	<0.001
Body mass index (kg/m2)	24.5 (22.5–26.7)	28.3 (25.5–31.8)	<0.001
<18.5	32 (1.7)	18 (0.3)
18.5–23	543 (28.2)	505 (8.6)
23–25	510 (26.5)	700 (11.9)
25–30	748 (38.8)	2,483 (42.3)
≥30	95 (4.9)	2,163 (36.9)
Systolic blood pressure (mm Hg)	126.0 (115.0–137.0)	136.0 (124.0–149.5)	<0.001
Diastolic blood pressure (mm Hg)	80.0 (70.0–86.0)	81.5 (74.0–89.0)	<0.001
CHA2DS2-VASc score	2.0 (1.0–3.0)	2.0 (1.0–3.0)	<0.001
Medical history
Hypertension	1,550 (80.4)	3,848 (65.0)	<0.001
Diabetes mellitus	364 (18.9)	786 (13.3)	<0.001
Dyslipidemia	1,218 (63.2)	2,285 (38.6)	<0.001
Ischemic stroke	374 (19.4)	358 (6.1)	<0.001
Myocardial infarction	222 (11.5)	617 (10.4)	0.19
Heart failure	580 (30.1)	890 (15.0)	<0.001
Peripheral arterial disease	174 (9.0)	113 (1.9)	<0.001
Chronic kidney disease	59 (3.1)	362 (6.1)	<0.001
End-stage kidney disease	1 (0.1)	50 (0.8)	<0.001
COPD	133 (6.9)	294 (5.0)	0.001
Malignancy	351 (18.2)	651 (11.0)	<0.001
Concurrent medication
Antithrombotic agents	946 (49.1)	1,358 (23.0)	<0.001
Warfarin	262 (13.6)	958 (16.2)	0.01
Aspirin	716 (37.1)	414 (7.0)	<0.001
P2Y12 inhibitors	120 (6.2)	74 (1.3)	<0.001
Statins	346 (17.9)	665 (11.2)	<0.001
ACEi/ARB	599 (31.1)	700 (11.8)	<0.001
DHP CCB	386 (20.0)	233 (3.9)	<0.001
Non-DHP CCB	158 (8.2)	115 (1.9)	<0.001
Beta blocker	509 (26.4)	598 (10.1)	<0.001
Loop diuretics	509 (26.4)	307 (5.2)	<0.001
K+ sparing diuretics	107 (5.5)	0 (0.0)	<0.001
Class IC AADs	77 (4.0)	126 (2.1)	<0.001
Class III AADs	65 (3.4)	161 (2.7)	0.16

Abbreviations: AAD, antiarrhythmic drug; ACEi, angiotensin converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; CCB, calcium channel blocker; COPD, chronic obstructive pulmonary disease; DHP, dihydropyridine.

Notes: Data are presented as medians (interquartile range) or N (%). Continuous variables were analyzed using the Wilcoxon rank-sum test, and categorical variables were analyzed using the chi-square test or Fisher's exact test.

 

Scroll to top

Racial Differences in Incidence Rates of Bleeding Events

Over the 5 years following enrollment, 126 primary outcome events were reported among East Asians and 587 among White Europeans, respectively ([Table 2]; [Fig. 2]). Accounting for differences in baseline characteristics, East Asians had a 42% lower incidence of the primary outcome compared with White Europeans (weighted IR 1.31 per 100 person-years vs. 2.24 per 100 person-years; IRR 0.58, 95% CI: 0.41–0.83). No significant differences in major bleeding events were observed between the two racial groups (weighted IR: 0.98 per 100 person-years vs. 1.18 per 100 person-years; IRR: 0.83, 95% CI: 0.54–1.29).

Table 2 Five-year incidence rates for primary and secondary outcomes by race

	East Asians	White Europeans
	(N = 1,928)	(N = 5,917)
Primary outcome
Number of events	126	587
Person-years	8,816	27,455
Crude incidence rate (95% CI)[a]	1.43 (1.18–1.68)	2.14 (1.97–2.31)
Weighted incidence rate (95% CI)[a]	1.31 (0.94–1.67)	2.24 (1.76–2.71)
Incidence rate ratio (95% CI)[b]	0.58 (0.41–0.83)	1 [Reference]
Secondary outcome
Major bleeding
Number of events	101	291
Person-years	8,884	28,181
Crude incidence rate (95% CI)[a]	1.14 (0.92–1.36)	1.03 (0.91–1.15)
Weighted incidence rate (95% CI)[a]	0.98 (0.66–1.30)	1.18 (0.83–1.52)
Incidence rate ratio (95% CI)[b]	0.83 (0.54–1.29)	1 [Reference]
Intracranial hemorrhage
Number of events	26	41
Person-years	9,027	28,753
Crude incidence rate (95% CI)[a]	0.29 (0.18–0.40)	0.14 (0.10–0.19)
Weighted incidence rate (95% CI)[a]	0.34 (0.15–0.52)	0.14 (0.02–0.26)
Incidence rate ratio (95% CI)[b]	2.36 (0.88–6.37)	1 [Reference]
Bleeding from gastrointestinal system
Number of events	59	254
Person-years	8,969	28,237
Crude incidence rate (95% CI)[a]	0.66 (0.49–0.83)	0.90 (0.79–1.01)
Weighted incidence rate (95% CI)[a]	0.49 (0.27–0.72)	1.03 (0.71–1.35)
Incidence rate ratio (95% CI)[b]	0.48 (0.28–0.83)	1 [Reference]
Bleeding from respiratory system
Number of events	19	129
Person-years	9,049	28,543
Crude incidence rate (95% CI)[a]	0.21 (0.12–0.30)	0.45 (0.37–0.53)
Weighted incidence rate (95% CI)[a]	0.21 (0.07–0.36)	0.49 (0.27–0.71)
Incidence rate ratio (95% CI)[b]	0.43 (0.19–0.99)	1 [Reference]
Bleeding from genitourinary system
Number of events	30	211
Person-years	9,000	28,302
Crude incidence rate (95% CI)[a]	0.33 (0.21–0.45)	0.75 (0.64–0.85)
Weighted incidence rate (95% CI)[a]	0.30 (0.12–0.47)	0.76 (0.48–1.03)
Incidence rate ratio (95% CI)[b]	0.39 (0.20–0.78)	1 [Reference]

Abbreviation: CI, confidence interval.

^a The incidence rates were calculated by dividing the number of first incident events by the total person-years over 5 years from enrollment, with 95% confidence intervals estimated using a Poisson distribution and reported per 100 person-years. Weighted incidence rates were calculated, incorporating each patient's follow-up duration and event occurrence, adjusted by their assigned weights.

^b The incidence rate ratios were calculated by dividing the weighted incidence rate of East Asians by that of White Europeans. The 95% confidence intervals were derived by calculating the standard error based on the weighted number of events in each racial group.

Fig. 2 Bar graph of weighted incidence rates by race. The weighted incidence rates for the primary outcome and each component of the secondary outcome in East Asians (N = 1,928) and White Europeans (N = 5,917) are displayed. Error bars represent the 95% confidence intervals for each weighted incidence rate, corresponding to the values shown in [Table 2]. ICH, intracranial hemorrhage; IRR, incidence rate ratio.

This variation was attributable to contrasting race-specific tendencies in bleeding. The incidence of ICH was over 2-fold higher in East Asians compared with White Europeans (weighted IR: 0.34 per 100 person-years vs. 0.14 per 100 person-years; IRR: 2.36, 95% CI: 0.88–6.37), whereas the incidence of bleeding from the gastrointestinal system was 52% lower in East Asians (weighted IR: 0.49 per 100 person-years vs. 1.03 per 100 person-years; IRR: 0.48, 95% CI: 0.28–0.83). Bleeding from the respiratory system (weighted IR: 0.21 per 100 person-years vs. 0.49 per 100 person-years; IRR: 0.43, 95% CI: 0.19–0.99) and the genitourinary system (weighted IR: 0.30 per 100 person-years vs. 0.76 per 100 person-years; IRR: 0.39, 95% CI: 0.20–0.78) also occurred 57 and 61% less frequently in East Asians, contributing to the results in the primary outcome.

The anatomical dependence of bleeding tendencies was also reflected in their relative proportions ([Supplementary Fig. S2], available in the online version). The total number of all reported bleeding events was 134 in East Asians and 635 in White Europeans. While both cohorts predominantly experienced bleeding from the gastrointestinal system and genitourinary system, followed by the third most common event being ICH in East Asians (19.4 vs. 6.5%), and bleeding from the respiratory system in White Europeans (14.2 vs. 20.3%).

 

Scroll to top

The Impact of Antithrombotic Agents on Bleeding Events

At baseline, 982 East Asians and 4,559 White Europeans were not using antithrombotic agents. During follow-up, 346 of these East Asians received a total of 452 antithrombotic prescriptions (99 for warfarin, 203 for aspirin, and 150 for P₂Y₁₂ inhibitors), and 859 of these White Europeans received a total of 1,130 prescriptions (648 for warfarin, 316 for aspirin, and 166 for P₂Y₁₂ inhibitors). Differences in the composition patterns of prescribed antithrombotic agents between the two races were also seen (p < 0.001). Even when censoring for the use of antithrombotic agents during follow-up, East Asians still exhibited a lower overall IR for the primary outcome compared with White Europeans (weighted IR: 0.79 per 100 person-years vs. 1.98 per 100 person-years; IRR: 0.40, 95% CI: 0.23–0.70; [Fig. 3]). While events of ICH were still relatively more frequent (weighted IR: 0.17 per 100 person-years vs. 0.08 per 100 person-years; IRR: 2.08, 95% CI: 0.37–11.60), whereas events of bleeding from the gastrointestinal system were less frequent (weighted IR: 0.33 per 100 person-years vs. 0.98 per 100 person-years; IRR: 0.34, 95% CI: 0.15–0.79), consistent with the main findings across all outcomes ([Supplementary Fig. S3], available in the online version).

Fig. 3 Results on the primary outcome by race according to the use of antithrombotic agents. To analyze the results censoring for the use of antithrombotic agents, 982 East Asians and 4,559 White Europeans not using antithrombotic agents at baseline were included. To analyze the results while on antithrombotic agents, 946 East Asians and 1,358 White Europeans who were already using antithrombotic agents at baseline were included. Additionally, separate analyses were performed for those using only warfarin (186 East Asians, 903 White Europeans) and those using only antiplatelet agents (684 East Asians, 400 White Europeans) at baseline. Error bars indicate 95% confidence intervals for the incidence rate ratios. ^aWeighted incidence rates were calculated by dividing the number of first incident events by the total person-years over 5 years from enrollment, with adjustments for assigned weights. The 95% confidence intervals were estimated using a Poisson distribution and reported per 100 person-years. ^bThe incidence rate ratios were calculated by dividing the weighted incidence rate of East Asians by that of White Europeans. The 95% CIs were derived by calculating the standard error based on the weighted number of events in each racial group. ^cThe results of the Cox proportional hazards model, including the interaction term for race and type of antithrombotic agent in the weighted population, showed no significant interaction (p for interaction = 0.26). CI, confidence interval.

For patients who were already prescribed antithrombotic agents at baseline, the relatively lower incidence of the primary outcome in East Asians was attenuated (weighted IR: 2.04 per 100 person-years vs. 2.36 per 100 person-years; IRR: 0.86, 95% CI: 0.48–1.55), and the higher incidence of ICH became more prominent (weighted IR: 0.49 per 100 person-years vs. 0.11 per 100 person-years; IRR: 4.54, 95% CI: 0.61–33.62). Detailed results are presented in [Fig. 3] and [Supplementary Fig. S3] (available in the online version). An exploratory analysis confined to OAC users only and antiplatelet users only at baseline, did not show significant statistical interactions between the two antithrombotic strategies, except for bleeding from the respiratory system, which had a very small number of events (p for interaction < 0.001).

 

Scroll to top

Sensitivity and Subgroup Analyses

After successfully adjusting for differences in baseline characteristics between the two cohorts through one-to-one propensity score matching and conducting the same analysis ([Supplementary Tables S4, S5], available in the online version), using only ICD-10 codes to define comorbidities in the UK Biobank ([Supplementary Table S6], available in the online version), calculating IRs based on the total number of events and person-years over the entire follow-up period ([Supplementary Table S7], available in the online version), and analyzing subgroups ([Supplementary Figs. S4–S9], available in the online version), the results were consistent with the main findings.

 

Scroll to top

 

Scroll to top

Discussion

The principal findings of this ecological epidemiological study, using patient-level data from South Korea and the United Kingdom, are as follows: First, East Asians with AF had a relatively lower incidence of overall bleeding, as represented by the primary outcome, compared with White Europeans. Second, similar to the general population, ICH was more frequent in East Asians than in White Europeans. Third, current users of antithrombotic agents for stroke prevention due to AF were associated with an attenuation of the relatively lower bleeding risk observed in East Asians, whereas the higher incidence of ICH became more prominent.

Racial Differences in Bleeding Tendencies and Patterns

Our study demonstrates that East Asians have a lower overall bleeding incidence compared with White Europeans, with notable variations based on the anatomical origin of the bleeding. Past studies have focused on the burden of cardiovascular comorbidities and the pharmacokinetic and pharmacodynamic characteristics of antithrombotic agents to explain the different bleeding tendencies among races.[6] Importantly, approaches targeting current users to find clues about racial differences are not free from the bias of selecting “survivors” with little or no bleeding tendencies. Additionally, the results of this study, which censored antithrombotic prescription initiation to provide a pure perspective on bleeding tendencies in untreated populations, support the need for further exploration of racial characteristics per se.

A lack of consistent methodologies and individual-level data has hindered comparisons between races.[15] Our series of studies indicates that classifying bleeding events by anatomical origin may provide clues to explaining this phenomenon.[8] Similar to the general population, the incidence of ICH was higher in East Asians, whereas the incidence of bleeding from the gastrointestinal system was lower in East Asians. The susceptibility of East Asians to ICH is a well-known phenomenon from pivotal trials, although the specific mechanisms remain unclear.[16] [17] Several factors are thought to be related. Intracranial atherosclerosis is about twice as prevalent in East Asians (30–50 vs. 15–30%),[18] and the prevalence of microbleeds, which can increase the risk of ICH by about 3-fold, is also significantly higher in Asian patients compared with non-Asian patients (68.4 vs. 56.9%).[19] Additionally, the incidence of cerebral arteriovenous malformations has been reported to be higher in Asians compared with Western populations,[20] all of which may contribute to the increased likelihood of ICH in East Asians. Additionally, genetic polymorphisms in the ACE, RAGE, and CD36 genes—implicated in vascular constriction, vascular remodeling, and endothelial dysfunction—have been reported to contribute to the development of hypertension and atherosclerosis, and are associated with an increased risk of ICH.[21] However, these genetic variants alone are insufficient to fully explain the risk of ICH in East Asians.

Regarding bleeding from the gastrointestinal system, the incidence of Helicobacter pylori infection is gradually decreasing in East Asians due to regular endoscopic examinations, which is common because of the high prevalence of gastric cancer.[22] In contrast, diverticulosis is still more commonly reported in Western populations, partially contributing to the differences in bleeding incidence between races.[23] From a public health perspective, effectively preventing bleeding events with limited medical resources requires considering race-specific epidemiological information on bleeding, including the anatomical origins of bleeding. In Korea, the government provides biennial upper endoscopy as a reimbursable service under the national health insurance for individuals aged over 40 years, serving as an example of a public health policy aimed at risk factor management.[24] This approach can ultimately lead to improved long-term outcomes.[1] [25]

 

Scroll to top

Influence of Antithrombotic Agents on Racial Differences in Bleeding Risk

First, it should be noted that White Europeans were more commonly treated with warfarin (16.2%) as an antithrombotic strategy, whereas East Asians primarily used antiplatelet agents such as aspirin (37.1%) and P₂Y₁₂ inhibitors (6.2%). A previous individual-patient data meta-analysis reported that when compared with aspirin, warfarin reduced the risk of stroke by 38%, while doubling the risk of major bleeding.[26] The fact that warfarin did not reduce vascular- or all-cause death led physicians to prescribe warfarin to patients with relatively high stroke risk and to use aspirin for those with lower risk.[26] Although a randomized clinical trial in elderly AF patients showed the noninferiority of warfarin compared with aspirin for bleeding risks,[27] several limitations—such as focusing only on Europeans, a low number of bleeding events, frequent crossover between treatment strategies, and a high proportion of patients already on warfarin at enrollment—combined with the reasons described below, were insufficient to alleviate the concerns of physicians managing the specific AF population subgroup referred to as East Asians.

Despite a tendency to maintaining lower international normalized ratio (INR) values, it is well-known that East Asians have a higher incidence of bleeding, particularly ICH, compared with White Europeans, and the sequelae and mortality of bleeding also tended to be greater.[28] Although the exact reasons for this phenomenon are not clearly understood, enhanced pharmacokinetics of warfarin in East Asians is often considered. The lower frequency of variant alleles such as CYP2C9*2 and CYP2C9*3, which are known to reduce the catalytic activity of cytochrome P450 2C9, in Japanese results in a reduced maintenance dose and increased bleeding.[29] Additionally, the haplotype frequency of the gene affecting hepatic vitamin K epoxide reductase, which influences the synthesis of various coagulation factors, was 89% in East Asians compared with only 42% in Caucasians, partly explaining the narrower therapeutic index of warfarin in East Asians.[30] Beyond genetic predisposition, it is also argued that the lower lean body mass, faster rate of renal function decline, and the relatively frequent intake of herbal materials in East Asian countries affects the blood levels of warfarin, making it difficult to maintain an appropriate time in therapeutic range.[31] [32] Reflecting these concerns, Japanese guidelines currently recommend an optimal INR of 1.6 to 2.6 for elderly AF patients over the age of 70.[33] Similarly, for nonvitamin K antagonist oral anticoagulants (NOACs), the differences in thrombotic and bleeding profiles observed between races in this study may generate the hypothesis of defining race-specific standard doses—illustrated by the example of Japan, where pharmacokinetic modeling data supported adopting a 15 mg standard dose of rivaroxaban instead of 20 mg.[1]

Antiplatelet agents are not effective alternatives for East Asians. The BAFTA and ACTIVE W studies revealed that antiplatelet agents do not reduce bleeding events compared with warfarin and were inferior for stroke prevention.[27] Moreover, for some P₂Y₁₂ inhibitors, the dose required to reach the same blood concentration was lower in East Asians than in Caucasians, leading to more frequent ICH and bleeding from gastrointestinal system in East Asians, supporting the findings of this study.[18] The most preferred NOACs in contemporary guidelines have demonstrated greater reductions in the risks of major bleeding, clinically relevant nonmajor bleeding, and gastrointestinal bleeding in Asians compared with warfarin.[34] However, the fact remains that the absolute ICH incidence in Asians taking NOACs is still higher than in Caucasians, highlighting the necessity for race-specific bleeding prevention strategies that should not be overlooked in the advancement of treatment options.[6] [28]

 

Scroll to top

 

Scroll to top

Limitations

Our study has several limitations. First, due to the nature of an ecological observational study, residual confounding factors may still exist. Because this study was based on claim-based data, it was not possible to obtain detailed clinical information such as laboratory test results (e.g., time in therapeutic range, hemoglobin levels during bleeding events), specific imaging findings, the success of hemostatic procedures or surgeries, or the presence of shock during bleeding events. As a result, there are inherent limitations in assessing bleeding severity and in establishing causality based on the observed results. Second, we relied on ICD-10 codes to define comorbidities, which introduces the possibility of coding errors. Additionally, the use of self-reported information to define comorbidities and medication use may result in further measurement errors. Third, unlike the K-NHIS-HealS data, which is considered representative of the general population, the UK Biobank data consists of voluntary participants who may not represent the general population.[35] Fourth, participants younger than 40 years or older than 70 years were excluded to ensure appropriate matching between the two racial groups and to minimize the influence of extreme values. Therefore, our findings cannot be generalized to these excluded age groups. Fifth, there was no information provided on the changes in anthropometric and comorbidity variables over time. Sixth, the East Asians in this study all resided in Korea, limiting the applicability of our conclusions to Asians in other regions.

 

Scroll to top

Conclusion

This ecological epidemiological study of patients with AF also showed that East Asians tend to have fewer overall bleeding events compared with Caucasians. However, East Asians still exhibited a greater susceptibility to ICH, and the lower overall bleeding incidence in East Asians was attenuated in prevalent users of antithrombotic agents. This suggests that the anatomical origin of bleeding and the use of antithrombotic agents are significant factors in explaining racial differences in bleeding tendencies.

 

Scroll to top

 

Scroll to top

Conflict of Interest

B.J. has served as a speaker for Bayer, BMS/Pfizer, Medtronic, and Daiichi-Sankyo and received research funds from Medtronic and Abbott. No fees have been received directly or personally. G.Y.H.L. has been a consultant and speaker for BMS/Pfizer, Boehringer Ingelheim, Daiichi-Sankyo, and Anthos. No fees are received personally. He is a National Institute for Health and Care Research (NIHR) Senior Investigator and co-PI of the AFFIRMO project on multimorbidity in AF (grant agreement number: 899871), TARGET project on digital twins for personalized management of atrial fibrillation and stroke (grant agreement number: 101136244) and ARISTOTELES project on artificial intelligence for management of chronic long-term conditions (grant agreement number: 101080189), which are all funded by the EU's Horizon Europe Research and Innovation program. Other authors have nothing to declare.

Acknowledgment

We would like to thank the National Health Insurance Service of Korea for their cooperation.

Data Availability Statement

Data from K-NHIS-HealS are available on the Web site (http://nhiss.nhis.or.kr). Data from UK Biobank are available upon application.

Authors' Contribution

D.S.K. and P.S.Y. contributed equally to this work. G.Y.H.L. and B.J. contributed to the conception and design of the work and critical revision of the manuscript. D.S.K. contributed to the conception and design of the work, analysis and interpretation of data, and drafting of the manuscript. P.S.Y., D.K., and E.J. contributed to data extraction and analysis. H.T.Y., T.H.K., J.H.S., and H.N.P. contributed to the conception and design of the work and revision of the manuscript. All authors approved the final version to be published. Also, they are responsible for ensuring that questions relating to accuracy or integrity of all parts of the work are properly investigated and resolved. The corresponding author certifies that all listed authors meet authorship criteria.

^* These authors contributed equally to this work.

^** These authors are joint senior authors.

The review process for this paper was fully handled by Christian Weber, Editor in Chief.

• References

• 1 Chao TF, Joung B, Takahashi Y. et al. 2021 focused update consensus guidelines of the Asia Pacific Heart Rhythm Society on stroke prevention in atrial fibrillation: executive summary. Thromb Haemost 2022; 122 (01) 20-47
• 2 Wang Y, Guo Y, Qin M. et al. Chinese expert consensus guidelines on the diagnosis and treatment of Atrial Fibrillation in the Elderly. Executive Summary. Thromb Haemost 2024; 2024
• 3 Ivany E, Lotto RR, Lip GYH, Lane DA. Managing uncertainty: physicians' decision making for stroke prevention for patients with atrial fibrillation and intracerebral hemorrhage. Thromb Haemost 2022; 122 (09) 1603-1611
• 4 Gorog DA, Gue YX, Chao TF. et al. Assessment and Mitigation of Bleeding Risk in Atrial Fibrillation and Venous Thromboembolism: Executive Summary of a European and Asia-Pacific Expert Consensus Paper. Thromb Haemost 2022; 122 (10) 1625-1652
• 5 Kang DS, Yang PS, Kim D. et al. Racial differences in ischemic and hemorrhagic stroke: an ecological epidemiological study. Thromb Haemost 2024; 124 (09) 883-892
• 6 Kim HK, Tantry US, Smith Jr SC. et al. The East Asian paradox: an updated position statement on the challenges to the current antithrombotic strategy in patients with cardiovascular disease. Thromb Haemost 2021; 121 (04) 422-432
• 7 Norby FL, Benjamin EJ, Alonso A, Chugh SS. Racial and ethnic considerations in patients with atrial fibrillation: JACC Focus Seminar 5/9. J Am Coll Cardiol 2021; 78 (25) 2563-2572
• 8 Kang DS, Yang PS, Kim D. et al. Racial differences in bleeding risk: an ecological epidemiological study comparing Korea and United Kingdom subjects. Thromb Haemost 2024; 124 (09) 842-851
• 9 Lee SS, Ae Kong K, Kim D. et al. Clinical implication of an impaired fasting glucose and prehypertension related to new onset atrial fibrillation in a healthy Asian population without underlying disease: a nationwide cohort study in Korea. Eur Heart J 2017; 38 (34) 2599-2607
• 10 Sudlow C, Gallacher J, Allen N. et al. UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med 2015; 12 (03) e1001779
• 11 Khurshid S, Weng LC, Al-Alusi MA. et al. Accelerometer-derived physical activity and risk of atrial fibrillation. Eur Heart J 2021; 42 (25) 2472-2483
• 12 Chao TF, Liu CJ, Tuan TC. et al. Lifetime risks, projected numbers, and adverse outcomes in Asian patients with atrial fibrillation: a report from the Taiwan nationwide AF cohort study. Chest 2018; 153 (02) 453-466
• 13 Kim BY, Kang SM, Kang JH. et al; Committee of Clinical Practice Guidelines, Korean Society for the Study of Obesity (KSSO). 2020 Korean Society for the study of obesity guidelines for the management of obesity in Korea. J Obes Metab Syndr 2021; 30 (02) 81-92
• 14 Wharton S, Lau DCW, Vallis M. et al. Obesity in adults: a clinical practice guideline. CMAJ 2020; 192 (31) E875-E891
• 15 Feigin VL, Krishnamurthi RV, Parmar P. et al; GBD 2013 Writing Group, GBD 2013 Stroke Panel Experts Group. Update on the Global Burden of Ischemic and Hemorrhagic Stroke in 1990-2013: the GBD 2013 Study. Neuroepidemiology 2015; 45 (03) 161-176
• 16 Yamashita T, Koretsune Y, Yang Y. et al. Edoxaban vs. Warfarin in East Asian Patients With Atrial Fibrillation - An ENGAGE AF-TIMI 48 Subanalysis. Circ J 2016; 80 (04) 860-869
• 17 Goto S, Zhu J, Liu L. et al; ARISTOTLE Investigators. Efficacy and safety of apixaban compared with warfarin for stroke prevention in patients with atrial fibrillation from East Asia: a subanalysis of the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) trial. Am Heart J 2014; 168 (03) 303-309
• 18 Kim JS, Kim YJ, Ahn SH, Kim BJ. Location of cerebral atherosclerosis: why is there a difference between East and West?. Int J Stroke 2018; 13 (01) 35-46
• 19 Lovelock CE, Cordonnier C, Naka H. et al. Antithrombotic drug use, cerebral microbleeds, and intracerebral hemorrhage: a systematic review of published and unpublished studies. Stroke 2010; 41 (06) 1222-1228
• 20 Kim T, Kwon OK, Bang JS. et al. Epidemiology of ruptured brain arteriovenous malformation: a National Cohort Study in Korea. J Neurosurg 2018; 130 (06) 1965-1970
• 21 Querfurth H, Marczak I, Rahimian N, Jijakli A, Green-LaRoche D. Spontaneous intracranial hemorrhages in a community of Asian Americans: case series and literature review. Cureus 2024; 16 (01) e52128
• 22 Cha JM, Moon JS, Chung IK. et al. National endoscopy quality improvement program remains suboptimal in Korea. Gut Liver 2016; 10 (05) 699-705
• 23 Wang FW, Chuang HY, Tu MS. et al. Prevalence and risk factors of asymptomatic colorectal diverticulosis in Taiwan. BMC Gastroenterol 2015; 15: 40
• 24 Cho YK. How to improve the quality of screening endoscopy in Korea: national endoscopy quality improvement program. Clin Endosc 2016; 49 (04) 312-317
• 25 Romiti GF, Pastori D, Rivera-Caravaca JM. et al. Adherence to the ‘Atrial Fibrillation Better Care’ pathway in patients with atrial fibrillation: impact on clinical outcomes-a systematic review and meta-analysis of 285,000 patients. Thromb Haemost 2022; 122 (03) 406-414
• 26 van Walraven C, Hart RG, Singer DE. et al. Oral anticoagulants vs aspirin in nonvalvular atrial fibrillation: an individual patient meta-analysis. JAMA 2002; 288 (19) 2441-2448
• 27 Mant J, Hobbs FD, Fletcher K. et al; BAFTA investigators, Midland Research Practices Network (MidReC). Warfarin versus aspirin for stroke prevention in an elderly community population with atrial fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): a randomised controlled trial. Lancet 2007; 370 (9586) 493-503
• 28 Chao TF, Chen SA, Ruff CT. et al. Clinical outcomes, edoxaban concentration, and anti-factor Xa activity of Asian patients with atrial fibrillation compared with non-Asians in the ENGAGE AF-TIMI 48 trial. Eur Heart J 2019; 40 (19) 1518-1527
• 29 Takahashi H, Wilkinson GR, Caraco Y. et al. Population differences in S-warfarin metabolism between CYP2C9 genotype-matched Caucasian and Japanese patients. Clin Pharmacol Ther 2003; 73 (03) 253-263
• 30 Takahashi H, Wilkinson GR, Nutescu EA. et al. Different contributions of polymorphisms in VKORC1 and CYP2C9 to intra- and inter-population differences in maintenance dose of warfarin in Japanese, Caucasians and African-Americans. Pharmacogenet Genomics 2006; 16 (02) 101-110
• 31 Chan HT, So LT, Li SW, Siu CW, Lau CP, Tse HF. Effect of herbal consumption on time in therapeutic range of warfarin therapy in patients with atrial fibrillation. J Cardiovasc Pharmacol 2011; 58 (01) 87-90
• 32 Gage BF, Eby C, Johnson JA. et al. Use of pharmacogenetic and clinical factors to predict the therapeutic dose of warfarin. Clin Pharmacol Ther 2008; 84 (03) 326-331
• 33 JCS Joint Working Group. JCS 2013. Circ J 2014; 78 (08) 1997-2021
• 34 Chiang CE, Wang KL, Lin SJ. Asian strategy for stroke prevention in atrial fibrillation. Europace 2015; 17 (Suppl. 02) ii31-ii39
• 35 Fry A, Littlejohns TJ, Sudlow C. et al. Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. Am J Epidemiol 2017; 186 (09) 1026-1034

Publikationsverlauf

Eingereicht: 06. Mai 2025

Angenommen: 25. August 2025

Accepted Manuscript online:
26. August 2025

Artikel online veröffentlicht:
09. September 2025